AuroraMiddleware/skia2
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
507c08efb2
skia2/tests/PathOpsSkpTest.cpp

4774 lines
227 KiB
C++
Raw Normal View History

path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
/*
* Copyright 2012 Google Inc.
*
* Use of this source code is governed by a BSD-style license that can be
* found in the LICENSE file.
*/
rewrite includes to not need so much -Ifoo Current strategy: everything from the top Things to look at first are the manual changes: - added tools/rewrite_includes.py - removed -Idirectives from BUILD.gn - various compile.sh simplifications - tweak tools/embed_resources.py - update gn/find_headers.py to write paths from the top - update gn/gn_to_bp.py SkUserConfig.h layout so that #include "include/config/SkUserConfig.h" always gets the header we want. No-Presubmit: true Change-Id: I73a4b181654e0e38d229bc456c0d0854bae3363e Reviewed-on: https://skia-review.googlesource.com/c/skia/+/209706 Commit-Queue: Mike Klein <mtklein@google.com> Reviewed-by: Hal Canary <halcanary@google.com> Reviewed-by: Brian Osman <brianosman@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2019-04-23 17:05:21 +00:00
#include "tests/PathOpsExtendedTest.h"
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
#define TEST(name) { name, #name }
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpcheeseandburger_com225(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(555, 468);
path.lineTo(555, 362);
path.lineTo(872, 362);
path.lineTo(872, 468);
path.lineTo(555, 468);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(859.11792f, 397.320343f);
pathB.cubicTo(855.523071f, 399.691284f, 853.721191f, 402.40863f, 853.721191f, 405.552216f);
pathB.cubicTo(853.721191f, 407.911163f, 854.727478f, 410.115387f, 857.043518f, 412.252716f);
pathB.cubicTo(859.920532f, 414.916138f, 862.704773f, 417.086426f, 864.679382f, 418.852386f);
pathB.cubicTo(866.382446f, 420.371765f, 867.19104f, 422.108795f, 867.19104f, 423.506378f);
pathB.cubicTo(867.19104f, 424.551605f, 866.741821f, 425.539886f, 865.935242f, 426.281616f);
pathB.cubicTo(865.250366f, 426.910553f, 864.662415f, 427.339813f, 864.139282f, 427.4646f);
pathB.cubicTo(863.536377f, 427.605347f, 862.259521f, 426.491272f, 860.366821f, 424.208191f);
pathB.cubicTo(858.345276f, 421.770355f, 857.317017f, 419.733856f, 857.317017f, 417.98587f);
pathB.cubicTo(857.317017f, 417.198212f, 857.942993f, 415.930389f, 857.942993f, 415.930389f);
pathB.cubicTo(857.942993f, 415.930389f, 852.106018f, 421.296173f, 852.279663f, 422.549042f);
pathB.cubicTo(852.462402f, 423.890747f, 853.669312f, 425.703613f, 855.876465f, 428.252258f);
pathB.cubicTo(858.038818f, 430.754944f, 859.4953f, 431.840088f, 860.190125f, 431.594513f);
pathB.cubicTo(862.571045f, 430.754944f, 865.48999f, 429.237549f, 868.44397f, 427.018372f);
pathB.cubicTo(870.505371f, 425.470032f, 871.582581f, 423.534332f, 871.582581f, 421.001678f);
pathB.cubicTo(871.582581f, 417.945923f, 870.056213f, 415.171692f, 867.015381f, 412.640045f);
pathB.cubicTo(863.683105f, 409.872803f, 861.445923f, 408.027954f, 860.551514f, 407.140503f);
pathB.cubicTo(858.660767f, 405.264709f, 857.765259f, 403.50174f, 857.765259f, 402.187988f);
pathB.cubicTo(857.765259f, 401.141785f, 858.339355f, 400.394073f, 859.476318f, 399.925873f);
pathB.cubicTo(860.004395f, 399.704254f, 861.270264f, 400.515869f, 863.156006f, 402.36969f);
pathB.cubicTo(865.094727f, 404.28241f, 866.203796f, 405.565186f, 866.383484f, 406.130219f);
pathB.cubicTo(868.250244f, 404.305359f, 869.179688f, 403.397919f, 871.046509f, 401.58902f);
pathB.cubicTo(868.26825f, 399.296967f, 864.431824f, 394.705841f, 863.156006f, 394.600037f);
pathB.cubicTo(863.145996f, 394.600037f, 863.136108f, 394.59903f, 863.126099f, 394.59903f);
pathB.cubicTo(862.352417f, 394.598022f, 859.909607f, 396.79425f, 859.11792f, 397.320343f);
pathB.moveTo(832.164246f, 394.307526f);
pathB.cubicTo(832.451721f, 394.425323f, 832.598511f, 394.486206f, 832.886963f, 394.605011f);
pathB.cubicTo(834.078979f, 395.474518f, 834.674927f, 395.90979f, 835.867859f, 396.781281f);
pathB.cubicTo(836.502808f, 397.325348f, 836.863159f, 398.000183f, 836.863159f, 398.964539f);
pathB.lineTo(836.863159f, 419.740845f);
pathB.cubicTo(836.863159f, 420.876923f, 836.319092f, 422.17868f, 835.055298f, 423.617188f);
pathB.cubicTo(836.39502f, 424.512665f, 837.063843f, 424.961884f, 838.39856f, 425.864349f);
pathB.cubicTo(839.477661f, 426.578125f, 841.37439f, 427.27594f, 842.275879f, 427.443634f);
pathB.cubicTo(842.999634f, 427.574402f, 843.82019f, 427.513519f, 844.354309f, 427.216034f);
pathB.cubicTo(846.956787f, 425.765503f, 848.689819f, 423.588257f, 848.58606f, 423.483429f);
pathB.cubicTo(848.58606f, 423.483429f, 846.877991f, 423.327698f, 845.971558f, 422.807587f);
pathB.cubicTo(845.253784f, 422.284485f, 844.892395f, 422.022949f, 844.171631f, 421.502838f);
pathB.cubicTo(843.361023f, 420.915833f, 842.907837f, 420.308899f, 842.907837f, 419.350525f);
pathB.lineTo(842.907837f, 399.445709f);
pathB.cubicTo(842.907837f, 398.053101f, 843.272217f, 397.417175f, 843.812256f, 397.518005f);
pathB.cubicTo(844.170654f, 397.583893f, 844.711731f, 398.122986f, 845.432495f, 398.782837f);
pathB.cubicTo(846.116333f, 399.402771f, 846.459717f, 399.709259f, 847.14856f, 400.3302f);
pathB.cubicTo(844.986206f, 402.099152f, 843.988892f, 403.926025f, 843.988892f, 405.932556f);
pathB.cubicTo(843.988892f, 410.209229f, 848.272583f, 410.951935f, 849.576355f, 408.394348f);
pathB.cubicTo(849.871826f, 407.816345f, 850.421875f, 406.214081f, 850.387939f, 406.196106f);
pathB.cubicTo(850.387939f, 406.196106f, 849.305786f, 406.771118f, 848.495239f, 406.615387f);
pathB.cubicTo(846.96582f, 406.316895f, 846.153198f, 405.46637f, 846.153198f, 403.89505f);
pathB.cubicTo(846.153198f, 401.796661f, 848.50116f, 399.09729f, 852.279663f, 396.270142f);
pathB.cubicTo(851.014893f, 395.315796f, 847.723511f, 391.546265f, 846.875f, 391.546265f);
pathB.cubicTo(846.330933f, 391.546265f, 843.988892f, 394.403351f, 843.273193f, 394.972382f);
pathB.cubicTo(840.889282f, 392.886963f, 839.700317f, 391.850739f, 837.312378f, 389.786285f);
pathB.cubicTo(835.257935f, 391.589203f, 834.225708f, 392.491638f, 832.164246f, 394.307526f);
pathB.moveTo(818.860107f, 392.707275f);
pathB.cubicTo(819.857361f, 393.382111f, 822.302124f, 395.764038f, 824.387573f, 397.051819f);
pathB.cubicTo(822.57666f, 398.249756f, 820.582092f, 399.687286f, 818.860107f, 400.827332f);
pathB.lineTo(818.860107f, 392.707275f);
pathB.close();
pathB.moveTo(810.69812f, 391.096039f);
pathB.cubicTo(810.69812f, 391.096039f, 812.786499f, 394.093903f, 812.786499f, 394.965393f);
pathB.lineTo(812.786499f, 415.743713f);
pathB.cubicTo(812.786499f, 417.753265f, 811.881042f, 418.497986f, 810.974609f, 419.769806f);
pathB.cubicTo(813.948486f, 421.160431f, 815.437988f, 421.864197f, 818.404846f, 423.283783f);
pathB.cubicTo(819.948181f, 423.95462f, 822.417969f, 424.592529f, 823.937317f, 423.782928f);
pathB.cubicTo(827.905518f, 421.663544f, 831.53125f, 417.600525f, 832.255005f, 415.191681f);
pathB.cubicTo(833.882263f, 409.877808f, 823.095825f, 411.495026f, 823.119751f, 411.518982f);
pathB.cubicTo(823.119751f, 411.518982f, 832.000488f, 411.874359f, 830.537964f, 416.29776f);
pathB.cubicTo(829.888123f, 418.253418f, 827.278564f, 420.292908f, 825.385864f, 419.55719f);
pathB.cubicTo(821.14209f, 417.915985f, 818.861023f, 417.414856f, 818.861023f, 414.970032f);
pathB.lineTo(818.861023f, 403.096436f);
pathB.cubicTo(822.126404f, 399.132233f, 831.289673f, 395.897797f, 831.356567f, 395.657227f);
pathB.cubicTo(831.356567f, 395.657227f, 823.022888f, 387.594055f, 821.763062f, 387.476257f);
pathB.cubicTo(821.755066f, 387.47525f, 821.746094f, 387.47525f, 821.737061f, 387.47525f);
pathB.cubicTo(820.793701f, 387.47525f, 810.72406f, 390.967255f, 810.69812f, 391.096039f);
pathB.moveTo(624.254211f, 390.498077f);
pathB.cubicTo(625.252502f, 390.893402f, 627.708252f, 392.592468f, 629.796692f, 393.307251f);
pathB.cubicTo(627.978821f, 395.006317f, 625.980225f, 397.000916f, 624.254211f, 398.618134f);
pathB.lineTo(624.254211f, 390.498077f);
pathB.close();
pathB.moveTo(627.160217f, 384.460449f);
pathB.cubicTo(626.286743f, 384.51535f, 616.076233f, 390.993225f, 616.086243f, 391.141968f);
pathB.cubicTo(616.086243f, 391.141968f, 618.173645f, 393.561798f, 618.173645f, 394.437317f);
pathB.lineTo(618.173645f, 415.216614f);
pathB.cubicTo(618.173645f, 417.222168f, 617.265198f, 418.219482f, 616.355774f, 419.742859f);
pathB.cubicTo(619.331665f, 420.307892f, 620.824097f, 420.599396f, 623.802979f, 421.198364f);
pathB.cubicTo(625.346313f, 421.437958f, 627.818115f, 421.39801f, 629.342468f, 420.166138f);
pathB.cubicTo(633.340576f, 416.939667f, 636.982361f, 411.871368f, 637.714111f, 409.263855f);
pathB.cubicTo(639.348267f, 403.500732f, 628.508911f, 408.111816f, 628.52887f, 408.126801f);
pathB.cubicTo(628.52887f, 408.126801f, 637.468506f, 405.998444f, 635.985046f, 410.844147f);
pathB.cubicTo(635.332153f, 412.984467f, 632.705688f, 415.748718f, 630.801941f, 415.541077f);
pathB.cubicTo(626.537292f, 415.072876f, 624.257202f, 415.202667f, 624.257202f, 412.755859f);
pathB.cubicTo(624.257202f, 408.007019f, 624.255188f, 405.636078f, 624.255188f, 400.884247f);
pathB.cubicTo(627.525574f, 396.016602f, 636.801636f, 390.283447f, 636.801636f, 389.97995f);
pathB.cubicTo(636.801636f, 389.97995f, 628.360168f, 384.458435f, 627.18219f, 384.458435f);
pathB.cubicTo(627.174194f, 384.460449f, 627.167236f, 384.460449f, 627.160217f, 384.460449f);
pathB.moveTo(796.530396f, 416.438538f);
pathB.cubicTo(795.892517f, 416.365662f, 794.527832f, 415.589996f, 792.348572f, 414.036652f);
pathB.lineTo(792.348572f, 391.425476f);
pathB.cubicTo(792.348572f, 390.465118f, 792.530273f, 390.047852f, 792.89563f, 390.088776f);
pathB.cubicTo(793.075317f, 390.109741f, 793.3479f, 390.317383f, 793.804077f, 390.629852f);
pathB.cubicTo(795.113831f, 391.585205f, 795.768738f, 392.059387f, 797.077515f, 393.018738f);
pathB.cubicTo(797.983948f, 393.648651f, 798.348267f, 394.219666f, 798.348267f, 394.742767f);
pathB.lineTo(798.348267f, 413.253998f);
pathB.cubicTo(798.348267f, 415.391327f, 797.783264f, 416.451508f, 796.728088f, 416.451508f);
pathB.cubicTo(796.664185f, 416.4505f, 796.598267f, 416.446533f, 796.530396f, 416.438538f);
pathB.moveTo(795.165771f, 383.714722f);
pathB.cubicTo(794.022705f, 383.851471f, 783.959961f, 388.652252f, 783.880127f, 388.873871f);
pathB.cubicTo(783.880127f, 388.873871f, 785.054077f, 389.871155f, 785.522339f, 390.606873f);
pathB.cubicTo(786.000488f, 391.361603f, 786.246094f, 391.9935f, 786.246094f, 392.427765f);
pathB.lineTo(786.246094f, 411.987183f);
pathB.cubicTo(786.246094f, 413.733185f, 784.160645f, 416.428558f, 784.246521f, 416.759979f);
pathB.cubicTo(784.258484f, 416.79892f, 785.432495f, 417.14032f, 785.793823f, 417.350952f);
pathB.cubicTo(786.739258f, 417.937958f, 787.213379f, 418.228455f, 788.161804f, 418.821442f);
pathB.cubicTo(789.342773f, 419.554199f, 790.619568f, 419.956482f, 791.892395f, 420.098236f);
pathB.cubicTo(794.533813f, 420.390747f, 796.717102f, 419.337555f, 798.349304f, 416.999573f);
pathB.lineTo(798.349304f, 425.212463f);
pathB.cubicTo(797.94696f, 425.47702f, 797.750305f, 425.609772f, 797.356018f, 425.874329f);
pathB.cubicTo(795.259583f, 423.619202f, 792.806824f, 422.286499f, 789.985657f, 421.984009f);
pathB.cubicTo(785.157959f, 421.463898f, 780.409119f, 428.344086f, 780.423096f, 428.346069f);
pathB.cubicTo(780.423096f, 428.346069f, 783.340088f, 424.960907f, 785.889709f, 425.218445f);
pathB.cubicTo(789.25592f, 425.565857f, 793.166199f, 430.745972f, 793.805115f, 430.790894f);
pathB.cubicTo(793.940857f, 430.798889f, 795.918457f, 429.091797f, 798.454102f, 427.383728f);
pathB.cubicTo(801.049683f, 425.635742f, 804.230225f, 423.886749f, 806.619141f, 423.980591f);
pathB.cubicTo(805.621826f, 423.586243f, 805.048828f, 423.074127f, 804.804199f, 422.609924f);
pathB.cubicTo(804.616577f, 422.25354f, 804.616577f, 421.539764f, 804.616577f, 420.31488f);
pathB.cubicTo(804.623535f, 411.732605f, 804.623535f, 403.147339f, 804.623535f, 394.562073f);
pathB.cubicTo(804.623535f, 392.464691f, 805.970215f, 391.000183f, 805.984192f, 390.896362f);
pathB.cubicTo(805.984192f, 390.896362f, 796.785034f, 383.7117f, 795.219666f, 383.7117f);
pathB.cubicTo(795.19873f, 383.712708f, 795.181763f, 383.712708f, 795.165771f, 383.714722f);
pathB.moveTo(648.092285f, 387.883545f);
pathB.cubicTo(649.095581f, 388.312805f, 651.55835f, 390.099762f, 653.655701f, 390.884399f);
pathB.cubicTo(651.831848f, 392.522583f, 649.82428f, 394.447296f, 648.092285f, 396.003601f);
pathB.lineTo(648.092285f, 387.883545f);
pathB.close();
pathB.moveTo(651.009277f, 381.943756f);
pathB.cubicTo(650.147766f, 381.983704f, 639.893372f, 388.105164f, 639.899353f, 388.254913f);
pathB.cubicTo(639.899353f, 388.254913f, 641.987793f, 390.744659f, 641.987793f, 391.617157f);
pathB.lineTo(641.987793f, 412.399475f);
pathB.cubicTo(641.987793f, 414.409027f, 641.082336f, 415.369354f, 640.169861f, 416.864807f);
pathB.cubicTo(643.155762f, 417.53064f, 644.650208f, 417.87207f, 647.638062f, 418.573853f);
pathB.cubicTo(649.188416f, 418.865356f, 651.666138f, 418.908295f, 653.19751f, 417.725311f);
pathB.cubicTo(657.204651f, 414.633636f, 660.859375f, 409.690125f, 661.590088f, 407.106567f);
pathB.cubicTo(663.231262f, 401.397339f, 652.356934f, 405.644073f, 652.375916f, 405.663025f);
pathB.cubicTo(652.375916f, 405.663025f, 661.338562f, 403.835175f, 659.857056f, 408.632935f);
pathB.cubicTo(659.199219f, 410.748291f, 656.568726f, 413.424713f, 654.656982f, 413.151184f);
pathB.cubicTo(650.381348f, 412.536224f, 648.092285f, 412.591125f, 648.092285f, 410.146332f);
pathB.lineTo(648.092285f, 398.270721f);
pathB.cubicTo(651.374634f, 393.5159f, 660.66571f, 388.09021f, 660.674683f, 387.791718f);
pathB.cubicTo(660.674683f, 387.791718f, 652.188232f, 381.941772f, 651.022278f, 381.942749f);
pathB.cubicTo(651.01825f, 381.942749f, 651.013245f, 381.942749f, 651.009277f, 381.943756f);
pathB.moveTo(761.636353f, 385.965851f);
pathB.cubicTo(761.927856f, 386.056702f, 762.071594f, 386.098633f, 762.363098f, 386.189453f);
pathB.cubicTo(763.570007f, 386.938171f, 764.175964f, 387.311554f, 765.376892f, 388.066254f);
pathB.cubicTo(766.019775f, 388.546417f, 766.384155f, 389.184326f, 766.384155f, 390.147675f);
pathB.lineTo(766.384155f, 410.924011f);
pathB.cubicTo(766.384155f, 412.057037f, 765.836121f, 413.410736f, 764.559326f, 414.979034f);
pathB.cubicTo(765.911987f, 415.738739f, 766.579834f, 416.12207f, 767.934509f, 416.887756f);
pathB.cubicTo(769.029602f, 417.495728f, 770.944336f, 418.000854f, 771.85675f, 418.075714f);
pathB.cubicTo(772.58551f, 418.134613f, 773.413086f, 417.987854f, 773.950195f, 417.638458f);
pathB.cubicTo(776.583618f, 415.917419f, 778.332642f, 413.564453f, 778.237793f, 413.473633f);
pathB.cubicTo(778.237793f, 413.473633f, 776.507812f, 413.497559f, 775.596313f, 413.066315f);
pathB.cubicTo(774.866577f, 412.61908f, 774.497253f, 412.39447f, 773.771484f, 411.951233f);
pathB.cubicTo(772.947876f, 411.444092f, 772.493652f, 410.877075f, 772.493652f, 409.919708f);
pathB.lineTo(772.493652f, 390.013885f);
pathB.cubicTo(772.493652f, 388.618286f, 772.860046f, 387.949432f, 773.407104f, 387.995361f);
pathB.cubicTo(773.771484f, 388.026306f, 774.318542f, 388.509491f, 775.049316f, 389.09848f);
pathB.cubicTo(775.742065f, 389.646515f, 776.088501f, 389.923065f, 776.77533f, 390.470123f);
pathB.cubicTo(774.590088f, 392.45871f, 773.589783f, 394.385376f, 773.589783f, 396.395935f);
pathB.cubicTo(773.589783f, 400.673584f, 777.907349f, 401.008026f, 779.237122f, 398.292694f);
pathB.cubicTo(779.539551f, 397.684723f, 780.089661f, 396.027557f, 780.058716f, 396.01358f);
pathB.cubicTo(780.058716f, 396.01358f, 778.970581f, 396.694427f, 778.149963f, 396.618561f);
pathB.cubicTo(776.598633f, 396.4758f, 775.775024f, 395.709106f, 775.775024f, 394.13681f);
pathB.cubicTo(775.775024f, 392.042419f, 778.149963f, 389.103455f, 781.973389f, 385.892975f);
pathB.cubicTo(780.697571f, 385.06839f, 777.326416f, 381.676208f, 776.506775f, 381.719147f);
pathB.cubicTo(775.908813f, 381.747101f, 773.588806f, 384.868744f, 772.860046f, 385.506622f);
pathB.cubicTo(770.451172f, 383.664795f, 769.248291f, 382.749359f, 766.843384f, 380.929504f);
pathB.cubicTo(764.758972f, 382.934052f, 763.716736f, 383.940338f, 761.636353f, 385.965851f);
pathB.moveTo(672.996521f, 379.821411f);
pathB.cubicTo(672.123047f, 379.891266f, 669.7052f, 382.898132f, 668.887573f, 383.64682f);
pathB.cubicTo(665.239868f, 386.999084f, 663.41095f, 390.213562f, 663.41095f, 393.356171f);
pathB.cubicTo(663.41095f, 395.715118f, 664.439209f, 397.642792f, 666.785156f, 399.150208f);
pathB.cubicTo(669.702148f, 401.02002f, 672.547302f, 402.439575f, 674.545837f, 403.655487f);
pathB.cubicTo(676.261902f, 404.697693f, 677.105469f, 406.231049f, 677.105469f, 407.625671f);
pathB.cubicTo(677.105469f, 408.671875f, 676.651245f, 409.777954f, 675.825684f, 410.7453f);
pathB.cubicTo(675.12384f, 411.569885f, 674.538879f, 412.145905f, 673.997803f, 412.417419f);
pathB.cubicTo(673.38385f, 412.724915f, 672.080078f, 411.958221f, 670.166382f, 410.198242f);
pathB.cubicTo(668.113892f, 408.319458f, 667.062683f, 406.55249f, 667.062683f, 404.808502f);
pathB.cubicTo(667.062683f, 404.020844f, 667.701599f, 402.580322f, 667.701599f, 402.580322f);
pathB.cubicTo(667.701599f, 402.580322f, 661.773804f, 409.542358f, 661.951477f, 410.7453f);
pathB.cubicTo(662.13916f, 412.037079f, 663.368042f, 413.524536f, 665.60321f, 415.469208f);
pathB.cubicTo(667.791443f, 417.368927f, 669.261963f, 418.074738f, 669.983704f, 417.630493f);
pathB.cubicTo(672.412537f, 416.138062f, 675.369446f, 413.822021f, 678.385254f, 410.790222f);
pathB.cubicTo(680.485657f, 408.677856f, 681.587769f, 406.446686f, 681.587769f, 403.917023f);
pathB.cubicTo(681.587769f, 400.859283f, 680.007446f, 398.490356f, 676.923767f, 396.806244f);
pathB.cubicTo(673.540588f, 394.957428f, 671.257507f, 393.71756f, 670.351074f, 393.075653f);
pathB.cubicTo(668.434326f, 391.71698f, 667.518921f, 390.193604f, 667.518921f, 388.88385f);
pathB.cubicTo(667.518921f, 387.837646f, 668.101929f, 386.934204f, 669.25592f, 386.156525f);
pathB.cubicTo(669.796997f, 385.788147f, 671.085815f, 386.257355f, 672.997498f, 387.592072f);
pathB.cubicTo(674.966125f, 388.968689f, 676.104187f, 389.951019f, 676.284851f, 390.465118f);
pathB.cubicTo(678.186584f, 388.130127f, 679.136963f, 386.966125f, 681.035706f, 384.646118f);
pathB.cubicTo(678.244507f, 383.133728f, 674.247375f, 379.819397f, 673.044434f, 379.819397f);
pathB.cubicTo(673.027466f, 379.819397f, 673.011475f, 379.820404f, 672.996521f, 379.821411f);
pathB.moveTo(732.95459f, 384.60318f);
pathB.cubicTo(733.246094f, 384.680054f, 733.391846f, 384.720001f, 733.689331f, 384.794861f);
pathB.cubicTo(735.072937f, 385.500641f, 735.769714f, 385.856049f, 737.162354f, 386.563812f);
pathB.cubicTo(737.891113f, 386.938171f, 738.164612f, 387.642975f, 738.164612f, 388.6073f);
pathB.lineTo(738.164612f, 408.510132f);
pathB.cubicTo(738.164612f, 410.257141f, 737.709412f, 411.893341f, 736.064209f, 413.416718f);
pathB.cubicTo(737.635498f, 414.235321f, 738.419189f, 414.651611f, 739.991455f, 415.475189f);
pathB.cubicTo(740.997742f, 416.034241f, 742.186707f, 416.344696f, 743.098145f, 416.379639f);
pathB.cubicTo(743.830872f, 416.410583f, 744.476807f, 416.175964f, 745.019836f, 415.851532f);
pathB.cubicTo(746.476318f, 414.977051f, 748.58075f, 413.571442f, 749.225647f, 413.079285f);
pathB.cubicTo(751.012573f, 414.253296f, 751.907043f, 414.845276f, 753.69696f, 416.028229f);
pathB.cubicTo(754.703247f, 416.610229f, 755.706543f, 416.84082f, 756.528076f, 416.892761f);
pathB.cubicTo(757.259827f, 416.93866f, 757.996582f, 416.807892f, 758.537659f, 416.494446f);
pathB.cubicTo(760.814758f, 415.174713f, 762.185425f, 413.509552f, 762.552734f, 412.830719f);
pathB.cubicTo(761.637329f, 412.681976f, 759.633789f, 411.58786f, 759.263428f, 411.387207f);
pathB.cubicTo(758.607544f, 410.994873f, 758.279114f, 410.803223f, 757.621216f, 410.413879f);
pathB.cubicTo(756.983276f, 410.020538f, 756.616943f, 409.301788f, 756.616943f, 408.343445f);
pathB.lineTo(756.616943f, 388.351746f);
pathB.cubicTo(756.616943f, 387.387421f, 757.164978f, 386.548859f, 758.627502f, 385.067383f);
pathB.cubicTo(755.523804f, 383.05484f, 753.97052f, 382.057556f, 750.862854f, 380.078949f);
pathB.cubicTo(749.001038f, 382.112457f, 748.069641f, 383.130707f, 746.207825f, 385.174194f);
pathB.cubicTo(746.501343f, 385.292999f, 746.647095f, 385.353912f, 746.939575f, 385.472687f);
pathB.cubicTo(747.996765f, 386.183472f, 748.525879f, 386.538879f, 749.587036f, 387.257629f);
pathB.cubicTo(750.224915f, 387.724823f, 750.498474f, 388.351746f, 750.498474f, 389.223267f);
pathB.lineTo(750.498474f, 407.822327f);
pathB.cubicTo(750.498474f, 408.694824f, 750.339722f, 409.955658f, 749.951416f, 410.847137f);
pathB.cubicTo(749.550049f, 411.761566f, 749.039978f, 411.585876f, 748.487915f, 411.560913f);
pathB.cubicTo(747.393799f, 411.503998f, 746.385498f, 410.53067f, 745.473083f, 410.022552f);
pathB.cubicTo(744.760254f, 409.627228f, 744.380981f, 409.013275f, 744.380981f, 407.965088f);
pathB.lineTo(744.380981f, 386.840363f);
pathB.cubicTo(744.380981f, 385.791138f, 744.833191f, 384.763916f, 745.657776f, 383.839508f);
pathB.cubicTo(742.656921f, 382.101501f, 741.161499f, 381.234985f, 738.162659f, 379.525909f);
pathB.cubicTo(736.083191f, 381.548431f, 735.039978f, 382.562683f, 732.95459f, 384.60318f);
pathB.moveTo(692.546936f, 385.171204f);
pathB.cubicTo(693.552246f, 385.667358f, 696.018005f, 387.607025f, 698.122375f, 388.521454f);
pathB.cubicTo(696.293518f, 390.043854f, 694.281982f, 391.844757f, 692.546936f, 393.294281f);
pathB.lineTo(692.546936f, 385.171204f);
pathB.close();
pathB.moveTo(695.4729f, 379.417084f);
pathB.cubicTo(694.635376f, 379.426086f, 684.32605f, 384.880707f, 684.322083f, 385.025452f);
pathB.cubicTo(684.322083f, 385.025452f, 686.422485f, 387.645966f, 686.422485f, 388.521454f);
pathB.lineTo(686.422485f, 409.300781f);
pathB.cubicTo(686.422485f, 411.312347f, 685.51001f, 412.21579f, 684.595581f, 413.65033f);
pathB.cubicTo(687.592468f, 414.504852f, 689.089905f, 414.945099f, 692.088745f, 415.833557f);
pathB.cubicTo(693.645081f, 416.221893f, 696.128784f, 416.420563f, 697.667114f, 415.334412f);
pathB.cubicTo(701.67926f, 412.494293f, 705.344971f, 407.783386f, 706.077698f, 405.240753f);
pathB.cubicTo(707.721924f, 399.638367f, 696.822632f, 403.198273f, 696.845581f, 403.216248f);
pathB.cubicTo(696.845581f, 403.216248f, 705.825134f, 401.960388f, 704.337708f, 406.658325f);
pathB.cubicTo(703.683838f, 408.733765f, 701.044373f, 411.241455f, 699.129639f, 410.847137f);
pathB.cubicTo(694.843018f, 409.968628f, 692.545959f, 409.876801f, 692.545959f, 407.432983f);
pathB.lineTo(692.545959f, 395.563354f);
pathB.cubicTo(695.838318f, 391.012177f, 705.134338f, 386.160522f, 705.162292f, 385.873993f);
pathB.cubicTo(705.162292f, 385.873993f, 696.635925f, 379.416107f, 695.473938f, 379.417084f);
pathB.cubicTo(695.474915f, 379.417084f, 695.473938f, 379.417084f, 695.4729f, 379.417084f);
pathB.moveTo(570.463562f, 420.81601f);
pathB.lineTo(570.463562f, 402.922729f);
pathB.cubicTo(571.039551f, 402.800934f, 571.327087f, 402.743042f, 571.901123f, 402.625244f);
pathB.lineTo(571.901123f, 423.142029f);
pathB.cubicTo(570.911804f, 422.823578f, 570.463562f, 422.123779f, 570.463562f, 420.81601f);
pathB.moveTo(570.463562f, 384.062134f);
pathB.cubicTo(571.039551f, 384.149963f, 571.327087f, 384.198883f, 571.901123f, 384.290741f);
pathB.lineTo(571.901123f, 401.580048f);
pathB.cubicTo(571.327087f, 401.695862f, 571.039551f, 401.756744f, 570.463562f, 401.874542f);
pathB.lineTo(570.463562f, 384.062134f);
pathB.close();
pathB.moveTo(573.880676f, 376.556f);
pathB.cubicTo(572.483093f, 376.996246f, 561.476013f, 385.624451f, 561.482971f, 385.70929f);
pathB.cubicTo(561.482971f, 385.70929f, 563.637268f, 388.554413f, 563.637268f, 389.688446f);
pathB.lineTo(563.637268f, 398.423462f);
pathB.cubicTo(556.411682f, 399.838043f, 555.429382f, 404.307373f, 555.418396f, 405.679993f);
pathB.lineTo(555.418396f, 405.724915f);
pathB.cubicTo(555.42041f, 405.94455f, 555.448364f, 406.073334f, 555.477295f, 406.083313f);
pathB.cubicTo(555.477295f, 406.083313f, 558.070862f, 404.250458f, 563.637268f, 403.222229f);
pathB.lineTo(563.637268f, 404.797516f);
pathB.cubicTo(556.993713f, 406.233063f, 555.191772f, 412.494293f, 555.569153f, 412.614105f);
pathB.cubicTo(555.569153f, 412.614105f, 561.572815f, 410.21521f, 563.637268f, 409.598267f);
pathB.lineTo(563.637268f, 424.00354f);
pathB.cubicTo(563.637268f, 426.357483f, 563.36676f, 427.901855f, 562.291565f, 429.70874f);
pathB.cubicTo(565.448181f, 430.067139f, 567.028442f, 430.256805f, 570.192017f, 430.653137f);
pathB.cubicTo(571.99292f, 430.893707f, 574.782166f, 430.669098f, 576.403381f, 429.136719f);
pathB.cubicTo(580.960571f, 424.828125f, 586.135681f, 419.346527f, 586.135681f, 416.115082f);
pathB.lineTo(586.135681f, 406.511566f);
pathB.cubicTo(586.135681f, 405.377533f, 586.047791f, 404.608856f, 586.678711f, 403.271149f);
pathB.cubicTo(584.151062f, 404.98819f, 582.888245f, 405.851715f, 580.362549f, 407.587738f);
pathB.cubicTo(579.281433f, 408.320465f, 579.192566f, 409.2948f, 579.192566f, 410.955933f);
pathB.lineTo(579.192566f, 421.869202f);
pathB.cubicTo(579.192566f, 423.180969f, 577.746033f, 423.273804f, 577.392639f, 423.266815f);
pathB.cubicTo(575.636658f, 423.228882f, 574.153259f, 423.295776f, 573.071106f, 423.077148f);
pathB.lineTo(573.071106f, 384.663086f);
pathB.cubicTo(575.230408f, 385.379852f, 576.309509f, 385.742249f, 578.473816f, 386.473999f);
pathB.cubicTo(579.373291f, 386.996094f, 579.553955f, 387.490234f, 579.553955f, 388.013336f);
pathB.cubicTo(581.861023f, 384.848785f, 583.015991f, 383.267487f, 585.325073f, 380.114899f);
pathB.cubicTo(581.680298f, 379.229431f, 575.865295f, 376.520081f, 574.157227f, 376.521057f);
pathB.cubicTo(574.047424f, 376.522064f, 573.955566f, 376.533051f, 573.880676f, 376.556f);
pathB.moveTo(593.447083f, 375.096527f);
pathB.cubicTo(592.363953f, 375.804291f, 591.821899f, 376.772644f, 591.821899f, 377.908691f);
pathB.lineTo(591.821899f, 419.46933f);
pathB.cubicTo(591.821899f, 420.517517f, 591.187012f, 422.018951f, 589.921143f, 423.991577f);
pathB.cubicTo(591.2948f, 424.412842f, 591.982605f, 424.622467f, 593.354248f, 425.050751f);
pathB.cubicTo(594.53125f, 425.462036f, 595.525513f, 425.555878f, 596.427979f, 425.404144f);
pathB.cubicTo(597.150757f, 425.279358f, 597.785645f, 424.914978f, 598.326721f, 424.475739f);
pathB.cubicTo(600.935242f, 422.385315f, 602.846985f, 419.809753f, 602.759094f, 419.749847f);
pathB.cubicTo(602.759094f, 419.749847f, 601.582153f, 419.935516f, 600.59082f, 419.831696f);
pathB.cubicTo(600.0448f, 419.74585f, 599.774231f, 419.700928f, 599.233154f, 419.615082f);
pathB.cubicTo(598.416565f, 419.484314f, 597.965332f, 418.860382f, 597.965332f, 417.988861f);
pathB.lineTo(597.965332f, 396.857147f);
pathB.cubicTo(597.965332f, 395.376678f, 598.326721f, 394.617004f, 598.867798f, 394.528137f);
pathB.cubicTo(599.232178f, 394.466248f, 599.773254f, 394.731812f, 600.59082f, 395.124115f);
pathB.cubicTo(601.601074f, 395.589325f, 602.111206f, 395.819946f, 603.123474f, 396.288116f);
pathB.cubicTo(603.93811f, 396.686432f, 603.93512f, 397.38324f, 603.93512f, 398.169891f);
pathB.cubicTo(603.93512f, 405.971497f, 603.93512f, 413.768127f, 603.93811f, 421.569702f);
pathB.cubicTo(603.93811f, 425.325256f, 601.109924f, 430.634155f, 601.133911f, 430.656128f);
pathB.cubicTo(601.133911f, 430.656128f, 605.184937f, 427.222015f, 607.017822f, 424.414825f);
pathB.cubicTo(609.118164f, 421.201355f, 610.280212f, 417.987854f, 610.280212f, 415.109802f);
pathB.lineTo(610.280212f, 394.593048f);
pathB.cubicTo(610.280212f, 393.890228f, 610.823242f, 393.112579f, 611.728699f, 392.020447f);
pathB.cubicTo(608.827698f, 390.960266f, 604.000977f, 387.703857f, 602.759094f, 387.967407f);
pathB.cubicTo(602.120239f, 388.104187f, 599.957947f, 391.29071f, 597.965332f, 393.27829f);
pathB.lineTo(597.965332f, 374.422668f);
pathB.cubicTo(597.965332f, 373.461334f, 598.326721f, 372.440063f, 598.867798f, 371.567566f);
pathB.cubicTo(596.701538f, 372.96817f, 595.616394f, 373.677948f, 593.447083f, 375.096527f);
pathB.moveTo(718.054138f, 409.318756f);
pathB.cubicTo(717.461182f, 408.789673f, 716.867188f, 408.178711f, 716.867188f, 407.218353f);
pathB.lineTo(716.867188f, 387.053986f);
pathB.cubicTo(716.867188f, 385.305969f, 717.323425f, 385.566528f, 718.328674f, 386.013763f);
pathB.cubicTo(719.645386f, 386.859314f, 720.307251f, 387.284576f, 721.622009f, 388.135132f);
pathB.cubicTo(722.266907f, 388.4935f, 722.903809f, 388.934753f, 722.903809f, 389.721405f);
pathB.lineTo(722.903809f, 407.794373f);
pathB.cubicTo(722.903809f, 408.66687f, 722.746094f, 410.490753f, 722.259888f, 410.758301f);
pathB.cubicTo(722.125122f, 410.83017f, 721.950439f, 410.862122f, 721.746826f, 410.862122f);
pathB.cubicTo(720.655701f, 410.864105f, 718.747925f, 409.936707f, 718.054138f, 409.318756f);
pathB.moveTo(711.928711f, 364.782227f);
pathB.cubicTo(711.195923f, 365.134613f, 710.648865f, 365.834412f, 710.648865f, 366.794769f);
pathB.lineTo(710.648865f, 407.392059f);
pathB.cubicTo(710.648865f, 409.397614f, 708.519531f, 411.37323f, 708.547485f, 411.684692f);
pathB.cubicTo(708.550476f, 411.745605f, 711.838867f, 413.067322f, 713.849365f, 414.368073f);
pathB.cubicTo(717.766663f, 416.906738f, 720.162537f, 415.845551f, 722.354797f, 414.073608f);
pathB.cubicTo(724.059875f, 412.69397f, 726.55957f, 410.981903f, 730.675537f, 410.124359f);
pathB.cubicTo(729.75708f, 409.143066f, 729.213013f, 407.993042f, 729.213013f, 406.683289f);
pathB.cubicTo(729.213013f, 399.630402f, 729.209045f, 396.103455f, 729.209045f, 389.047546f);
pathB.cubicTo(729.209045f, 387.648956f, 730.577698f, 385.292023f, 730.583679f, 385.149261f);
pathB.cubicTo(730.583679f, 385.149261f, 720.888306f, 378.762207f, 719.609497f, 378.947906f);
pathB.cubicTo(719.275085f, 378.996826f, 717.872498f, 381.118164f, 716.868225f, 381.896851f);
pathB.lineTo(716.868225f, 365.046783f);
pathB.cubicTo(716.868225f, 363.740021f, 716.960083f, 363.043213f, 717.597961f, 362);
pathB.cubicTo(715.331848f, 363.104095f, 714.19873f, 363.657166f, 711.928711f, 364.782227f);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpeverytechpro_blogspot_com100(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(1074.29285f, 627.292786f);
path.quadTo(1074.58582f, 627, 1075, 627);
path.lineTo(1117, 627);
path.quadTo(1124.04163f, 627, 1129.02246f, 631.9776f);
path.quadTo(1134, 636.958374f, 1134, 644);
path.lineTo(1134, 645);
path.quadTo(1134, 652.041626f, 1129.02246f, 657.0224f);
path.quadTo(1124.04163f, 662, 1117, 662);
path.lineTo(1075, 662);
path.quadTo(1074.58582f, 662, 1074.29285f, 661.707214f);
path.quadTo(1074, 661.414185f, 1074, 661);
path.lineTo(1074, 628);
path.quadTo(1074, 627.585815f, 1074.29285f, 627.292786f);
path.close();
path.moveTo(1076, 629);
path.lineTo(1117, 629);
path.cubicTo(1125.2843f, 629, 1132, 635.715698f, 1132, 644);
path.lineTo(1132, 645);
path.cubicTo(1132, 653.284302f, 1125.2843f, 660, 1117, 660);
path.lineTo(1076, 660);
path.lineTo(1076, 629);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(1074, 627);
pathB.lineTo(1075, 628);
pathB.lineTo(1116.5f, 644.5f);
pathB.lineTo(1134, 627);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpflite_com41(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(301.464081f, 424);
path.lineTo(296, 433.46405f);
path.lineTo(296, 433.810822f);
path.lineTo(303.25589f, 438);
path.lineTo(304.729736f, 438);
path.lineTo(311, 427.139557f);
path.lineTo(311, 426.305237f);
path.lineTo(307.007202f, 424);
path.lineTo(301.464081f, 424);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(302.849854f, 421.599762f);
pathB.lineTo(311.510101f, 426.599762f);
pathB.lineTo(304.510101f, 438.724121f);
pathB.lineTo(295.849854f, 433.724121f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpilkoora_com37(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(818, 157);
path.cubicTo(818, 148.715729f, 824.715698f, 142, 833, 142);
path.lineTo(909, 142);
path.lineTo(909, 143);
path.lineTo(833, 143);
path.cubicTo(825.268005f, 143, 819, 149.268005f, 819, 157);
path.lineTo(819, 926);
path.lineTo(818, 926);
path.lineTo(818, 157);
path.close();
path.moveTo(1184, 926);
path.lineTo(1185, 926);
path.lineTo(1185, 157);
path.cubicTo(1185, 148.715729f, 1178.2843f, 142, 1170, 142);
path.lineTo(1093, 142);
path.lineTo(1093, 143);
path.lineTo(1170, 143);
path.cubicTo(1177.73193f, 143, 1184, 149.268005f, 1184, 157);
path.lineTo(1184, 926);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(1185, 142);
pathB.lineTo(1001.5f, 325.5f);
pathB.lineTo(1001.5f, 782.5f);
pathB.lineTo(1185, 966);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpmm4everfriends_com43(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(540.74231f, 215.922546f);
path.cubicTo(540.893127f, 215.391159f, 541.443909f, 215.090134f, 541.972473f, 215.250168f);
path.lineTo(581.213318f, 227.131104f);
path.cubicTo(581.741882f, 227.291153f, 582.048157f, 227.851654f, 581.897339f, 228.383041f);
path.lineTo(576.708923f, 246.663925f);
path.cubicTo(576.558167f, 247.195297f, 576.007324f, 247.496338f, 575.47876f, 247.336288f);
path.lineTo(536.237915f, 235.455353f);
path.cubicTo(535.709351f, 235.295319f, 535.403137f, 234.734802f, 535.553894f, 234.20343f);
path.lineTo(540.74231f, 215.922546f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(541.015381f, 214.960388f);
pathB.lineTo(582.17041f, 227.420883f);
pathB.lineTo(576.435852f, 247.626068f);
pathB.lineTo(535.280823f, 235.165573f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpmtrk_uz27(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(33, 787);
path.lineTo(33, 412);
path.lineTo(1233, 412);
path.lineTo(1233, 787);
path.quadTo(1233, 793.213196f, 1228.60803f, 797.607971f);
path.quadTo(1224.21326f, 802, 1218, 802);
path.lineTo(48, 802);
path.quadTo(41.7867966f, 802, 37.3919983f, 797.607971f);
path.quadTo(33, 793.213196f, 33, 787);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(33, 412);
pathB.lineTo(1233, 412);
pathB.lineTo(1233, 787);
pathB.quadTo(1233, 793.213196f, 1228.60791f, 797.608032f);
pathB.quadTo(1224.21313f, 802, 1218, 802);
pathB.lineTo(48, 802);
pathB.quadTo(41.7867432f, 802, 37.3919678f, 797.608032f);
pathB.quadTo(33, 793.213196f, 33, 787);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpfrauen_magazin_com83(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(808, 886);
path.cubicTo(805.581055f, 886, 803.563293f, 887.717773f, 803.100037f, 890);
path.lineTo(1122.90002f, 890);
path.cubicTo(1122.43677f, 887.717773f, 1120.41895f, 886, 1118, 886);
path.lineTo(808, 886);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(808, 886);
pathB.lineTo(1118, 886);
pathB.cubicTo(1120.76147f, 886, 1123, 888.238586f, 1123, 891);
pathB.lineTo(1123, 1521);
pathB.cubicTo(1123, 1523.20911f, 1120.76147f, 1525, 1118, 1525);
pathB.lineTo(808, 1525);
pathB.cubicTo(805.238586f, 1525, 803, 1523.20911f, 803, 1521);
pathB.lineTo(803, 891);
pathB.cubicTo(803, 888.238586f, 805.238586f, 886, 808, 886);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpi_gino_com16(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(184, 734);
path.quadTo(133.051727f, 734, 97.0258636f, 770.025879f);
path.quadTo(61, 806.051758f, 61, 857);
path.quadTo(61, 895.835083f, 81.9317017f, 926);
path.lineTo(286.068298f, 926);
path.quadTo(307, 895.835083f, 307, 857);
path.quadTo(307, 806.051758f, 270.974121f, 770.025879f);
path.quadTo(234.948273f, 734, 184, 734);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(185, 734);
pathB.cubicTo(252.93103f, 734, 308, 789.06897f, 308, 857);
pathB.cubicTo(308, 924.93103f, 252.93103f, 980, 185, 980);
pathB.lineTo(184, 980);
pathB.cubicTo(116.068977f, 980, 61, 924.93103f, 61, 857);
pathB.cubicTo(61, 789.06897f, 116.068977f, 734, 184, 734);
pathB.lineTo(185, 734);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skppchappy_com_au102(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(363, 493);
path.cubicTo(360.790863f, 493, 359, 494.790863f, 359, 497);
path.lineTo(359, 656);
path.cubicTo(359, 658.209106f, 360.790863f, 660, 363, 660);
path.lineTo(623.001709f, 660);
path.cubicTo(624.657776f, 659.999023f, 626, 658.65625f, 626, 657);
path.lineTo(626, 496);
path.cubicTo(626, 494.343872f, 624.657959f, 493.00116f, 623.002075f, 493);
path.lineTo(363, 493);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(362, 494);
pathB.lineTo(623, 494);
pathB.cubicTo(624.65686f, 494, 626, 494.895416f, 626, 496);
pathB.lineTo(626, 657);
pathB.cubicTo(626, 658.65686f, 624.65686f, 660, 623, 660);
pathB.lineTo(362, 660);
pathB.cubicTo(360.34314f, 660, 359, 658.65686f, 359, 657);
pathB.lineTo(359, 496);
pathB.cubicTo(359, 494.895416f, 360.34314f, 494, 362, 494);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpsciality_com161(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(656, 728);
path.cubicTo(653.790833f, 728, 652, 729.790833f, 652, 732);
path.lineTo(652, 789);
path.cubicTo(652, 791.209106f, 653.790833f, 793, 656, 793);
path.lineTo(769.001282f, 793);
path.cubicTo(770.657532f, 792.999268f, 772, 791.656433f, 772, 790);
path.lineTo(772, 731);
path.cubicTo(772, 729.34314f, 770.65686f, 728, 769, 728);
path.lineTo(656, 728);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(655, 729);
pathB.lineTo(769, 729);
pathB.cubicTo(770.65686f, 729, 772, 729.895447f, 772, 731);
pathB.lineTo(772, 790);
pathB.cubicTo(772, 791.65686f, 770.65686f, 793, 769, 793);
pathB.lineTo(655, 793);
pathB.cubicTo(653.34314f, 793, 652, 791.65686f, 652, 790);
pathB.lineTo(652, 731);
pathB.cubicTo(652, 729.895447f, 653.34314f, 729, 655, 729);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpsudoestenegocios_com186(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(0, 495);
path.lineTo(1.23685242e-14f, 293);
path.lineTo(44, 293);
path.quadTo(45.6568527f, 293, 46.8288002f, 294.171204f);
path.quadTo(48, 295.34314f, 48, 297);
path.lineTo(48, 491);
path.quadTo(48, 492.65686f, 46.8288002f, 493.828796f);
path.quadTo(45.6568527f, 495, 44, 495);
path.lineTo(0, 495);
path.close();
path.moveTo(1, 294);
path.lineTo(44, 294);
path.cubicTo(45.6568565f, 294, 47, 295.34314f, 47, 297);
path.lineTo(47, 491);
path.cubicTo(47, 492.65686f, 45.6568565f, 494, 44, 494);
path.lineTo(1, 494);
path.lineTo(1, 294);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(48, 495);
pathB.lineTo(24, 471);
pathB.lineTo(24, 317);
pathB.lineTo(48, 293);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpthesuburbanite_com213(skiatest::Reporter* reporter, const char* filename) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
path.moveTo(863.439026f, 692);
path.lineTo(863.283264f, 692);
path.lineTo(802, 708.420837f);
path.lineTo(802, 718.773621f);
path.lineTo(866, 701.624817f);
path.lineTo(866, 701.557922f);
path.lineTo(863.439026f, 692);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
pathB.moveTo(783.256775f, 713.443054f);
pathB.lineTo(863.428589f, 691.96106f);
pathB.lineTo(866.016724f, 701.620361f);
pathB.lineTo(785.84491f, 723.102356f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skphostloco_com11(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(6.66133815e-16f, 648);
path.lineTo(25.8522835f, 648);
path.quadTo(27.5087376f, 647.999634f, 28.6807098f, 646.82843f);
path.quadTo(29.8518829f, 645.656433f, 29.8522835f, 644);
path.lineTo(29.8522835f, 467);
path.quadTo(29.8518829f, 465.343536f, 28.6807098f, 464.17157f);
path.quadTo(27.5087376f, 463.000397f, 25.8522835f, 463);
path.lineTo(2.22044605e-16f, 463);
path.lineTo(6.66133815e-16f, 648);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(0, 463);
pathB.lineTo(30, 463);
pathB.lineTo(30, 648);
pathB.lineTo(0, 648);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpsergeychunkevich_com8(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(0, 926);
path.lineTo(0, 0);
path.lineTo(1265, 0);
path.lineTo(1265, 926);
path.lineTo(0, 926);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(37, 374);
pathB.lineTo(37, 535);
pathB.cubicTo(37, 536.65686f, 35.6568565f, 538, 34, 538);
pathB.lineTo(1.02866934e-14f, 538);
pathB.lineTo(6.12303177e-17f, 371);
pathB.lineTo(34, 371);
pathB.cubicTo(35.6568565f, 371, 37, 372.34314f, 37, 374);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skptracksflow_com9(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(16, 56);
path.lineTo(32, 56);
path.lineTo(32, 72);
path.lineTo(16, 72);
path.lineTo(16, 56);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(31.65625f, 70.0555649f);
pathB.lineTo(31.65625f, 70.0554962f);
pathB.lineTo(26.9727192f, 65.3615341f);
pathB.cubicTo(27.6210003f, 64.4029694f, 28.0048752f, 63.2470932f, 28.0048752f, 62.0027809f);
pathB.cubicTo(28.0048752f, 58.6875305f, 25.3199062f, 56, 22.0046558f, 56);
pathB.cubicTo(18.6894073f, 56, 16.0031872f, 58.6875305f, 16.0031872f, 62.0027809f);
pathB.cubicTo(16.0031872f, 65.3180008f, 18.6913433f, 68.0055618f, 22.0066261f, 68.0055618f);
pathB.cubicTo(23.2509995f, 68.0055618f, 24.4072189f, 67.6187515f, 25.3657818f, 66.9704056f);
pathB.lineTo(30.0599365f, 71.65625f);
pathB.lineTo(30.0600014f, 71.65625f);
pathB.cubicTo(30.2668133f, 71.875f, 30.5524693f, 71.9992828f, 30.868f, 71.9992828f);
pathB.cubicTo(31.4994049f, 71.9992828f, 32.0014687f, 71.4909363f, 32.0014687f, 70.8595276f);
pathB.cubicTo(32.0015335f, 70.5439072f, 31.875f, 70.2623444f, 31.65625f, 70.0555649f);
pathB.close();
pathB.moveTo(18.0054054f, 62.0027809f);
pathB.cubicTo(18.0054054f, 59.7925949f, 19.7970943f, 58.0009079f, 22.0072823f, 58.0009079f);
pathB.cubicTo(24.2174377f, 58.0009079f, 26.0091248f, 59.7925949f, 26.0091248f, 62.0027809f);
pathB.cubicTo(26.0091248f, 64.2129364f, 24.2174377f, 66.0046234f, 22.0072803f, 66.0046234f);
pathB.cubicTo(19.7970943f, 66.0045929f, 18.0054054f, 64.2129059f, 18.0054054f, 62.0027809f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpautobutler_dk29(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(0, 926);
path.lineTo(0, 0);
path.lineTo(1265, 0);
path.lineTo(1265, 926);
path.lineTo(0, 926);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(21, 162);
pathB.lineTo(21, 301);
pathB.lineTo(8.57224448e-15f, 301);
pathB.lineTo(6.12303177e-17f, 162);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skponlinecollege_org144(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(179, 407);
path.cubicTo(177.34314f, 407, 176, 408.34314f, 176, 410);
path.lineTo(176, 436);
path.cubicTo(176, 437.65686f, 177.34314f, 439, 179, 439);
path.lineTo(337.002289f, 439);
path.cubicTo(338.105835f, 438.998779f, 339, 438.103821f, 339, 437);
path.lineTo(339, 409);
path.cubicTo(339, 407.896362f, 338.10611f, 407.001526f, 337.002838f, 407);
path.lineTo(179, 407);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(179, 408);
pathB.lineTo(337, 408);
pathB.cubicTo(338.65686f, 408, 340, 408.895416f, 340, 410);
pathB.lineTo(340, 436);
pathB.cubicTo(340, 437.65686f, 338.65686f, 439, 337, 439);
pathB.lineTo(179, 439);
pathB.cubicTo(177.895432f, 439, 177, 437.65686f, 177, 436);
pathB.lineTo(177, 410);
pathB.cubicTo(177, 408.895416f, 177.895432f, 408, 179, 408);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpnational_com_au81(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(807, 817);
path.quadTo(806.585876f, 817.000122f, 806.292908f, 817.292908f);
path.quadTo(806.000122f, 817.585876f, 806, 818);
path.lineTo(806, 881);
path.lineTo(1111, 881);
path.lineTo(1111, 818);
path.quadTo(1110.99988f, 817.585876f, 1110.70715f, 817.292908f);
path.quadTo(1110.41406f, 817.000122f, 1110, 817);
path.lineTo(807, 817);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(807, 817);
pathB.lineTo(1110, 817);
pathB.cubicTo(1110.55225f, 817, 1111, 817.447693f, 1111, 818);
pathB.lineTo(1111, 880);
pathB.lineTo(806, 880);
pathB.lineTo(806, 818);
pathB.cubicTo(806, 817.447693f, 806.447693f, 817, 807, 817);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skprentacheat_com30(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(967, 263);
path.quadTo(966.585876f, 263.000092f, 966.292908f, 263.292908f);
path.quadTo(966.000122f, 263.585876f, 966, 264);
path.lineTo(966, 301);
path.lineTo(1214, 301);
path.lineTo(1214, 264);
path.quadTo(1213.99988f, 263.585876f, 1213.70715f, 263.292908f);
path.quadTo(1213.41406f, 263.000092f, 1213, 263);
path.lineTo(967, 263);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(967, 263);
pathB.lineTo(1213, 263);
pathB.cubicTo(1213.55225f, 263, 1214, 263.447723f, 1214, 264);
pathB.lineTo(1214, 300);
pathB.lineTo(966, 300);
pathB.lineTo(966, 264);
pathB.cubicTo(966, 263.447723f, 966.447693f, 263, 967, 263);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpbreakmystyle_com10(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(271.032867f, -5.32907052e-15f);
path.lineTo(56.9671326f, -5.16253706e-15f);
path.quadTo(52.7835083f, 3.69968891f, 48.7416f, 7.74160004f);
path.quadTo(1, 55.4831848f, 1, 123);
path.quadTo(1, 190.516815f, 48.7416f, 238.258392f);
path.quadTo(96.4831848f, 286, 164, 286);
path.quadTo(231.516815f, 286, 279.258392f, 238.258392f);
path.quadTo(327, 190.516815f, 327, 123);
path.quadTo(327, 55.4831848f, 279.258392f, 7.74160004f);
path.quadTo(275.216431f, 3.69964004f, 271.032867f, -5.32907052e-15f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(327, 123);
pathB.quadTo(327, 190.516815f, 279.258392f, 238.258392f);
pathB.quadTo(231.516815f, 286, 164, 286);
pathB.quadTo(96.4831848f, 286, 48.7416f, 238.258392f);
pathB.quadTo(1, 190.516815f, 1, 123);
pathB.quadTo(1, 55.4831848f, 48.7416f, 7.74160004f);
pathB.quadTo(96.4831848f, -40, 164, -40);
pathB.quadTo(231.516815f, -40, 279.258392f, 7.74160004f);
pathB.quadTo(327, 55.4831848f, 327, 123);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpsd_graphic_net104(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(475.421448f, 836.985962f);
path.lineTo(461.280975f, 841.990662f);
path.cubicTo(466.80899f, 857.609802f, 458.62854f, 874.752991f, 443.009399f, 880.281006f);
path.cubicTo(435.199829f, 883.044983f, 427.009247f, 882.381897f, 420.080048f, 879.075378f);
path.lineTo(413.620056f, 892.613037f);
path.quadTo(430.419983f, 900.629761f, 447.96701f, 894.43811f);
path.quadTo(448.00708f, 894.42395f, 448.014038f, 894.421509f);
path.quadTo(448.043976f, 894.410889f, 448.061066f, 894.404846f);
path.quadTo(465.596313f, 888.179932f, 473.613037f, 871.379944f);
path.quadTo(477.351227f, 863.546143f, 478, 855.549866f);
path.lineTo(478, 848.804321f);
path.quadTo(477.528076f, 842.93811f, 475.421448f, 836.985962f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(405.592621f, 909.435547f);
pathB.lineTo(390.578583f, 867.014099f);
pathB.lineTo(433, 852.000061f);
pathB.lineTo(490.435486f, 879.40741f);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpnaoxrane_ru23(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(458.703552f, 275.050262f);
path.quadTo(487.41687f, 273.000702f, 528, 273);
path.lineTo(529, 273);
path.quadTo(530.242371f, 273.000305f, 531.121338f, 273.878693f);
path.quadTo(531.999695f, 274.75766f, 532, 276);
path.lineTo(532, 378);
path.quadTo(531.990173f, 380.898804f, 503.296448f, 382.949738f);
path.quadTo(474.58313f, 384.999298f, 434, 385);
path.lineTo(433, 385);
path.quadTo(431.75766f, 384.999695f, 430.878693f, 384.121307f);
path.quadTo(430.000305f, 383.24234f, 430, 382);
path.lineTo(430, 280);
path.quadTo(430.009796f, 277.101196f, 458.703552f, 275.050262f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(528, 278);
pathB.lineTo(529, 278);
pathB.cubicTo(530.65686f, 278, 532, 278, 532, 278);
pathB.lineTo(532, 378);
pathB.cubicTo(532, 379.104584f, 488.123932f, 380, 434, 380);
pathB.lineTo(433, 380);
pathB.cubicTo(431.34314f, 380, 430, 380, 430, 380);
pathB.lineTo(430, 280);
pathB.cubicTo(430, 278.895416f, 473.876068f, 278, 528, 278);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skptcmevents_org23(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(465.503998f, 546);
path.lineTo(347, 546);
path.lineTo(347, 632);
path.lineTo(469.104248f, 632);
path.quadTo(470.79007f, 627.638672f, 471.833496f, 621.036255f);
path.quadTo(474.902588f, 601.562866f, 470.591064f, 574.024353f);
path.lineTo(468.507751f, 560.724426f);
path.quadTo(467.275146f, 552.856262f, 465.84668f, 547.288391f);
path.quadTo(465.670349f, 546.601501f, 465.503998f, 546);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(363.052246f, 542.495361f);
pathB.lineTo(463.779907f, 542.671143f);
pathB.cubicTo(464.829529f, 542.672974f, 466.946289f, 550.755676f, 468.507751f, 560.724426f);
pathB.lineTo(470.591064f, 574.024353f);
pathB.cubicTo(476.26178f, 610.226624f, 471.498932f, 639.557922f, 459.953003f, 639.537781f);
pathB.lineTo(368.727936f, 639.378601f);
pathB.cubicTo(351.933868f, 639.349304f, 337.053741f, 631.244324f, 335.492249f, 621.275574f);
pathB.lineTo(325.968597f, 560.475708f);
pathB.cubicTo(324.407104f, 550.506958f, 341.01001f, 542.456909f, 363.052246f, 542.495361f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpredbullskatearcade_es16(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(936.765625f, 458.965302f);
path.cubicTo(937.028442f, 453.863251f, 933.145813f, 449.864502f, 928.093445f, 450.033905f);
path.lineTo(661.882263f, 458.958862f);
path.lineTo(661.875366f, 458.959106f);
path.cubicTo(656.828369f, 459.13205f, 652.525085f, 463.399719f, 652.258545f, 468.496124f);
path.lineTo(652.258179f, 468.503662f);
path.lineTo(649.021729f, 531.322754f);
path.cubicTo(648.75885f, 536.424805f, 652.641479f, 540.423523f, 657.693848f, 540.25415f);
path.lineTo(923.905029f, 531.329163f);
path.cubicTo(928.955017f, 531.159851f, 933.262268f, 526.890442f, 933.528809f, 521.791565f);
path.lineTo(933.529175f, 521.784363f);
path.lineTo(936.765625f, 458.965302f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(661.882263f, 458.958862f);
pathB.lineTo(928.093445f, 450.033905f);
pathB.cubicTo(929.103882f, 450, 929.709961f, 454.108612f, 929.447144f, 459.210663f);
pathB.lineTo(926.210693f, 522.029724f);
pathB.cubicTo(926.079224f, 524.58075f, 925.153442f, 526.676208f, 924.143066f, 526.710083f);
pathB.lineTo(657.931885f, 535.635071f);
pathB.cubicTo(652.879456f, 535.804443f, 648.890259f, 533.873779f, 649.021729f, 531.322754f);
pathB.lineTo(652.258179f, 468.503662f);
pathB.cubicTo(652.520996f, 463.401611f, 656.829834f, 459.128235f, 661.882263f, 458.958862f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpfinanzasdigital_com9(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(156, 126);
path.quadTo(154.343552f, 126.000397f, 153.17157f, 127.17157f);
path.quadTo(152.000397f, 128.343552f, 152, 130);
path.lineTo(152, 174);
path.lineTo(1114, 174);
path.lineTo(1114, 130);
path.quadTo(1113.99963f, 128.343552f, 1112.82837f, 127.17157f);
path.quadTo(1111.65649f, 126.000397f, 1110, 126);
path.lineTo(156, 126);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(156, 126);
pathB.lineTo(1110, 126);
pathB.cubicTo(1111.65686f, 126, 1113, 127.790863f, 1113, 130);
pathB.lineTo(1113, 174);
pathB.lineTo(153, 174);
pathB.lineTo(153, 130);
pathB.cubicTo(153, 127.790863f, 154.34314f, 126, 156, 126);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skppartainasdemo250_org56(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(182.000015f, 645);
path.lineTo(182, 640);
path.cubicTo(174.322327f, 640, 166.644669f, 637.071045f, 160.786804f, 631.213196f);
path.cubicTo(149.071075f, 619.497437f, 149.071075f, 600.502563f, 160.786804f, 588.786804f);
path.lineTo(157.251266f, 585.251221f);
path.quadTo(147, 595.502502f, 147.000015f, 610);
path.quadTo(147, 624.482605f, 157.230255f, 634.727722f);
path.quadTo(157.251251f, 634.748779f, 157.251282f, 634.748779f);
path.quadTo(157.282852f, 634.780334f, 157.272263f, 634.769775f);
path.quadTo(167.517334f, 645, 182.000015f, 645);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(182, 659.497498f);
pathB.lineTo(206.748749f, 634.748718f);
pathB.lineTo(182.000015f, 610);
pathB.lineTo(132.502533f, 610);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpmlk_com326(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(154, 670);
path.cubicTo(151.238571f, 670, 149, 672.238586f, 149, 675);
path.lineTo(149, 710.001465f);
path.cubicTo(149.000809f, 712.209961f, 150.791367f, 714, 153, 714);
path.lineTo(189, 714);
path.cubicTo(191.209137f, 714, 193, 712.209167f, 193, 710);
path.lineTo(193, 675);
path.cubicTo(193, 672.238586f, 190.761429f, 670, 188, 670);
path.lineTo(154, 670);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(154, 671);
pathB.lineTo(188, 671);
pathB.cubicTo(190.761429f, 671, 193, 672.790833f, 193, 675);
pathB.lineTo(193, 710);
pathB.cubicTo(193, 712.761414f, 190.761429f, 715, 188, 715);
pathB.lineTo(154, 715);
pathB.cubicTo(151.238571f, 715, 149, 712.761414f, 149, 710);
pathB.lineTo(149, 675);
pathB.cubicTo(149, 672.790833f, 151.238571f, 671, 154, 671);
pathB.close();
Now, path ops natively intersect conics, quads, and cubics in any combination. There are still a class of cubic tests that fail and a handful of undiagnosed failures from skps and fuzz tests, but things are much better overall. Extended tests (150M+) run to completion in release in about 6 minutes; the standard test suite exceeds 100K and finishes in a few seconds on desktops. TBR=reed BUG=skia:3588 Review URL: https://codereview.chromium.org/1037953004
2015-04-20 15:31:59 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpcyclist_friends_gr52(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(50, 182);
path.lineTo(1215, 182);
path.lineTo(1215, 202);
path.quadTo(1214.99951f, 204.070572f, 1213.53552f, 205.535538f);
path.quadTo(1212.07056f, 206.999496f, 1210, 207);
path.lineTo(55, 207);
path.quadTo(52.9294319f, 206.999496f, 51.4644661f, 205.535538f);
path.quadTo(50.0004997f, 204.070572f, 50, 202);
path.lineTo(50, 182);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(50, 183);
pathB.lineTo(1215, 183);
pathB.lineTo(1215, 202);
pathB.cubicTo(1215, 204.761429f, 1212.76147f, 207, 1210, 207);
pathB.lineTo(55, 207);
pathB.cubicTo(52.238575f, 207, 50, 204.761429f, 50, 202);
pathB.lineTo(50, 183);
pathB.close();
Path ops formerly found the topmost unprocessed edge and determined its angle sort order to initialize the winding. This never worked correctly with cubics and was flaky with paths consisting mostly of vertical edges. This replacement shoots axis-aligned rays through all intersecting edges to find the outermost one either horizontally or vertically. The resulting code is smaller and twice as fast. To support this, most of the horizontal / vertical intersection code was rewritten and standardized, and old code supporting the top-directed winding was deleted. Contours were pointed to by an SkTDArray. Instead, put them in a linked list, and designate the list head with its own class to ensure that methods that take lists of contours start at the top. This change removed a large percentage of memory allocations used by path ops. TBR=reed@google.com BUG=skia:3588 Review URL: https://codereview.chromium.org/1111333002
2015-05-11 14:21:27 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_fj_p_com_22(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(172, 201);
path.lineTo(172, 202);
path.lineTo(220, 202);
path.cubicTo(221.65686f, 202, 223, 200.65686f, 223, 199);
path.cubicTo(223, 200.104568f, 221.65686f, 201, 220, 201);
path.lineTo(172, 201);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(161, 202);
pathB.lineTo(161, 199);
pathB.lineTo(223, 199.000015f);
pathB.lineTo(223, 202);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_lavoixdunord_fr_11(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(806, 57);
path.cubicTo(806, 55.3431473f, 807.34314f, 54, 809, 54);
path.lineTo(930, 54);
path.cubicTo(931.65686f, 54, 933, 55.3431473f, 933, 57);
path.lineTo(933, 91);
path.cubicTo(933, 92.6568527f, 931.65686f, 94, 930, 94);
path.lineTo(809, 94);
path.cubicTo(807.34314f, 94, 806, 92.6568527f, 806, 91);
path.lineTo(806, 57);
path.close();
path.moveTo(808, 58);
path.cubicTo(808, 56.8954315f, 808.895447f, 56, 810, 56);
path.lineTo(929, 56);
path.cubicTo(930.104553f, 56, 931, 56.8954315f, 931, 58);
path.lineTo(931, 90);
path.cubicTo(931, 91.1045685f, 930.104553f, 92, 929, 92);
path.lineTo(810, 92);
path.cubicTo(808.895447f, 92, 808, 91.1045685f, 808, 90);
path.lineTo(808, 58);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(806, 54);
pathB.lineTo(808, 56);
pathB.lineTo(935.02002f, 56.0200005f);
pathB.lineTo(933, 54);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skppptv_com_62(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(173, 5342);
path.quadTo(171.343536f, 5342.00049f, 170.17157f, 5343.17139f);
path.quadTo(169.000397f, 5344.34375f, 169, 5346);
path.lineTo(169, 5372);
path.lineTo(234, 5372);
path.lineTo(234, 5346);
path.quadTo(233.999603f, 5344.34375f, 232.82843f, 5343.17139f);
path.quadTo(231.656464f, 5342.00049f, 230, 5342);
path.lineTo(173, 5342);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(173, 5342);
pathB.lineTo(230, 5342);
pathB.cubicTo(231.65686f, 5342, 233, 5343.79102f, 233, 5346);
pathB.lineTo(233, 5372);
pathB.lineTo(169, 5372);
pathB.lineTo(169, 5346);
pathB.cubicTo(169, 5343.79102f, 170.790863f, 5342, 173, 5342);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_booking_com_68(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(90, 187);
path.cubicTo(90, 185.34314f, 91.3431473f, 184, 93, 184);
path.lineTo(588, 184);
path.cubicTo(589.65686f, 184, 591, 185.34314f, 591, 187);
path.lineTo(591, 218);
path.cubicTo(591, 219.65686f, 589.65686f, 221, 588, 221);
path.lineTo(93, 221);
path.cubicTo(91.3431473f, 221, 90, 219.65686f, 90, 218);
path.lineTo(90, 187);
path.close();
path.moveTo(92, 188);
path.cubicTo(92, 186.895432f, 92.8954315f, 186, 94, 186);
path.lineTo(587, 186);
path.cubicTo(588.104553f, 186, 589, 186.895432f, 589, 188);
path.lineTo(589, 217);
path.cubicTo(589, 218.104568f, 588.104553f, 219, 587, 219);
path.lineTo(94, 219);
path.cubicTo(92.8954315f, 219, 92, 218.104568f, 92, 217);
path.lineTo(92, 188);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(90, 184);
pathB.lineTo(92, 186);
pathB.lineTo(593.02002f, 186.020004f);
pathB.lineTo(591, 184);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_despegar_com_mx_272(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(635, 1788);
path.cubicTo(635, 1786.34314f, 636.34314f, 1785, 638, 1785);
path.lineTo(832, 1785);
path.cubicTo(833.65686f, 1785, 835, 1786.34314f, 835, 1788);
path.lineTo(835, 1812);
path.cubicTo(835, 1813.65686f, 833.65686f, 1815, 832, 1815);
path.lineTo(638, 1815);
path.cubicTo(636.34314f, 1815, 635, 1813.65686f, 635, 1812);
path.lineTo(635, 1788);
path.close();
path.moveTo(637, 1789);
path.cubicTo(637, 1787.89539f, 637.895447f, 1787, 639, 1787);
path.lineTo(831, 1787);
path.cubicTo(832.104553f, 1787, 833, 1787.89539f, 833, 1789);
path.lineTo(833, 1811);
path.cubicTo(833, 1812.10461f, 832.104553f, 1813, 831, 1813);
path.lineTo(639, 1813);
path.cubicTo(637.895447f, 1813, 637, 1812.10461f, 637, 1811);
path.lineTo(637, 1789);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(835, 1785);
pathB.lineTo(833, 1787);
pathB.lineTo(832.97998f, 1817.02002f);
pathB.lineTo(835, 1815);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_joomla_org_23(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(320, 347);
path.cubicTo(320, 344.238586f, 322.238586f, 342, 325, 342);
path.lineTo(416, 342);
path.cubicTo(418.761414f, 342, 421, 344.238586f, 421, 347);
path.cubicTo(421, 344.790863f, 418.761414f, 343, 416, 343);
path.lineTo(325, 343);
path.cubicTo(322.238586f, 343, 320, 344.790863f, 320, 347);
path.close();
path.moveTo(320, 378);
path.cubicTo(320, 380.761414f, 322.238586f, 383, 325, 383);
path.lineTo(416, 383);
path.cubicTo(418.761414f, 383, 421, 380.761414f, 421, 378);
path.cubicTo(421, 380.209137f, 418.761414f, 382, 416, 382);
path.lineTo(325, 382);
path.cubicTo(322.238586f, 382, 320, 380.209137f, 320, 378);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(320, 383);
pathB.lineTo(320, 378);
pathB.lineTo(421, 378.000031f);
pathB.lineTo(421, 383);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_macrumors_com_131(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(136, 14089);
path.lineTo(136, 14056);
path.lineTo(778, 14056);
path.lineTo(778, 14089);
path.quadTo(777.999573f, 14090.6562f, 776.82843f, 14091.8281f);
path.quadTo(775.656433f, 14093, 774, 14093);
path.lineTo(140, 14093);
path.quadTo(138.343552f, 14093, 137.17157f, 14091.8281f);
path.quadTo(136.000397f, 14090.6562f, 136, 14089);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(136, 14057);
pathB.lineTo(778, 14057);
pathB.lineTo(778, 14089);
pathB.cubicTo(778, 14091.209f, 776.209167f, 14093, 774, 14093);
pathB.lineTo(140, 14093);
pathB.cubicTo(137.790863f, 14093, 136, 14091.209f, 136, 14089);
pathB.lineTo(136, 14057);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_leadpages_net_84(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(377.1716f, 5910.17139f);
path.cubicTo(376.447723f, 5910.89551f, 376, 5911.89551f, 376, 5913);
path.lineTo(376, 5972);
path.cubicTo(376, 5974.20898f, 377.790863f, 5976, 380, 5976);
path.cubicTo(378.34314f, 5976, 377, 5974.20898f, 377, 5972);
path.lineTo(377, 5913);
path.cubicTo(377, 5912.17139f, 377.335785f, 5911.42188f, 377.878693f, 5910.87891f);
path.lineTo(377.1716f, 5910.17139f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(376, 5909);
pathB.lineTo(378.481873f, 5909);
pathB.lineTo(379.999878f, 5976);
pathB.lineTo(376, 5976);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_briian_com_34(skiatest::Reporter* reporter, const char* filename) {
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
path.moveTo(843, 216);
path.cubicTo(843, 213.238571f, 845.238586f, 211, 848, 211);
path.lineTo(1191, 211);
path.cubicTo(1193.76147f, 211, 1196, 213.238571f, 1196, 216);
path.lineTo(1196, 779);
path.cubicTo(1196, 781.761414f, 1193.76147f, 784, 1191, 784);
path.lineTo(848, 784);
path.cubicTo(845.238586f, 784, 843, 781.761414f, 843, 779);
path.lineTo(843, 216);
path.close();
path.moveTo(844, 217);
path.cubicTo(844, 214.238571f, 846.238586f, 212, 849, 212);
path.lineTo(1190, 212);
path.cubicTo(1192.76147f, 212, 1195, 214.238571f, 1195, 217);
path.lineTo(1195, 778);
path.cubicTo(1195, 779.65686f, 1192.76147f, 781, 1190, 781);
path.lineTo(849, 781);
path.cubicTo(846.238586f, 781, 844, 779.65686f, 844, 778);
path.lineTo(844, 217);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
pathB.moveTo(843, 784);
pathB.lineTo(843, 779);
pathB.lineTo(1196, 779.000061f);
pathB.lineTo(1196, 784);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_sciality_com_100(skiatest::Reporter* reporter, const char* filename) {
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
path.moveTo(162, 468);
path.cubicTo(159.790863f, 468, 158, 469.790863f, 158, 472);
path.lineTo(158, 528);
path.cubicTo(158, 530.209106f, 159.790863f, 532, 162, 532);
path.lineTo(275, 532);
path.cubicTo(277.209137f, 532, 279, 530.209106f, 279, 528);
path.lineTo(279, 472);
path.cubicTo(279, 469.790863f, 277.209137f, 468, 275, 468);
path.lineTo(162, 468);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
pathB.moveTo(275, 468);
pathB.cubicTo(276.65686f, 468, 278, 469.34314f, 278, 471);
pathB.lineTo(278, 529);
pathB.cubicTo(278, 530.65686f, 276.65686f, 532, 275, 532);
pathB.lineTo(161, 532);
pathB.cubicTo(159.34314f, 532, 158, 530.65686f, 158, 529);
pathB.lineTo(158, 471);
pathB.cubicTo(158, 469.34314f, 159.34314f, 468, 161, 468);
pathB.lineTo(275, 468);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_sciality_com_101(skiatest::Reporter* reporter, const char* filename) {
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
path.moveTo(162, 468);
path.cubicTo(159.790863f, 468, 158, 469.790863f, 158, 472);
path.lineTo(158, 528);
path.cubicTo(158, 530.209106f, 159.790863f, 532, 162, 532);
path.lineTo(275.009186f, 532);
path.cubicTo(276.661774f, 531.994995f, 278, 530.653748f, 278, 529);
path.lineTo(278, 471);
path.cubicTo(278, 469.346375f, 276.662079f, 468.005249f, 275.009705f, 468);
path.lineTo(162, 468);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
pathB.moveTo(161, 469);
pathB.lineTo(275, 469);
pathB.cubicTo(276.65686f, 469, 278, 469.895416f, 278, 471);
pathB.lineTo(278, 529);
pathB.cubicTo(278, 530.65686f, 276.65686f, 532, 275, 532);
pathB.lineTo(161, 532);
pathB.cubicTo(159.34314f, 532, 158, 530.65686f, 158, 529);
pathB.lineTo(158, 471);
pathB.cubicTo(158, 469.895416f, 159.34314f, 469, 161, 469);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
}
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_meb_gov_tr_5(skiatest::Reporter* reporter, const char* filename) {
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
path.moveTo(137.34314f, 145.34314f);
path.quadTo(139.687088f, 143.000793f, 143, 143);
path.lineTo(242, 143);
path.quadTo(245.312912f, 143.000793f, 247.65686f, 145.34314f);
path.quadTo(249.999207f, 147.687088f, 250, 151);
path.lineTo(250, 177);
path.lineTo(135, 177);
path.lineTo(135, 151);
path.quadTo(135.000793f, 147.687088f, 137.34314f, 145.34314f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
pathB.moveTo(135, 143);
pathB.lineTo(250, 143);
pathB.lineTo(250, 177);
pathB.lineTo(135, 177);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpwww_meb_gov_tr_6(skiatest::Reporter* reporter, const char* filename) {
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
path.moveTo(143, 143);
path.quadTo(139.687088f, 143.000793f, 137.34314f, 145.34314f);
path.quadTo(135.000793f, 147.687088f, 135, 151);
path.lineTo(135, 177);
path.lineTo(250, 177);
path.lineTo(250, 151);
path.quadTo(249.999207f, 147.687088f, 247.65686f, 145.34314f);
path.quadTo(245.312912f, 143.000793f, 242, 143);
path.lineTo(143, 143);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
pathB.moveTo(143, 143);
pathB.lineTo(242, 143);
pathB.cubicTo(245.865997f, 143, 249, 146.581726f, 249, 151);
pathB.lineTo(249, 177);
pathB.lineTo(135, 177);
pathB.lineTo(135, 151);
pathB.cubicTo(135, 146.581726f, 138.581726f, 143, 143, 143);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
}
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpgithub_io_25(skiatest::Reporter* reporter, const char* filename) {
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
path.moveTo(1001.87866f, 14.8786793f);
path.quadTo(1002.75745f, 14.0001001f, 1004, 14);
path.lineTo(1105, 14);
path.quadTo(1106.24255f, 14.0001001f, 1107.12134f, 14.8786793f);
path.quadTo(1107.99988f, 15.7574596f, 1108, 17);
path.lineTo(1108, 41);
path.quadTo(1107.99988f, 42.2425423f, 1107.12134f, 43.1213188f);
path.quadTo(1106.24255f, 43.9999008f, 1105, 44);
path.lineTo(1004, 44);
path.quadTo(1002.75745f, 43.9999008f, 1001.87866f, 43.1213188f);
path.quadTo(1001.00012f, 42.2425423f, 1001, 41);
path.lineTo(1001, 17);
path.quadTo(1001.00012f, 15.7574596f, 1001.87866f, 14.8786793f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
pathB.moveTo(1005, 16);
pathB.lineTo(1104, 16);
pathB.cubicTo(1105.10461f, 16, 1106, 16.8954296f, 1106, 18);
pathB.lineTo(1106, 40);
pathB.cubicTo(1106, 41.1045685f, 1105.10461f, 42, 1104, 42);
pathB.lineTo(1005, 42);
pathB.cubicTo(1003.89545f, 42, 1003, 41.1045685f, 1003, 40);
pathB.lineTo(1003, 18);
pathB.cubicTo(1003, 16.8954296f, 1003.89545f, 16, 1005, 16);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpgithub_io_26(skiatest::Reporter* reporter, const char* filename) {
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
path.moveTo(1001.87866f, 14.8786793f);
path.quadTo(1002.75745f, 14.0001001f, 1004, 14);
path.lineTo(1105, 14);
path.quadTo(1106.24255f, 14.0001001f, 1107.12134f, 14.8786793f);
path.quadTo(1107.99988f, 15.7574596f, 1108, 17);
path.lineTo(1108, 41);
path.quadTo(1107.99988f, 42.2425423f, 1107.12134f, 43.1213188f);
path.quadTo(1106.24255f, 43.9999008f, 1105, 44);
path.lineTo(1004, 44);
path.quadTo(1002.75745f, 43.9999008f, 1001.87866f, 43.1213188f);
path.quadTo(1001.00012f, 42.2425423f, 1001, 41);
path.lineTo(1001, 17);
path.quadTo(1001.00012f, 15.7574596f, 1001.87866f, 14.8786793f);
path.close();
path.moveTo(1003, 18);
path.cubicTo(1003, 16.8954296f, 1003.89545f, 16, 1005, 16);
path.lineTo(1104, 16);
path.cubicTo(1105.10461f, 16, 1106, 16.8954296f, 1106, 18);
path.lineTo(1106, 40);
path.cubicTo(1106, 41.1045685f, 1105.10461f, 42, 1104, 42);
path.lineTo(1005, 42);
path.cubicTo(1003.89545f, 42, 1003, 41.1045685f, 1003, 40);
path.lineTo(1003, 18);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
pathB.moveTo(1108, 14);
pathB.lineTo(1106, 16);
pathB.lineTo(1105.97998f, 46.0200005f);
pathB.lineTo(1108, 44);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12101 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-02 07:02:02 +00:00
}
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpskpicture14(skiatest::Reporter* reporter, const char* filename) {
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12208 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-09 07:02:23 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12208 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-09 07:02:23 +00:00
path.moveTo(0, 994);
path.lineTo(0, 0);
path.lineTo(512, 0);
path.lineTo(512, 994);
path.lineTo(0, 994);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12208 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-09 07:02:23 +00:00
pathB.moveTo(-317, 168);
pathB.quadTo(-317, 166.757385f, -316.121338f, 165.878662f);
pathB.quadTo(-315.242645f, 165, -314, 165);
pathB.lineTo(320, 165);
pathB.quadTo(321.242615f, 165, 322.121338f, 165.878662f);
pathB.quadTo(323, 166.757385f, 323, 168);
pathB.lineTo(323, 193);
pathB.lineTo(-317, 193);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12208 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-09 07:02:23 +00:00
}
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpskpicture15(skiatest::Reporter* reporter, const char* filename) {
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12208 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-09 07:02:23 +00:00
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12208 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-09 07:02:23 +00:00
path.moveTo(0, 193);
path.lineTo(323, 193);
path.lineTo(323, 168);
path.quadTo(323, 166.757385f, 322.121338f, 165.878662f);
path.quadTo(321.242615f, 165, 320, 165);
path.lineTo(0, 165);
path.lineTo(0, 193);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kInverseWinding);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12208 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-09 07:02:23 +00:00
pathB.moveTo(-314, 165);
pathB.lineTo(320, 165);
pathB.cubicTo(321.65686f, 165, 323, 166.34314f, 323, 168);
pathB.lineTo(323, 192);
pathB.lineTo(-317, 192);
pathB.lineTo(-317, 168);
pathB.cubicTo(-317, 166.34314f, -315.65686f, 165, -314, 165);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
}
static void skpelpais_com_18(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
path.moveTo(183, 8507);
path.lineTo(552, 8506.99023f);
path.lineTo(552, 8508);
path.lineTo(183, 8508);
path.lineTo(183, 8507);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
pathB.moveTo(183, 8508);
pathB.lineTo(183, 8506.99023f);
pathB.lineTo(552, 8507);
pathB.lineTo(552, 8508);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Sanitizing source files in Housekeeper-Nightly git-svn-id: http://skia.googlecode.com/svn/trunk@12208 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-09 07:02:23 +00:00
}
fix line intersection git-svn-id: http://skia.googlecode.com/svn/trunk@12197 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-08 18:00:01 +00:00
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
static void skpnamecheap_com_405(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
path.moveTo(140, 1000);
path.lineTo(140, 842);
path.lineTo(141, 842);
path.lineTo(141.14502f, 1000);
path.lineTo(140, 1000);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
pathB.moveTo(140, 842);
pathB.lineTo(141.008835f, 837.9646f);
pathB.lineTo(141.235291f, 1109.05884f);
pathB.lineTo(140, 1114);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
}
static void skpwww_alrakoba_net_62(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
path.moveTo(134.34314f, 9802.34277f);
path.quadTo(132, 9804.68652f, 132, 9808);
path.lineTo(132, 9822);
path.quadTo(132, 9825.31348f, 134.34314f, 9827.65723f);
path.quadTo(136.686295f, 9830, 140, 9830);
path.lineTo(140.028473f, 9830);
path.lineTo(139.877213f, 9828.90723f);
path.quadTo(137.692032f, 9828.5332f, 136.050247f, 9826.65723f);
path.quadTo(134, 9824.31348f, 134, 9821);
path.lineTo(134, 9809);
path.quadTo(134, 9806.10059f, 136.050247f, 9804.0498f);
path.lineTo(134.34314f, 9802.34277f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
pathB.moveTo(132, 9800);
pathB.lineTo(135.962357f, 9800);
pathB.lineTo(140, 9830);
pathB.lineTo(132, 9830);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
}
static void skpwww_cityads_ru_249(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
path.moveTo(817.464478f, 11.4644661f);
path.quadTo(818.928955f, 10, 821, 10);
path.lineTo(998, 10);
path.quadTo(999.082947f, 10, 1000, 10.4003992f);
path.lineTo(1000, 13.3527431f);
path.quadTo(999.917603f, 13.2607508f, 999.82843f, 13.1715727f);
path.quadTo(998.65686f, 12, 997, 12);
path.lineTo(822, 12);
path.quadTo(820.34314f, 12, 819.17157f, 13.1715727f);
path.quadTo(818, 14.3431454f, 818, 16);
path.lineTo(818, 28);
path.quadTo(818, 29.6568546f, 819.17157f, 30.8284264f);
path.quadTo(820.34314f, 32, 822, 32);
path.lineTo(997, 32);
path.quadTo(998.65686f, 32, 999.82843f, 30.8284264f);
path.quadTo(999.917603f, 30.7392426f, 1000, 30.6472569f);
path.lineTo(1000, 33.5996017f);
path.quadTo(999.082947f, 34, 998, 34);
path.lineTo(821, 34);
path.quadTo(818.928955f, 34, 817.464478f, 32.5355339f);
path.quadTo(816, 31.0710678f, 816, 29);
path.lineTo(816, 15);
path.quadTo(816, 12.9289322f, 817.464478f, 11.4644661f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
pathB.moveTo(1003, 10);
pathB.lineTo(1000, 13);
pathB.lineTo(999.969971f, 37.0299988f);
pathB.lineTo(1003, 34);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
}
static void skpwww_dealnews_com_315(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
path.moveTo(966.464478f, 4261.46436f);
path.quadTo(965, 4262.92871f, 965, 4265);
path.lineTo(965, 4276);
path.quadTo(965, 4278.07129f, 966.464478f, 4279.53564f);
path.quadTo(967.928955f, 4281, 970, 4281);
path.lineTo(970.020325f, 4281);
path.lineTo(969.887512f, 4279.81641f);
path.quadTo(968.928284f, 4279.48145f, 968.17157f, 4278.53564f);
path.quadTo(967, 4277.07129f, 967, 4275);
path.lineTo(967, 4266);
path.quadTo(967, 4264.44287f, 968.035217f, 4263.31396f);
path.lineTo(968, 4263);
path.lineTo(966.464478f, 4261.46436f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
pathB.moveTo(965, 4260);
pathB.lineTo(967.716675f, 4260);
pathB.lineTo(970, 4281);
pathB.lineTo(965, 4281);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
}
static void skpwww_inmotionhosting_com_9(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
path.moveTo(991.633911f, 1839);
path.lineTo(964.265015f, 1839);
path.lineTo(963.734985f, 1893.73242f);
path.lineTo(991.3703f, 1894);
path.lineTo(1018.23492f, 1894);
path.lineTo(1018.76501f, 1839.2627f);
path.lineTo(991.638184f, 1839);
path.lineTo(991.633911f, 1839);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
pathB.moveTo(964.267578f, 1838.73499f);
pathB.lineTo(1019.26501f, 1839.26758f);
pathB.lineTo(1018.73242f, 1894.26501f);
pathB.lineTo(963.734985f, 1893.73242f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
}
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
static void skpwww_alucinados_net_101(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(1129.53552f, 1164.46448f);
path.lineTo(1128, 1166);
path.lineTo(1128.12231f, 1166.49548f);
path.quadTo(1129, 1167.56592f, 1129, 1169);
path.lineTo(1129, 1170.05054f);
path.lineTo(1130.34509f, 1175.49878f);
path.quadTo(1131, 1174.38513f, 1131, 1173);
path.lineTo(1131, 1168);
path.quadTo(1131, 1165.92896f, 1129.53552f, 1164.46448f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(1131, 1163);
pathB.lineTo(-43515.8555f, -177415.594f);
pathB.lineTo(1129.76465f, 1173.05884f);
pathB.lineTo(1131, 1178);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_hairjobsearch_com_31(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(143.292892f, 0.707106769f);
path.quadTo(143, 0.414213538f, 143, 0);
path.lineTo(1123, 0);
path.quadTo(1123, 0.414213538f, 1122.70715f, 0.707106769f);
path.quadTo(1122.41418f, 1, 1122, 1);
path.lineTo(144, 1);
path.quadTo(143.585785f, 1, 143.292892f, 0.707106769f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(143, 1);
pathB.lineTo(144, 0);
pathB.lineTo(1122, 0);
pathB.lineTo(1123, 1);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_heartiste_wordpress_com_86(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(741, 9432);
path.lineTo(761, 9431.99023f);
path.lineTo(761, 9433);
path.lineTo(741, 9433);
path.lineTo(741, 9432);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(741, 9433);
pathB.lineTo(741, 9431.99023f);
pathB.lineTo(761, 9432);
pathB.lineTo(761, 9433);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_argus_presse_fr_41(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(1000, 343);
path.lineTo(165, 343);
path.lineTo(165, 364.869873f);
path.lineTo(1000, 364.869873f);
path.lineTo(1000, 343);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(165, 343.000031f);
pathB.lineTo(1000, 343.000031f);
pathB.lineTo(1000, 364.869904f);
pathB.lineTo(165, 364.869904f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_320kbps_net_2231(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(838, 9125);
path.lineTo(862, 9124.99023f);
path.lineTo(862, 9126);
path.lineTo(838, 9126);
path.lineTo(838, 9125);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(838, 9126);
pathB.lineTo(838, 9124.99023f);
pathB.lineTo(862, 9125);
pathB.lineTo(862, 9126);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_exystence_net_61(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(143, 9075);
path.lineTo(316, 9075);
path.lineTo(316, 9073.99023f);
path.lineTo(143, 9074);
path.lineTo(143, 9075);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(143, 9075);
pathB.lineTo(143, 9073.99023f);
pathB.lineTo(316, 9074);
pathB.lineTo(316, 9075);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_trashness_com_36(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(541.5f, 4835.99512f);
path.lineTo(91.5f, 4836);
path.lineTo(91.5f, 4836.5f);
path.lineTo(541.5f, 4836.5f);
path.lineTo(541.5f, 4835.99512f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(91.5f, 4836.5f);
pathB.lineTo(91.5f, 4835.99512f);
pathB.lineTo(541.5f, 4836);
pathB.lineTo(541.5f, 4836.5f);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_getgold_jp_731(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(284.878693f, 10134.8789f);
path.quadTo(284, 10135.7578f, 284, 10137);
path.lineTo(284, 10216);
path.quadTo(284, 10217.2422f, 284.878693f, 10218.1211f);
path.quadTo(285.125122f, 10218.3672f, 285.40213f, 10218.5459f);
path.lineTo(286, 10138);
path.lineTo(286, 10136);
path.lineTo(284.878693f, 10134.8789f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(284, 10134);
pathB.lineTo(286.05957f, 10129.8809f);
pathB.lineTo(285.399994f, 10216.2002f);
pathB.lineTo(284, 10219);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_maturesupertube_com_21(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(3.17157292f, 11831.1719f);
path.quadTo(4.34314585f, 11830, 6, 11830);
path.lineTo(1259, 11830);
path.quadTo(1260.65686f, 11830, 1261.82837f, 11831.1719f);
path.quadTo(1263, 11832.3428f, 1263, 11834);
path.lineTo(1263, 11848);
path.quadTo(1263, 11849.6572f, 1261.82837f, 11850.8281f);
path.quadTo(1260.65686f, 11852, 1259, 11852);
path.lineTo(6, 11852);
path.quadTo(4.34314585f, 11852, 3.17157292f, 11850.8281f);
path.quadTo(2, 11849.6572f, 2, 11848);
path.lineTo(2, 11834);
path.quadTo(2, 11832.3428f, 3.17157292f, 11831.1719f);
path.close();
path.moveTo(3.87867975f, 11831.8789f);
path.quadTo(4.7573595f, 11831, 6, 11831);
path.lineTo(1259, 11831);
path.quadTo(1260.24268f, 11831, 1261.12134f, 11831.8789f);
path.quadTo(1262, 11832.7578f, 1262, 11834);
path.lineTo(1262, 11848);
path.quadTo(1262, 11849.2422f, 1261.12134f, 11850.1211f);
path.quadTo(1260.24268f, 11851, 1259, 11851);
path.lineTo(6, 11851);
path.quadTo(4.7573595f, 11851, 3.87867975f, 11850.1211f);
path.quadTo(3, 11849.2422f, 3, 11848);
path.lineTo(3, 11834);
path.quadTo(3, 11832.7578f, 3.87867975f, 11831.8789f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(2, 11830);
pathB.lineTo(4.5f, 11832.5f);
pathB.lineTo(1260.5f, 11832.5f);
pathB.lineTo(1263, 11830);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_hubbyscook_com_22(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(1000, 902.329346f);
path.quadTo(998, 905.250427f, 998, 909);
path.lineTo(998, 910);
path.quadTo(998, 913.749573f, 1000, 916.670654f);
path.lineTo(1000, 902.329346f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(998, 910);
pathB.lineTo(998, 909);
pathB.quadTo(998, 904.029419f, 1001.51471f, 900.514709f);
pathB.quadTo(1005.02942f, 897, 1010, 897);
pathB.lineTo(1011, 897);
pathB.quadTo(1015.14215f, 897, 1018.07104f, 900.514709f);
pathB.quadTo(1021, 904.029419f, 1021, 909);
pathB.lineTo(1021, 910);
pathB.quadTo(1021, 914.142151f, 1018.07104f, 917.071045f);
pathB.quadTo(1015.14215f, 920, 1011, 920);
pathB.lineTo(1010, 920);
pathB.quadTo(1005.02942f, 920, 1001.51471f, 917.071045f);
pathB.quadTo(998, 914.142151f, 998, 910);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kDifference_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_gruposejaumdivulgador_com_br_4(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(610.5f, 5.78626502e-14f);
path.lineTo(1083.5f, -6.12303177e-17f);
path.lineTo(1083.5f, 469);
path.lineTo(610.5f, 469);
path.lineTo(610.5f, 5.78626502e-14f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(611, 0);
pathB.lineTo(1084, 0);
pathB.lineTo(1084, 469);
pathB.lineTo(611, 469);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_phototransferapp_com_24(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(85.6091843f, 5.92893219f);
path.quadTo(89.6041641f, 3, 93.7462997f, 3);
path.lineTo(1212.74634f, 3);
path.quadTo(1216.88843f, 3, 1218.75134f, 5.92893219f);
path.quadTo(1220.61414f, 8.85775471f, 1219.10669f, 12.9996767f);
path.quadTo(1220.46338f, 9.27196693f, 1218.4939f, 6.63603878f);
path.quadTo(1216.52441f, 4, 1212.38232f, 4);
path.lineTo(93.3823318f, 4);
path.quadTo(89.2401962f, 4, 85.3518219f, 6.63603878f);
path.quadTo(81.4634476f, 9.27207756f, 80.1065979f, 13);
path.quadTo(81.614212f, 8.85786438f, 85.6091843f, 5.92893219f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(83.7462997f, 3);
pathB.lineTo(1222.74634f, 3);
pathB.lineTo(1219.10657f, 13);
pathB.lineTo(80.1065979f, 13);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
static void skpwww_phototransferapp_com_24x(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(85.6091843f, 5.92893219f);
path.quadTo(89.6041641f, 3, 93.7462997f, 3);
path.lineTo(112.74634f, 3);
path.quadTo(116.88843f, 3, 118.75134f, 5.92893219f);
path.quadTo(120.61414f, 8.85775471f, 119.10669f, 12.9996767f);
path.quadTo(120.46338f, 9.27196693f, 118.4939f, 6.63603878f);
path.quadTo(116.52441f, 4, 112.38232f, 4);
path.lineTo(93.3823318f, 4);
path.quadTo(89.2401962f, 4, 85.3518219f, 6.63603878f);
path.quadTo(81.4634476f, 9.27207756f, 80.1065979f, 13);
path.quadTo(81.614212f, 8.85786438f, 85.6091843f, 5.92893219f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(83.7462997f, 3);
pathB.lineTo(122.74634f, 3);
pathB.lineTo(119.10657f, 13);
pathB.lineTo(80.1065979f, 13);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
static void skpwww_helha_be_109(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(117.686981f, 3339.08423f);
path.lineTo(109.533035f, 3350.72925f);
path.quadTo(107.120552f, 3354.17456f, 103.879379f, 3354.41821f);
path.quadTo(100.638504f, 3354.66187f, 98.4674301f, 3351.56177f);
path.quadTo(100.87973f, 3355.00635f, 104.291222f, 3355.00635f);
path.quadTo(107.70298f, 3355.00635f, 110.115463f, 3351.56104f);
path.lineTo(118, 3340.30078f);
path.lineTo(118, 3339.53125f);
path.lineTo(117.686981f, 3339.08423f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(118.269409f, 3339.91602f);
pathB.lineTo(117.686981f, 3339.08423f);
pathB.lineTo(98.4669647f, 3351.56104f);
pathB.lineTo(104.291214f, 3359.87891f);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_cooksnaps_com_32(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(509.34021f, 176);
path.lineTo(505, 176);
path.quadTo(500.964233f, 176, 497.299988f, 176.896912f);
path.quadTo(493.678162f, 177.952286f, 490.183014f, 179.9702f);
path.lineTo(489.316986f, 180.4702f);
path.quadTo(485.175385f, 182.861359f, 482.115265f, 186.082397f);
path.quadTo(479.69455f, 188.700821f, 477.995941f, 191.618286f);
path.quadTo(476.316772f, 194.550476f, 475.258759f, 197.959335f);
path.quadTo(474, 202.218979f, 474, 207);
path.lineTo(474, 208);
path.quadTo(474, 212.03569f, 474.896851f, 215.699799f);
path.quadTo(475.950256f, 219.315002f, 477.962708f, 222.803986f);
path.lineTo(477.970215f, 222.816986f);
path.lineTo(478.470215f, 223.683014f);
path.quadTo(478.474915f, 223.691162f, 478.479645f, 223.69931f);
path.quadTo(480.867981f, 227.831055f, 484.082947f, 230.885254f);
path.quadTo(486.701447f, 233.305939f, 489.61908f, 235.004517f);
path.quadTo(492.550232f, 236.682983f, 495.957611f, 237.740738f);
path.quadTo(500.217987f, 239, 505, 239);
path.lineTo(509.482178f, 239);
path.quadTo(515.299133f, 238.212051f, 520.801941f, 235.038513f);
path.quadTo(520.809509f, 235.034164f, 520.817017f, 235.0298f);
path.lineTo(521.683044f, 234.5298f);
path.quadTo(521.692078f, 234.524582f, 521.701111f, 234.519348f);
path.quadTo(532.80603f, 228.09938f, 536.126709f, 215.70639f);
path.quadTo(539.450134f, 203.303314f, 533.029785f, 192.183014f);
path.lineTo(532.529785f, 191.316986f);
path.quadTo(526.109497f, 180.196686f, 513.706421f, 176.873276f);
path.quadTo(511.503082f, 176.282898f, 509.34021f, 176);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(478.470215f, 223.683014f);
pathB.lineTo(477.970215f, 222.816986f);
pathB.quadTo(471.549896f, 211.696686f, 474.873322f, 199.293594f);
pathB.quadTo(478.196686f, 186.890503f, 489.316986f, 180.4702f);
pathB.lineTo(490.183014f, 179.9702f);
pathB.quadTo(501.303345f, 173.549896f, 513.706421f, 176.873276f);
pathB.quadTo(526.109497f, 180.196686f, 532.529785f, 191.316986f);
pathB.lineTo(533.029785f, 192.183014f);
pathB.quadTo(539.450134f, 203.303314f, 536.126709f, 215.70639f);
pathB.quadTo(532.803345f, 228.109497f, 521.683044f, 234.5298f);
pathB.lineTo(520.817017f, 235.0298f);
pathB.quadTo(509.696686f, 241.450104f, 497.29361f, 238.126709f);
pathB.quadTo(484.890533f, 234.803314f, 478.470215f, 223.683014f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
static void skpwww_cooksnaps_com_32a(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(497.299988f, 176.896912f);
path.quadTo(493.678162f, 177.952286f, 490.183014f, 179.9702f);
path.lineTo(489.316986f, 180.4702f);
path.quadTo(485.175385f, 182.861359f, 482.115265f, 186.082397f);
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(474.873322f, 199.293594f);
pathB.quadTo(478.196686f, 186.890503f, 489.316986f, 180.4702f);
pathB.lineTo(490.183014f, 179.9702f);
pathB.quadTo(501.303345f, 173.549896f, 513.706421f, 176.873276f);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
static void skpwww_contextualnewsfeeds_com_346(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(460.257355f, 1202.27808f);
path.lineTo(460.257355f, 1204.27808f);
path.quadTo(461.081207f, 1204.27808f, 461.665161f, 1203.69873f);
path.lineTo(461.67157f, 1203.69238f);
path.lineTo(466.621307f, 1198.74268f);
path.quadTo(466.623993f, 1198.73999f, 466.626648f, 1198.7373f);
path.quadTo(466.914185f, 1198.44604f, 466.914185f, 1198.03552f);
path.quadTo(466.914215f, 1197.62122f, 466.621307f, 1197.32837f);
path.lineTo(465.914215f, 1196.62122f);
path.lineTo(460.257355f, 1202.27808f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(460.257355f, 1205.10657f);
pathB.lineTo(458.828979f, 1203.67822f);
pathB.lineTo(465.914215f, 1196.62122f);
pathB.lineTo(467.32843f, 1198.03552f);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_pindosiya_com_99(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(899.17157f, 548.17157f);
path.quadTo(898, 549.34314f, 898, 551);
path.lineTo(898, 556);
path.lineTo(899.027283f, 556);
path.lineTo(900.02356f, 551.602844f);
path.quadTo(900.06073f, 551.297058f, 900.156555f, 551.015747f);
path.lineTo(900.5f, 549.5f);
path.lineTo(899.17157f, 548.17157f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(898, 547);
pathB.lineTo(901.086914f, 547);
pathB.lineTo(899, 556);
pathB.lineTo(898, 556);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_karnivool_com_au_11(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(0, 1431);
path.lineTo(0, 775);
path.lineTo(1265, 775);
path.lineTo(1265, 1431);
path.lineTo(0, 1431);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(32.3243904f, 851);
pathB.lineTo(459.324402f, 851);
pathB.lineTo(427, 1081);
pathB.lineTo(-3.81469727e-06f, 1081);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_tunero_de_24(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(1020.79303f, 2252);
path.quadTo(1018.72198f, 2252, 1016.86798f, 2253.46436f);
path.quadTo(1015.02032f, 2254.92383f, 1014.4668f, 2256.98584f);
path.lineTo(1014.46301f, 2257);
path.lineTo(1011.53705f, 2268);
path.quadTo(1010.98615f, 2270.07104f, 1012.06104f, 2271.53564f);
path.quadTo(1013.13599f, 2273, 1015.20703f, 2273);
path.lineTo(1083.20703f, 2273);
path.quadTo(1085.27808f, 2273, 1087.13208f, 2271.53564f);
path.quadTo(1088.97144f, 2270.08252f, 1089.52832f, 2268.03271f);
path.lineTo(1089.53711f, 2268);
path.lineTo(1092.46301f, 2257);
path.lineTo(1092.4679f, 2256.98145f);
path.quadTo(1093.00916f, 2254.92236f, 1091.93909f, 2253.46436f);
path.quadTo(1090.86414f, 2252, 1088.79297f, 2252);
path.lineTo(1020.79303f, 2252);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(1011.53705f, 2268);
pathB.lineTo(1014.46301f, 2257);
pathB.quadTo(1015.01392f, 2254.92896f, 1016.86798f, 2253.46436f);
pathB.quadTo(1018.72198f, 2252, 1020.79303f, 2252);
pathB.lineTo(1088.79297f, 2252);
pathB.quadTo(1090.86414f, 2252, 1091.93909f, 2253.46436f);
pathB.quadTo(1093.01392f, 2254.92896f, 1092.46301f, 2257);
pathB.lineTo(1089.53711f, 2268);
pathB.quadTo(1088.98608f, 2270.07104f, 1087.13208f, 2271.53564f);
pathB.quadTo(1085.27808f, 2273, 1083.20703f, 2273);
pathB.lineTo(1015.20703f, 2273);
pathB.quadTo(1013.13599f, 2273, 1012.06104f, 2271.53564f);
pathB.quadTo(1010.98615f, 2270.07104f, 1011.53705f, 2268);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_docgelo_com_66(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(22.5f, 24174.5f);
path.lineTo(185.5f, 24174.498f);
path.lineTo(185.5f, 24174.75f);
path.lineTo(22.5f, 24174.75f);
path.lineTo(22.5f, 24174.5f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(22.5f, 24174.75f);
pathB.lineTo(22.5f, 24174.498f);
pathB.lineTo(185.5f, 24174.5f);
pathB.lineTo(185.5f, 24174.75f);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_kpopexplorer_net_22(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(1000, 866.329346f);
path.quadTo(998, 869.250427f, 998, 873);
path.lineTo(998, 874);
path.quadTo(998, 877.749573f, 1000, 880.670654f);
path.lineTo(1000, 866.329346f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(998, 874);
pathB.lineTo(998, 873);
pathB.quadTo(998, 868.029419f, 1001.51471f, 864.514709f);
pathB.quadTo(1005.02942f, 861, 1010, 861);
pathB.lineTo(1011, 861);
pathB.quadTo(1015.14215f, 861, 1018.07104f, 864.514709f);
pathB.quadTo(1021, 868.029419f, 1021, 873);
pathB.lineTo(1021, 874);
pathB.quadTo(1021, 878.142151f, 1018.07104f, 881.071045f);
pathB.quadTo(1015.14215f, 884, 1011, 884);
pathB.lineTo(1010, 884);
pathB.quadTo(1005.02942f, 884, 1001.51471f, 881.071045f);
pathB.quadTo(998, 878.142151f, 998, 874);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kDifference_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_artblart_com_8(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(22.5f, 24527.25f);
path.lineTo(45, 24527.248f);
path.lineTo(45, 24527.5f);
path.lineTo(22.5f, 24527.5f);
path.lineTo(22.5f, 24527.25f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(22.5f, 24527.5f);
pathB.lineTo(22.5f, 24527.248f);
pathB.lineTo(45, 24527.25f);
pathB.lineTo(45, 24527.5f);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_jessicaslens_wordpress_com_222(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(1000, 844.329346f);
path.quadTo(998, 847.250427f, 998, 851);
path.lineTo(998, 852);
path.quadTo(998, 855.749573f, 1000, 858.670654f);
path.lineTo(1000, 844.329346f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(998, 852);
pathB.lineTo(998, 851);
pathB.quadTo(998, 846.029419f, 1001.51471f, 842.514709f);
pathB.quadTo(1005.02942f, 839, 1010, 839);
pathB.lineTo(1011, 839);
pathB.quadTo(1015.14215f, 839, 1018.07104f, 842.514709f);
pathB.quadTo(1021, 846.029419f, 1021, 851);
pathB.lineTo(1021, 852);
pathB.quadTo(1021, 856.142151f, 1018.07104f, 859.071045f);
pathB.quadTo(1015.14215f, 862, 1011, 862);
pathB.lineTo(1010, 862);
pathB.quadTo(1005.02942f, 862, 1001.51471f, 859.071045f);
pathB.quadTo(998, 856.142151f, 998, 852);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kDifference_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_simplysaru_com_40(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(1000, 866.329346f);
path.quadTo(998, 869.250427f, 998, 873);
path.lineTo(998, 874);
path.quadTo(998, 877.749573f, 1000, 880.670654f);
path.lineTo(1000, 866.329346f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(998, 874);
pathB.lineTo(998, 873);
pathB.quadTo(998, 868.029419f, 1001.51471f, 864.514709f);
pathB.quadTo(1005.02942f, 861, 1010, 861);
pathB.lineTo(1011, 861);
pathB.quadTo(1015.14215f, 861, 1018.07104f, 864.514709f);
pathB.quadTo(1021, 868.029419f, 1021, 873);
pathB.lineTo(1021, 874);
pathB.quadTo(1021, 878.142151f, 1018.07104f, 881.071045f);
pathB.quadTo(1015.14215f, 884, 1011, 884);
pathB.lineTo(1010, 884);
pathB.quadTo(1005.02942f, 884, 1001.51471f, 881.071045f);
pathB.quadTo(998, 878.142151f, 998, 874);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kDifference_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_partsdata_de_53(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(407, 119);
path.lineTo(407, 28);
path.lineTo(647, 28);
path.lineTo(647, 119);
path.lineTo(407, 119);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(471.228394f, 64.443222f);
pathB.cubicTo(471.193878f, 60.953373f, 470.234985f, 52.4797592f, 462.201569f, 46.6231461f);
pathB.cubicTo(454.168152f, 40.7665405f, 446.592804f, 41.993145f, 443.033936f, 42.8568878f);
pathB.cubicTo(439.475098f, 43.7206268f, 436.978638f, 45.4052658f, 436.304871f, 45.7851906f);
pathB.cubicTo(435.631104f, 46.1651154f, 435.156006f, 45.8026352f, 434.957306f, 45.6384506f);
pathB.cubicTo(434.758636f, 45.4742737f, 431.460999f, 42.0105858f, 430.916901f, 41.4664841f);
pathB.cubicTo(430.372803f, 40.9223785f, 429.731628f, 41.3303604f, 429.375214f, 41.6996689f);
pathB.cubicTo(429.375214f, 41.6996689f, 420.610016f, 50.5190887f, 419.918793f, 51.1754227f);
pathB.cubicTo(419.22757f, 51.8321419f, 419.331482f, 52.2465706f, 419.884308f, 52.8339005f);
pathB.cubicTo(420.437134f, 53.4212303f, 425.170624f, 58.0167313f, 425.619568f, 58.4660416f);
pathB.cubicTo(426.068512f, 58.9153557f, 426.137878f, 59.2952805f, 426.137878f, 59.8136024f);
pathB.cubicTo(426.137878f, 60.3319244f, 426.103394f, 103.141357f, 426.103394f, 103.970215f);
pathB.cubicTo(426.103394f, 104.799835f, 426.310791f, 105.525177f, 427.17453f, 105.525177f);
pathB.cubicTo(428.038269f, 105.525177f, 440.131409f, 105.491432f, 440.85675f, 105.491432f);
pathB.cubicTo(441.582489f, 105.491432f, 441.996918f, 104.903343f, 441.996918f, 104.17762f);
pathB.cubicTo(441.996918f, 103.452271f, 441.996918f, 89.5622559f, 441.996918f, 87.8002701f);
pathB.cubicTo(441.996918f, 86.0382843f, 442.342712f, 86.2453079f, 442.826141f, 86.3491974f);
pathB.cubicTo(443.309601f, 86.4530945f, 445.832184f, 87.040802f, 448.700195f, 87.040802f);
pathB.cubicTo(466.87442f, 86.6949997f, 471.815338f, 69.4876556f, 471.228394f, 64.443222f);
pathB.close();
pathB.moveTo(448.873108f, 72.4595642f);
pathB.cubicTo(444.636658f, 72.4595642f, 441.202545f, 69.0254517f, 441.202545f, 64.7890244f);
pathB.cubicTo(441.202545f, 60.5525932f, 444.636658f, 57.1184845f, 448.873108f, 57.1184845f);
pathB.cubicTo(453.109528f, 57.1184845f, 456.54364f, 60.552597f, 456.54364f, 64.7890244f);
pathB.cubicTo(456.54364f, 69.0254517f, 453.109528f, 72.4595642f, 448.873108f, 72.4595642f);
pathB.close();
pathB.moveTo(520.242859f, 82.5275803f);
pathB.cubicTo(519.733276f, 81.6547394f, 494.845154f, 41.3379478f, 494.263123f, 40.3923073f);
pathB.cubicTo(493.681122f, 39.4462852f, 493.244324f, 39.5918846f, 492.807495f, 40.3195038f);
pathB.cubicTo(492.515167f, 40.8071136f, 467.238037f, 81.6790085f, 466.528625f, 82.8373566f);
pathB.cubicTo(465.819183f, 83.9945831f, 466.497528f, 84.2975311f, 466.904358f, 84.5792542f);
pathB.cubicTo(466.904358f, 84.5792542f, 477.399689f, 91.2036743f, 478.235748f, 91.6981049f);
pathB.cubicTo(479.071838f, 92.1929092f, 479.396393f, 91.9028473f, 479.669403f, 91.5104141f);
pathB.cubicTo(479.942383f, 91.1175995f, 484.106384f, 84.1549606f, 484.481781f, 83.5570221f);
pathB.cubicTo(484.857147f, 82.9602051f, 485.198792f, 82.8741379f, 485.966583f, 82.8741379f);
pathB.cubicTo(486.734406f, 82.8741379f, 499.635925f, 82.8062668f, 500.830322f, 82.8062668f);
pathB.cubicTo(502.024658f, 82.8062668f, 502.229797f, 83.4038391f, 502.400452f, 83.6427078f);
pathB.cubicTo(502.571045f, 83.8815842f, 506.615631f, 90.7414703f, 506.990997f, 91.2870865f);
pathB.cubicTo(507.651123f, 92.2475128f, 508.017395f, 92.1333847f, 508.672241f, 91.6962051f);
pathB.cubicTo(509.327423f, 91.2597809f, 519.442444f, 85.0744324f, 520.243286f, 84.4920349f);
pathB.cubicTo(521.042908f, 83.9100189f, 520.751709f, 83.4004211f, 520.242859f, 82.5275803f);
pathB.close();
pathB.moveTo(493.720551f, 72.4811783f);
pathB.cubicTo(491.265442f, 72.4811783f, 489.275574f, 70.4905548f, 489.275574f, 68.0362091f);
pathB.cubicTo(489.275574f, 65.5811005f, 491.265808f, 63.5912399f, 493.720551f, 63.5912399f);
pathB.cubicTo(496.175262f, 63.5912399f, 498.165527f, 65.5814819f, 498.165527f, 68.0362091f);
pathB.cubicTo(498.165894f, 70.4905548f, 496.175659f, 72.4811783f, 493.720551f, 72.4811783f);
pathB.close();
pathB.moveTo(526.561707f, 42.132679f);
pathB.cubicTo(526.876404f, 41.786499f, 527.537292f, 41.1881752f, 528.512878f, 42.1956215f);
pathB.cubicTo(529.488892f, 43.2030678f, 531.691833f, 45.2486725f, 531.881042f, 45.4693451f);
pathB.cubicTo(532.07019f, 45.6896439f, 532.762939f, 46.2875862f, 533.517883f, 45.7525826f);
pathB.cubicTo(534.272827f, 45.2172012f, 539.497681f, 41.9753265f, 545.384277f, 42.132679f);
pathB.cubicTo(546.203247f, 42.2270889f, 546.580566f, 42.5421791f, 546.612f, 43.8324814f);
pathB.cubicTo(546.643494f, 45.1227837f, 546.674561f, 55.1957283f, 546.612f, 56.2657356f);
pathB.cubicTo(546.549072f, 57.3361206f, 546.549072f, 57.7452469f, 544.975525f, 57.8396568f);
pathB.cubicTo(543.401978f, 57.9340706f, 538.742798f, 59.5709305f, 538.774658f, 64.6070251f);
pathB.cubicTo(538.806458f, 69.6431122f, 538.806091f, 86.2631226f, 538.806091f, 86.9554901f);
pathB.cubicTo(538.806091f, 87.6478424f, 538.585815f, 88.4349976f, 537.484314f, 88.4349976f);
pathB.cubicTo(536.382446f, 88.4349976f, 524.79895f, 88.4031372f, 524.04364f, 88.4031372f);
pathB.cubicTo(523.287964f, 88.4031372f, 522.910706f, 88.3402023f, 522.910706f, 86.8606949f);
pathB.cubicTo(522.910706f, 85.3823242f, 522.910706f, 60.6094704f, 522.910706f, 60.0740852f);
pathB.cubicTo(522.910706f, 59.5387039f, 522.81665f, 58.9092865f, 522.470093f, 58.6260452f);
pathB.cubicTo(522.123901f, 58.3428116f, 517.465088f, 53.621048f, 517.181885f, 53.3378105f);
pathB.cubicTo(516.898621f, 53.0545731f, 516.300659f, 52.330368f, 517.213318f, 51.5121231f);
pathB.cubicTo(518.125977f, 50.6942635f, 526.561707f, 42.132679f, 526.561707f, 42.132679f);
pathB.close();
pathB.moveTo(550.169006f, 43.7373123f);
pathB.cubicTo(550.169006f, 42.6358337f, 550.767334f, 42.2263336f, 551.49115f, 42.2263336f);
pathB.cubicTo(552.214966f, 42.2263336f, 552.781433f, 42.2263336f, 552.938843f, 42.2263336f);
pathB.cubicTo(553.096191f, 42.2263336f, 553.725586f, 41.8801537f, 553.662598f, 41.124855f);
pathB.cubicTo(553.600098f, 40.369175f, 553.662598f, 31.4614124f, 553.662598f, 30.8005257f);
pathB.cubicTo(553.662598f, 30.1396389f, 553.75708f, 29.1951332f, 555.110657f, 29.1951332f);
pathB.cubicTo(556.463928f, 29.1951332f, 566.536865f, 29.1951332f, 567.859009f, 29.1951332f);
pathB.cubicTo(569.180786f, 29.1951332f, 569.463623f, 30.1711082f, 569.463623f, 30.7690544f);
pathB.cubicTo(569.463623f, 31.3670006f, 569.463623f, 40.2122002f, 569.463623f, 41.0619125f);
pathB.cubicTo(569.463623f, 41.9116249f, 569.746887f, 42.3207474f, 570.659912f, 42.3207474f);
pathB.cubicTo(571.572571f, 42.3207474f, 575.507568f, 42.2263336f, 576.199951f, 42.2263336f);
pathB.cubicTo(576.892273f, 42.2263336f, 577.742004f, 42.5725098f, 577.742004f, 43.8631973f);
pathB.cubicTo(577.742004f, 45.1538811f, 577.742004f, 55.8877106f, 577.742004f, 56.5800705f);
pathB.cubicTo(577.742004f, 57.2724266f, 577.616455f, 58.0595779f, 576.45166f, 58.0595779f);
pathB.cubicTo(575.286865f, 58.0595779f, 570.943115f, 58.0595779f, 570.471069f, 58.0595779f);
pathB.cubicTo(569.999023f, 58.0595779f, 569.479919f, 57.8389015f, 569.479919f, 59.0510979f);
pathB.cubicTo(569.479919f, 60.2629128f, 569.479919f, 66.2124176f, 569.479919f, 67.1880188f);
pathB.cubicTo(569.479919f, 68.1636047f, 570.628418f, 73.6406708f, 575.917053f, 73.6406708f);
pathB.cubicTo(577.018921f, 73.6406708f, 577.742737f, 73.9242859f, 577.742737f, 75.4348907f);
pathB.cubicTo(577.742737f, 76.945488f, 577.742737f, 87.0813751f, 577.742737f, 87.8366699f);
pathB.cubicTo(577.742737f, 88.5915909f, 577.648315f, 89.4416809f, 576.295044f, 89.4416809f);
pathB.cubicTo(574.028809f, 89.6312714f, 553.978088f, 88.0254974f, 553.631897f, 65.928421f);
pathB.cubicTo(553.631897f, 65.928421f, 553.631897f, 59.6964378f, 553.631897f, 58.972229f);
pathB.cubicTo(553.631897f, 58.2484055f, 553.034363f, 58.0436554f, 552.703735f, 58.0436554f);
pathB.cubicTo(552.372681f, 58.0436554f, 551.522949f, 58.0436554f, 551.208252f, 58.0436554f);
pathB.cubicTo(550.893921f, 58.0436554f, 550.170105f, 58.0906677f, 550.170105f, 56.6115417f);
pathB.cubicTo(550.170105f, 55.1324081f, 550.169006f, 43.7373123f, 550.169006f, 43.7373123f);
pathB.close();
pathB.moveTo(611.203857f, 39.5509338f);
pathB.cubicTo(612.084961f, 39.5509338f, 612.620422f, 40.0544662f, 612.620422f, 40.8097687f);
pathB.cubicTo(612.620422f, 41.5650673f, 612.620422f, 53.1486092f, 612.620422f, 53.6839905f);
pathB.cubicTo(612.620422f, 54.2193718f, 612.651489f, 55.2264404f, 611.612976f, 55.2264404f);
pathB.cubicTo(610.574463f, 55.2264404f, 604.404663f, 55.9817429f, 604.404663f, 61.899395f);
pathB.cubicTo(604.404663f, 65.4878235f, 604.373169f, 66.6211548f, 604.373169f, 67.5338135f);
pathB.cubicTo(604.373169f, 68.0684357f, 604.414124f, 74.3353043f, 599.934631f, 80.4702148f);
pathB.cubicTo(596.765564f, 84.8109131f, 590.664368f, 89.3942871f, 582.150208f, 89.630127f);
pathB.cubicTo(580.028015f, 89.630127f, 580.230469f, 88.5286484f, 580.230469f, 88.1820908f);
pathB.cubicTo(580.230469f, 87.5845184f, 580.198975f, 75.8436279f, 580.198975f, 75.2142105f);
pathB.cubicTo(580.198975f, 74.5844116f, 580.608154f, 73.8605804f, 581.58374f, 73.8605804f);
pathB.cubicTo(583.40979f, 73.8605804f, 588.603271f, 72.7905731f, 588.603271f, 66.9043884f);
pathB.cubicTo(588.603271f, 61.3958588f, 588.603271f, 61.8679237f, 588.603271f, 61.0496788f);
pathB.cubicTo(588.603271f, 60.2314377f, 588.666565f, 54.3137856f, 593.230591f, 48.3961296f);
pathB.cubicTo(597.794617f, 42.4784775f, 604.814087f, 39.5509338f, 611.203857f, 39.5509338f);
pathB.close();
pathB.moveTo(635.22937f, 81.9311447f);
pathB.cubicTo(635.057617f, 81.6475296f, 634.869141f, 81.3851471f, 634.664429f, 81.1439972f);
pathB.lineTo(635.039001f, 81.0385895f);
pathB.cubicTo(635.447754f, 80.9233246f, 635.729858f, 80.5509796f, 635.729858f, 80.1263123f);
pathB.lineTo(635.729858f, 78.6149597f);
pathB.cubicTo(635.729858f, 78.3309631f, 635.602417f, 78.0617523f, 635.382935f, 77.8816452f);
pathB.cubicTo(635.162598f, 77.7015381f, 634.873291f, 77.6295013f, 634.595764f, 77.6852341f);
pathB.cubicTo(633.906799f, 77.8232498f, 633.194397f, 77.9017334f, 632.478149f, 77.9191818f);
pathB.cubicTo(631.714844f, 77.9373779f, 630.851501f, 77.9464874f, 629.911133f, 77.9464874f);
pathB.lineTo(615.131226f, 77.9464874f);
pathB.cubicTo(614.607605f, 77.9464874f, 614.18335f, 78.3707733f, 614.18335f, 78.8944016f);
pathB.lineTo(614.18335f, 81.1337585f);
pathB.cubicTo(614.18335f, 81.6573868f, 614.607605f, 82.0816803f, 615.131226f, 82.0816803f);
pathB.lineTo(619.693787f, 82.0816803f);
pathB.cubicTo(619.680908f, 82.1423492f, 619.669128f, 82.2026367f, 619.657776f, 82.2629166f);
pathB.cubicTo(619.571289f, 82.728157f, 619.529602f, 83.3200302f, 619.529602f, 84.0730591f);
pathB.cubicTo(619.529602f, 86.3196259f, 620.260254f, 88.1236954f, 621.701477f, 89.4348602f);
pathB.cubicTo(623.116516f, 90.7225037f, 625.163269f, 91.3754272f, 627.784058f, 91.3754272f);
pathB.cubicTo(630.525024f, 91.3754272f, 632.517944f, 90.8669662f, 633.876099f, 89.8208466f);
pathB.cubicTo(635.291565f, 88.7314987f, 636.009705f, 87.0798492f, 636.009705f, 84.9129181f);
pathB.cubicTo(636.010071f, 83.7905807f, 635.747314f, 82.7873077f, 635.22937f, 81.9311447f);
pathB.close();
pathB.moveTo(631.880554f, 85.7326736f);
pathB.cubicTo(631.690552f, 86.0545807f, 631.436157f, 86.307869f, 631.102844f, 86.5076904f);
pathB.cubicTo(630.736206f, 86.7279816f, 630.277039f, 86.8906479f, 629.737854f, 86.9903717f);
pathB.cubicTo(629.146362f, 87.1003265f, 628.488892f, 87.1564484f, 627.783997f, 87.1564484f);
pathB.cubicTo(626.159668f, 87.1564484f, 624.996399f, 86.8656235f, 624.327881f, 86.293457f);
pathB.cubicTo(623.693604f, 85.7489777f, 623.385315f, 84.995575f, 623.385315f, 83.9896393f);
pathB.cubicTo(623.385315f, 83.3655396f, 623.431519f, 82.8718567f, 623.522583f, 82.5215149f);
pathB.cubicTo(623.563477f, 82.3645325f, 623.616943f, 82.2189331f, 623.684448f, 82.0824356f);
pathB.lineTo(630.008179f, 82.0824356f);
pathB.cubicTo(630.758911f, 82.247757f, 631.311401f, 82.5256805f, 631.650757f, 82.9101562f);
pathB.cubicTo(631.990112f, 83.2942505f, 632.154663f, 83.8303986f, 632.154663f, 84.549675f);
pathB.cubicTo(632.154663f, 85.02742f, 632.062927f, 85.4251709f, 631.880554f, 85.7326736f);
pathB.close();
pathB.moveTo(635.667664f, 69.5979919f);
pathB.cubicTo(635.518311f, 69.0645142f, 635.325684f, 68.5818329f, 635.093994f, 68.1620941f);
pathB.cubicTo(634.940796f, 67.8856812f, 634.770569f, 67.6316376f, 634.586304f, 67.4026184f);
pathB.lineTo(635.054565f, 67.2619476f);
pathB.cubicTo(635.455322f, 67.1417542f, 635.729858f, 66.7724457f, 635.729858f, 66.3538437f);
pathB.lineTo(635.729858f, 64.7021942f);
pathB.cubicTo(635.729858f, 64.4045486f, 635.590332f, 64.1243439f, 635.3526f, 63.9449997f);
pathB.cubicTo(635.115234f, 63.7660294f, 634.807373f, 63.7087784f, 634.521057f, 63.7906761f);
pathB.cubicTo(634.059998f, 63.9226265f, 633.544678f, 64.0155258f, 632.988831f, 64.0659485f);
pathB.cubicTo(631.578735f, 64.1941071f, 629.921387f, 64.1565704f, 628.141968f, 64.0632935f);
pathB.cubicTo(627.067383f, 64.0068054f, 625.948853f, 63.9779854f, 624.81665f, 63.9779854f);
pathB.cubicTo(624.253601f, 63.9779854f, 623.681396f, 64.0359955f, 623.116089f, 64.1512604f);
pathB.cubicTo(622.479126f, 64.2809448f, 621.888367f, 64.5437012f, 621.35907f, 64.9315872f);
pathB.cubicTo(620.807007f, 65.3365402f, 620.360352f, 65.9159088f, 620.031189f, 66.6548996f);
pathB.cubicTo(619.712708f, 67.3722839f, 619.557983f, 68.2625656f, 619.557983f, 69.3769379f);
pathB.cubicTo(619.557983f, 70.4655304f, 619.669128f, 71.5268097f, 619.887878f, 72.5323639f);
pathB.cubicTo(620.11499f, 73.573555f, 620.473694f, 74.5040283f, 620.954468f, 75.2983856f);
pathB.cubicTo(621.196411f, 75.6976471f, 621.693481f, 75.861824f, 622.124939f, 75.6847534f);
pathB.lineTo(623.832336f, 74.9851913f);
pathB.cubicTo(624.086365f, 74.8809204f, 624.282776f, 74.6716156f, 624.370728f, 74.4111328f);
pathB.cubicTo(624.45874f, 74.15065f, 624.429138f, 73.8651276f, 624.290405f, 73.6281509f);
pathB.cubicTo(624.166382f, 73.416954f, 624.051147f, 73.1644287f, 623.947205f, 72.875885f);
pathB.cubicTo(623.836853f, 72.5702744f, 623.741333f, 72.2407837f, 623.663574f, 71.8968811f);
pathB.cubicTo(623.584717f, 71.549942f, 623.522217f, 71.2018585f, 623.477417f, 70.8621292f);
pathB.cubicTo(623.434937f, 70.5409775f, 623.41333f, 70.2391663f, 623.41333f, 69.9646454f);
pathB.cubicTo(623.41333f, 68.8229752f, 623.672729f, 68.4748993f, 623.75116f, 68.3960266f);
pathB.cubicTo(623.853577f, 68.2940369f, 624.20166f, 68.0574341f, 625.236755f, 68.0574341f);
pathB.cubicTo(625.39679f, 68.0574341f, 625.566284f, 68.0616074f, 625.744446f, 68.0695648f);
pathB.lineTo(625.744446f, 68.7331085f);
pathB.cubicTo(625.744446f, 69.8065338f, 625.819153f, 70.8048782f, 625.967041f, 71.70047f);
pathB.cubicTo(626.12323f, 72.6483841f, 626.392456f, 73.4825516f, 626.767456f, 74.1794586f);
pathB.cubicTo(627.173523f, 74.9328613f, 627.730957f, 75.5292969f, 628.424438f, 75.9528198f);
pathB.cubicTo(629.123596f, 76.3790054f, 629.981628f, 76.5951309f, 630.975464f, 76.5951309f);
pathB.cubicTo(631.722046f, 76.5951309f, 632.406799f, 76.4597626f, 633.009644f, 76.1924591f);
pathB.cubicTo(633.611816f, 75.9262848f, 634.136536f, 75.5543213f, 634.567688f, 75.0875626f);
pathB.cubicTo(634.998779f, 74.6223297f, 635.333191f, 74.0672302f, 635.561096f, 73.4370575f);
pathB.cubicTo(635.78479f, 72.8212891f, 635.898193f, 72.1520538f, 635.898193f, 71.4479446f);
pathB.cubicTo(635.898193f, 70.7688599f, 635.820496f, 70.1462708f, 635.667664f, 69.5979919f);
pathB.close();
pathB.moveTo(631.656494f, 71.9905396f);
pathB.cubicTo(631.416077f, 72.2574692f, 631.13739f, 72.3765259f, 630.751404f, 72.3765259f);
pathB.cubicTo(630.390015f, 72.3765259f, 630.239502f, 72.2536774f, 630.190247f, 72.2127228f);
pathB.cubicTo(630.002197f, 72.0587845f, 629.853149f, 71.8483429f, 629.735596f, 71.5704193f);
pathB.cubicTo(629.594177f, 71.2348557f, 629.494507f, 70.8310394f, 629.439453f, 70.3714905f);
pathB.cubicTo(629.379211f, 69.8607559f, 629.348511f, 69.3284073f, 629.348511f, 68.7892303f);
pathB.cubicTo(629.348511f, 68.5765228f, 629.351929f, 68.3603973f, 629.357971f, 68.1416168f);
pathB.lineTo(630.581177f, 68.1416168f);
pathB.cubicTo(630.702515f, 68.2026672f, 630.831787f, 68.2841797f, 630.967163f, 68.3857956f);
pathB.cubicTo(631.149902f, 68.523056f, 631.322815f, 68.703537f, 631.480225f, 68.922699f);
pathB.cubicTo(631.639038f, 69.1437531f, 631.77478f, 69.4186478f, 631.884399f, 69.7390442f);
pathB.cubicTo(631.989807f, 70.0488281f, 632.04364f, 70.4169922f, 632.04364f, 70.8329391f);
pathB.cubicTo(632.042847f, 71.3228302f, 631.916565f, 71.7012329f, 631.656494f, 71.9905396f);
pathB.close();
pathB.moveTo(622.689575f, 63.4953079f);
pathB.lineTo(620.72998f, 63.4953079f);
pathB.cubicTo(620.206421f, 63.4953079f, 619.782104f, 63.0710182f, 619.782104f, 62.54739f);
pathB.lineTo(619.782104f, 61.3116837f);
pathB.lineTo(617.958679f, 61.3116837f);
pathB.cubicTo(617.536255f, 61.3116837f, 617.164307f, 61.0318604f, 617.047913f, 60.6257744f);
pathB.lineTo(616.404114f, 58.3864136f);
pathB.cubicTo(616.321411f, 58.1001472f, 616.378662f, 57.7922592f, 616.557678f, 57.5541458f);
pathB.cubicTo(616.737061f, 57.3164062f, 617.017212f, 57.1764946f, 617.31488f, 57.1764946f);
pathB.lineTo(619.782104f, 57.1764946f);
pathB.lineTo(619.782104f, 54.3171997f);
pathB.cubicTo(619.782104f, 53.7935715f, 620.206421f, 53.3692818f, 620.730042f, 53.3692818f);
pathB.lineTo(622.689575f, 53.3692818f);
pathB.cubicTo(623.213196f, 53.3692818f, 623.637512f, 53.7935715f, 623.637512f, 54.3171997f);
pathB.lineTo(623.637512f, 57.1764946f);
pathB.lineTo(630.443176f, 57.1764946f);
pathB.cubicTo(631.548828f, 57.1764946f, 631.921936f, 57.0028381f, 632.009888f, 56.9493713f);
pathB.cubicTo(632.057617f, 56.9205589f, 632.154724f, 56.8621674f, 632.154724f, 56.5288773f);
pathB.cubicTo(632.154724f, 56.139473f, 632.116821f, 55.8179398f, 632.04248f, 55.5737572f);
pathB.cubicTo(631.949219f, 55.268528f, 631.822205f, 54.9193192f, 631.665588f, 54.5363579f);
pathB.cubicTo(631.563599f, 54.288002f, 631.572327f, 54.0085602f, 631.688354f, 53.7670288f);
pathB.cubicTo(631.804749f, 53.5251198f, 632.017456f, 53.3438797f, 632.274536f, 53.2680435f);
pathB.lineTo(633.9823f, 52.7641296f);
pathB.cubicTo(634.430115f, 52.6325607f, 634.90332f, 52.8437576f, 635.105042f, 53.2623596f);
pathB.cubicTo(635.375f, 53.8235245f, 635.59491f, 54.4404297f, 635.758301f, 55.0956268f);
pathB.cubicTo(635.925537f, 55.7656174f, 636.010498f, 56.4928589f, 636.010498f, 57.2565002f);
pathB.cubicTo(636.010498f, 58.7109833f, 635.553589f, 59.7790947f, 634.653076f, 60.4316406f);
pathB.cubicTo(633.835938f, 61.0235176f, 632.596069f, 61.311306f, 630.863647f, 61.311306f);
pathB.lineTo(623.637878f, 61.311306f);
pathB.lineTo(623.637878f, 62.5470123f);
pathB.cubicTo(623.637451f, 63.071022f, 623.213196f, 63.4953079f, 622.689575f, 63.4953079f);
pathB.close();
pathB.moveTo(635.667664f, 46.2246475f);
pathB.cubicTo(635.518311f, 45.6911545f, 635.325684f, 45.2084808f, 635.093994f, 44.7887421f);
pathB.cubicTo(634.940796f, 44.5123291f, 634.770569f, 44.2582855f, 634.586304f, 44.029274f);
pathB.lineTo(635.054565f, 43.8886032f);
pathB.cubicTo(635.455322f, 43.7684059f, 635.729858f, 43.3990974f, 635.729858f, 42.9804955f);
pathB.lineTo(635.729858f, 41.328846f);
pathB.cubicTo(635.729858f, 41.0312004f, 635.590332f, 40.7509995f, 635.3526f, 40.5716515f);
pathB.cubicTo(635.115234f, 40.3926849f, 634.807373f, 40.3354301f, 634.521057f, 40.4173279f);
pathB.cubicTo(634.059998f, 40.5492783f, 633.544678f, 40.6421738f, 632.988831f, 40.6926041f);
pathB.cubicTo(631.578735f, 40.8207626f, 629.921387f, 40.7836075f, 628.141968f, 40.689949f);
pathB.cubicTo(627.067383f, 40.6334534f, 625.948853f, 40.6046371f, 624.81665f, 40.6046371f);
pathB.cubicTo(624.253601f, 40.6046371f, 623.681396f, 40.6626511f, 623.116089f, 40.777916f);
pathB.cubicTo(622.479126f, 40.9075928f, 621.888367f, 41.1703568f, 621.35907f, 41.5582428f);
pathB.cubicTo(620.807007f, 41.963192f, 620.360352f, 42.5425606f, 620.031189f, 43.2815552f);
pathB.cubicTo(619.712708f, 43.9989395f, 619.557983f, 44.8892212f, 619.557983f, 46.0035934f);
pathB.cubicTo(619.557983f, 47.0921783f, 619.669128f, 48.1534653f, 619.887878f, 49.1590195f);
pathB.cubicTo(620.11499f, 50.2002106f, 620.473694f, 51.1306839f, 620.954468f, 51.9250374f);
pathB.cubicTo(621.196411f, 52.3243027f, 621.693481f, 52.4880981f, 622.124939f, 52.311409f);
pathB.lineTo(623.832336f, 51.6118431f);
pathB.cubicTo(624.086365f, 51.5075722f, 624.282776f, 51.2982712f, 624.370728f, 51.0377846f);
pathB.cubicTo(624.45874f, 50.777298f, 624.429138f, 50.4917831f, 624.290405f, 50.2548065f);
pathB.cubicTo(624.166382f, 50.0436096f, 624.051147f, 49.7910843f, 623.947205f, 49.5025406f);
pathB.cubicTo(623.836853f, 49.1969299f, 623.741333f, 48.8674355f, 623.663574f, 48.5235291f);
pathB.cubicTo(623.584717f, 48.1765938f, 623.522217f, 47.8285179f, 623.477417f, 47.4887848f);
pathB.cubicTo(623.434937f, 47.1676331f, 623.41333f, 46.8658142f, 623.41333f, 46.5912971f);
pathB.cubicTo(623.41333f, 45.4496269f, 623.672729f, 45.1015511f, 623.75116f, 45.0226822f);
pathB.cubicTo(623.853577f, 44.9206886f, 624.20166f, 44.6840897f, 625.236755f, 44.6840897f);
pathB.cubicTo(625.39679f, 44.6840897f, 625.566284f, 44.6882591f, 625.744446f, 44.6962204f);
pathB.lineTo(625.744446f, 45.3597641f);
pathB.cubicTo(625.744446f, 46.4331856f, 625.819153f, 47.43153f, 625.967041f, 48.3271217f);
pathB.cubicTo(626.12323f, 49.2750397f, 626.392456f, 50.1092072f, 626.767456f, 50.8061142f);
pathB.cubicTo(627.173523f, 51.5595169f, 627.730957f, 52.1559486f, 628.424438f, 52.5794754f);
pathB.cubicTo(629.123596f, 53.005661f, 629.981628f, 53.2217865f, 630.975464f, 53.2217865f);
pathB.cubicTo(631.722046f, 53.2217865f, 632.406799f, 53.086422f, 633.009644f, 52.8191147f);
pathB.cubicTo(633.611816f, 52.5529366f, 634.136536f, 52.1809769f, 634.567688f, 51.7142181f);
pathB.cubicTo(634.998779f, 51.2489815f, 635.333191f, 50.693882f, 635.561096f, 50.0637054f);
pathB.cubicTo(635.78479f, 49.4479408f, 635.898193f, 48.7787094f, 635.898193f, 48.0746002f);
pathB.cubicTo(635.898193f, 47.3958893f, 635.820496f, 46.7733002f, 635.667664f, 46.2246475f);
pathB.close();
pathB.moveTo(631.656494f, 48.6171875f);
pathB.cubicTo(631.416077f, 48.8841209f, 631.13739f, 49.0031815f, 630.751404f, 49.0031815f);
pathB.cubicTo(630.390015f, 49.0031815f, 630.239502f, 48.8803291f, 630.190247f, 48.8393784f);
pathB.cubicTo(630.002197f, 48.6854401f, 629.853149f, 48.4749985f, 629.735596f, 48.1970711f);
pathB.cubicTo(629.594177f, 47.8615112f, 629.494507f, 47.457695f, 629.439453f, 46.9981461f);
pathB.cubicTo(629.379211f, 46.4874115f, 629.348511f, 45.9550591f, 629.348511f, 45.4158859f);
pathB.cubicTo(629.348511f, 45.2031708f, 629.351929f, 44.9870453f, 629.357971f, 44.7682648f);
pathB.lineTo(630.581177f, 44.7682648f);
pathB.cubicTo(630.702515f, 44.8293152f, 630.831787f, 44.9108353f, 630.967163f, 45.0124512f);
pathB.cubicTo(631.149902f, 45.1497116f, 631.322815f, 45.3301926f, 631.480225f, 45.5493507f);
pathB.cubicTo(631.639038f, 45.7704048f, 631.77478f, 46.0453033f, 631.884399f, 46.3656998f);
pathB.cubicTo(631.989807f, 46.6754761f, 632.04364f, 47.0436478f, 632.04364f, 47.4595947f);
pathB.cubicTo(632.042847f, 47.949852f, 631.916565f, 48.3282623f, 631.656494f, 48.6171875f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_seopack_blogspot_com_2153(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(999.892212f, 246);
path.lineTo(927.340759f, 245.505722f);
path.quadTo(928.068054f, 246, 929, 246);
path.lineTo(999.892212f, 246);
path.close();
path.moveTo(927.340759f, 245.505722f);
path.lineTo(926.5f, 245.5f);
path.lineTo(925.17157f, 246.82843f);
path.quadTo(926.34314f, 248, 928, 248);
path.lineTo(1000, 248);
path.lineTo(1000, 246);
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(924, 248);
pathB.lineTo(924, 245.472672f);
pathB.lineTo(1143, 247);
pathB.lineTo(1143, 248);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_lokado_de_173(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(1000, 896.991394f);
path.quadTo(999.789917f, 896.718872f, 999.535522f, 896.464478f);
path.quadTo(998.071045f, 895, 996, 895);
path.lineTo(956, 895);
path.quadTo(951.857849f, 895, 948.928955f, 897.928955f);
path.quadTo(946, 900.857849f, 946, 905);
path.lineTo(946, 906);
path.quadTo(946, 910.142151f, 948.928955f, 913.071045f);
path.quadTo(951.857849f, 916, 956, 916);
path.lineTo(996, 916);
path.quadTo(998.071045f, 916, 999.535522f, 914.535522f);
path.quadTo(999.789917f, 914.281128f, 1000, 914.008606f);
path.lineTo(1000, 896.991394f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(946, 906);
pathB.lineTo(946, 905);
pathB.quadTo(946, 901.272095f, 948.928955f, 898.636047f);
pathB.quadTo(951.857849f, 896, 956, 896);
pathB.lineTo(996, 896);
pathB.quadTo(998.071045f, 896, 999.535522f, 897.17157f);
pathB.quadTo(1001, 898.34314f, 1001, 900);
pathB.lineTo(1001, 911);
pathB.quadTo(1001, 913.071045f, 999.535522f, 914.535522f);
pathB.quadTo(998.071045f, 916, 996, 916);
pathB.lineTo(956, 916);
pathB.quadTo(951.857849f, 916, 948.928955f, 913.071045f);
pathB.quadTo(946, 910.142151f, 946, 906);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kDifference_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_wartepop_blogspot_com_br_6(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(90.9763107f, 153.309662f);
path.quadTo(91.9526215f, 152.333344f, 93.3333359f, 152.333344f);
path.lineTo(124.666664f, 152.333344f);
path.quadTo(126.047379f, 152.333344f, 127.023689f, 153.309662f);
path.quadTo(128, 154.285965f, 128, 155.666672f);
path.lineTo(128, 163.666672f);
path.lineTo(90, 163.666672f);
path.lineTo(90, 155.666672f);
path.quadTo(90, 154.285965f, 90.9763107f, 153.309662f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(90, 163.666672f);
pathB.lineTo(90, 155.666672f);
pathB.quadTo(90, 154.285965f, 90.9763107f, 153.309662f);
pathB.quadTo(91.9526215f, 152.333344f, 93.3333359f, 152.333344f);
pathB.lineTo(124.666672f, 152.333344f);
pathB.quadTo(125.909309f, 152.333344f, 126.787994f, 153.309662f);
pathB.quadTo(127.666672f, 154.285965f, 127.666672f, 155.666672f);
pathB.lineTo(127.666672f, 163.666672f);
pathB.lineTo(127.666672f, 163.666672f);
pathB.lineTo(127.666672f, 163.666672f);
pathB.lineTo(90, 163.666672f);
pathB.lineTo(90, 163.666672f);
pathB.lineTo(90, 163.666672f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kDifference_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
static void skpwww_wartepop_blogspot_com_br_6a(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(90.9763107f, 153.309662f);
path.quadTo(91.9526215f, 152.333344f, 93.3333359f, 152.333344f);
path.lineTo(124.666672f, 152.333344f);
path.quadTo(126.047379f, 152.333344f, 127.023689f, 153.309662f);
path.quadTo(128, 154.285965f, 128, 155.666672f);
path.lineTo(128, 163.666672f);
path.lineTo(90, 163.666672f);
path.lineTo(90, 155.666672f);
path.quadTo(90, 154.285965f, 90.9763107f, 153.309662f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(90, 163.666672f);
pathB.lineTo(90, 155.666672f);
pathB.quadTo(90, 154.285965f, 90.9763107f, 153.309662f);
pathB.quadTo(91.9526215f, 152.333344f, 93.3333359f, 152.333344f);
pathB.lineTo(124.666672f, 152.333344f);
pathB.quadTo(125.909309f, 152.333344f, 126.787994f, 153.309662f);
pathB.quadTo(127.666672f, 154.285965f, 127.666672f, 155.666672f);
pathB.lineTo(127.666672f, 163.666672f);
pathB.lineTo(127.666672f, 163.666672f);
pathB.lineTo(127.666672f, 163.666672f);
pathB.lineTo(90, 163.666672f);
pathB.lineTo(90, 163.666672f);
pathB.lineTo(90, 163.666672f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kDifference_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
static void skpwww_odia_com_br_26(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(360.740479f, 741.040771f);
path.quadTo(360.378967f, 741, 360, 741);
path.quadTo(359.159821f, 741, 358.403076f, 741.200745f);
path.quadTo(357.649658f, 741.415833f, 356.92746f, 741.846436f);
path.quadTo(356.600769f, 742.041199f, 356.310211f, 742.262268f);
path.quadTo(356.025513f, 742.489197f, 355.757355f, 742.757385f);
path.quadTo(355.16394f, 743.350769f, 354.770874f, 744.027283f);
path.quadTo(354.389618f, 744.71283f, 354.183258f, 745.528564f);
path.quadTo(354.090027f, 745.897095f, 354.040833f, 746.259277f);
path.quadTo(354, 746.621216f, 354, 747);
path.quadTo(354, 747.839844f, 354.200653f, 748.596497f);
path.quadTo(354.415771f, 749.35022f, 354.846466f, 750.072632f);
path.quadTo(355.040741f, 750.398438f, 355.261444f, 750.688721f);
path.quadTo(355.488861f, 750.974121f, 355.757355f, 751.242615f);
path.quadTo(356.352142f, 751.837402f, 357.030304f, 752.230896f);
path.quadTo(357.714539f, 752.610901f, 358.528564f, 752.816833f);
path.quadTo(358.895294f, 752.909607f, 359.25528f, 752.95874f);
path.quadTo(359.618896f, 753, 360, 753);
path.quadTo(360.842285f, 753, 361.600952f, 752.798157f);
path.quadTo(362.352386f, 752.583008f, 363.072601f, 752.153625f);
path.quadTo(363.397339f, 751.960022f, 363.686829f, 751.740051f);
path.quadTo(363.973297f, 751.511963f, 364.242645f, 751.242615f);
path.quadTo(364.837799f, 750.647461f, 365.231354f, 749.968933f);
path.quadTo(365.610992f, 749.285034f, 365.816803f, 748.471497f);
path.quadTo(365.909668f, 748.104431f, 365.958832f, 747.743713f);
path.quadTo(366, 747.380371f, 366, 747);
path.quadTo(366, 746.158997f, 365.798767f, 745.401367f);
path.quadTo(365.583618f, 744.648682f, 365.153595f, 743.927429f);
path.quadTo(364.959442f, 743.601807f, 364.738678f, 743.311462f);
path.quadTo(364.511108f, 743.025818f, 364.242645f, 742.757385f);
path.quadTo(363.649017f, 742.163757f, 362.972168f, 741.770569f);
path.quadTo(362.286835f, 741.389526f, 361.471497f, 741.183289f);
path.quadTo(361.102509f, 741.089966f, 360.740479f, 741.040771f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(355.654724f, 739.711792f);
pathB.lineTo(367.288269f, 742.654724f);
pathB.lineTo(364.345337f, 754.288269f);
pathB.lineTo(352.711792f, 751.345337f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_evolvehq_com_210(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(172, 972);
path.quadTo(170.757355f, 972, 169.878677f, 972.878662f);
path.quadTo(169, 973.757385f, 169, 975);
path.lineTo(169, 1171);
path.quadTo(169, 1172.24268f, 169.878677f, 1173.12134f);
path.quadTo(170.757355f, 1174, 172, 1174);
path.lineTo(308, 1174);
path.quadTo(309.242645f, 1174, 310.121307f, 1173.12134f);
path.quadTo(310.337311f, 1172.9054f, 310.5f, 1172.66772f);
path.lineTo(310.5f, 973.332336f);
path.quadTo(310.337219f, 973.094604f, 310.121307f, 972.878662f);
path.quadTo(309.242645f, 972, 308, 972);
path.lineTo(172, 972);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(170, 1171);
pathB.lineTo(170, 975);
pathB.quadTo(170, 974.17157f, 170.585785f, 973.585815f);
pathB.quadTo(171.17157f, 973, 172, 973);
pathB.lineTo(308, 973);
pathB.quadTo(309.242645f, 973, 310.121307f, 973.585815f);
pathB.quadTo(311, 974.17157f, 311, 975);
pathB.lineTo(311, 1171);
pathB.quadTo(311, 1172.24268f, 310.121307f, 1173.12134f);
pathB.quadTo(309.242645f, 1174, 308, 1174);
pathB.lineTo(172, 1174);
pathB.quadTo(171.17157f, 1174, 170.585785f, 1173.12134f);
pathB.quadTo(170, 1172.24268f, 170, 1171);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kDifference_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_catingueiraonline_com_352(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(443, 8292);
path.lineTo(443, 8140);
path.lineTo(444, 8140);
path.lineTo(444.01001f, 8292);
path.lineTo(443, 8292);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(443, 8140);
pathB.lineTo(444.01001f, 8140);
pathB.lineTo(444, 8292);
pathB.lineTo(443, 8292);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_galaxystwo_com_4(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(10105, 2510);
path.lineTo(10123, 2509.98999f);
path.lineTo(10123, 2511);
path.lineTo(10105, 2511);
path.lineTo(10105, 2510);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(10105, 2511);
pathB.lineTo(10105, 2509.98999f);
pathB.lineTo(10123, 2510);
pathB.lineTo(10123, 2511);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_thaienews_blogspot_com_36(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(429.994995f, 6268);
path.lineTo(430, 2187);
path.lineTo(430.5f, 2187);
path.lineTo(430.5f, 6268);
path.lineTo(429.994995f, 6268);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(430.5f, 2187);
pathB.lineTo(429.994995f, 2187);
pathB.lineTo(430, 6268);
pathB.lineTo(430.5f, 6268);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_fashionscandal_com_94(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(25.9107456f, 272.577423f);
path.quadTo(26.1548233f, 272.333344f, 26.5000019f, 272.333344f);
path.lineTo(131.166672f, 272.333344f);
path.quadTo(131.511841f, 272.333344f, 131.75592f, 272.577423f);
path.quadTo(132, 272.821503f, 132, 273.166687f);
path.lineTo(132, 417.166656f);
path.quadTo(132, 417.511841f, 131.75592f, 417.75592f);
path.quadTo(131.511841f, 418, 131.166672f, 418);
path.lineTo(26.5000019f, 418);
path.quadTo(26.1548233f, 418, 25.9107456f, 417.75592f);
path.quadTo(25.6666679f, 417.511841f, 25.6666679f, 417.166656f);
path.lineTo(25.6666679f, 273.166687f);
path.quadTo(25.6666679f, 272.821503f, 25.9107456f, 272.577423f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(25.833334f, 417.166656f);
pathB.lineTo(25.833334f, 273.166656f);
pathB.quadTo(25.833334f, 272.890533f, 26.0285969f, 272.695251f);
pathB.quadTo(26.2238579f, 272.5f, 26.5f, 272.5f);
pathB.lineTo(131.166672f, 272.5f);
pathB.quadTo(131.442825f, 272.5f, 131.638077f, 272.695251f);
pathB.quadTo(131.833344f, 272.890533f, 131.833344f, 273.166656f);
pathB.lineTo(131.833344f, 417.166656f);
pathB.quadTo(131.833344f, 417.511841f, 131.638077f, 417.75592f);
pathB.quadTo(131.442825f, 418, 131.166672f, 418);
pathB.lineTo(26.5f, 418);
pathB.quadTo(26.2238579f, 418, 26.0285969f, 417.75592f);
pathB.quadTo(25.833334f, 417.511841f, 25.833334f, 417.166656f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kDifference_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_kenlevine_blogspot_com_28(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(276, 9506);
path.lineTo(276, 7531);
path.lineTo(277, 7531);
path.lineTo(277.01001f, 9506);
path.lineTo(276, 9506);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(276, 7531);
pathB.lineTo(277.01001f, 7531);
pathB.lineTo(277, 9506);
pathB.lineTo(276, 9506);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
static void skpwww_defense_studies_blogspot_com_64(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
path.moveTo(276, 9600);
path.lineTo(276, 7703);
path.lineTo(277, 7703);
path.lineTo(277.01001f, 9600);
path.lineTo(276, 9600);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
pathB.moveTo(276, 7703);
pathB.lineTo(277.01001f, 7703);
pathB.lineTo(277, 9600);
pathB.lineTo(276, 9600);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
}
fix cubic/line intersection; add skp tests BUG=skia:2488 TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/252243003 git-svn-id: http://skia.googlecode.com/svn/trunk@14458 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-30 13:37:48 +00:00
static void skpwww_uniquefx_net_442(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix cubic/line intersection; add skp tests BUG=skia:2488 TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/252243003 git-svn-id: http://skia.googlecode.com/svn/trunk@14458 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-30 13:37:48 +00:00
path.moveTo(960, 306);
path.lineTo(960, 305);
path.lineTo(1000, 305);
path.lineTo(1000, 306.708527f);
path.lineTo(960, 306);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix cubic/line intersection; add skp tests BUG=skia:2488 TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/252243003 git-svn-id: http://skia.googlecode.com/svn/trunk@14458 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-30 13:37:48 +00:00
pathB.moveTo(960, 305);
pathB.lineTo(958.997253f, 306.002747f);
pathB.lineTo(1017, 307);
pathB.lineTo(1019, 305);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix cubic/line intersection; add skp tests BUG=skia:2488 TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/252243003 git-svn-id: http://skia.googlecode.com/svn/trunk@14458 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-30 13:37:48 +00:00
}
static void skpwww_kitcheninspirations_wordpress_com_32(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix cubic/line intersection; add skp tests BUG=skia:2488 TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/252243003 git-svn-id: http://skia.googlecode.com/svn/trunk@14458 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-30 13:37:48 +00:00
path.moveTo(47.1666679f, 19651.334f);
path.lineTo(65.8333359f, 19651.332f);
path.lineTo(65.8333359f, 19651.5f);
path.lineTo(47.1666679f, 19651.5f);
path.lineTo(47.1666679f, 19651.334f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix cubic/line intersection; add skp tests BUG=skia:2488 TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/252243003 git-svn-id: http://skia.googlecode.com/svn/trunk@14458 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-30 13:37:48 +00:00
pathB.moveTo(47.1666679f, 19651.5f);
pathB.lineTo(47.1666679f, 19651.332f);
pathB.lineTo(65.8333359f, 19651.334f);
pathB.lineTo(65.8333359f, 19651.5f);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix cubic/line intersection; add skp tests BUG=skia:2488 TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/252243003 git-svn-id: http://skia.googlecode.com/svn/trunk@14458 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-30 13:37:48 +00:00
}
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
static void skpwww_educationalcraft_com_4(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(941, 1494);
path.lineTo(941, 1464);
path.lineTo(985, 1464);
path.lineTo(985, 1494);
path.lineTo(941, 1494);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(979.211975f, 1480.45496f);
pathB.cubicTo(979.211975f, 1480.45496f, 976.348999f, 1479.68506f, 977.495972f, 1475.59497f);
pathB.cubicTo(977.497009f, 1475.59497f, 981.072021f, 1477.88501f, 979.211975f, 1480.45496f);
pathB.close();
pathB.moveTo(977.854004f, 1484.453f);
pathB.cubicTo(977.854004f, 1484.453f, 975.265991f, 1483.26099f, 976.713989f, 1479.35205f);
pathB.cubicTo(976.713989f, 1479.35303f, 979.84198f, 1482.23499f, 977.854004f, 1484.453f);
pathB.close();
pathB.moveTo(980.226013f, 1476.229f);
pathB.cubicTo(980.226013f, 1476.229f, 977.078003f, 1476.349f, 977.234985f, 1471.97095f);
pathB.cubicTo(977.234985f, 1471.97095f, 980.666992f, 1473.12903f, 980.226013f, 1476.229f);
pathB.close();
pathB.moveTo(984.546021f, 1478.31494f);
pathB.cubicTo(984.546021f, 1478.31494f, 983.187988f, 1481.93396f, 980.026001f, 1481.276f);
pathB.cubicTo(980.026978f, 1481.276f, 979.554993f, 1478.38904f, 984.546021f, 1478.31494f);
pathB.close();
pathB.moveTo(978.989014f, 1484.198f);
pathB.cubicTo(978.989014f, 1484.198f, 979.094971f, 1481.33496f, 983.786011f, 1481.823f);
pathB.cubicTo(983.786011f, 1481.823f, 982.070007f, 1485.49805f, 978.989014f, 1484.198f);
pathB.close();
pathB.moveTo(976.393005f, 1486.86804f);
pathB.cubicTo(976.393005f, 1486.86804f, 976.719971f, 1484.06494f, 981.679016f, 1485.37f);
pathB.cubicTo(981.679016f, 1485.37f, 979.169983f, 1488.40796f, 976.393005f, 1486.86804f);
pathB.close();
pathB.moveTo(969.156982f, 1490.40002f);
pathB.cubicTo(969.156982f, 1490.40002f, 971.478027f, 1488.23596f, 974.869995f, 1491.21399f);
pathB.cubicTo(974.869995f, 1491.21497f, 970.828003f, 1493.026f, 969.156982f, 1490.40002f);
pathB.close();
pathB.moveTo(972.825012f, 1483.93701f);
pathB.cubicTo(972.825012f, 1483.93701f, 973.971985f, 1487.98401f, 971.161987f, 1488.94604f);
pathB.cubicTo(971.161987f, 1488.94495f, 969.278015f, 1486.37097f, 972.825012f, 1483.93701f);
pathB.close();
pathB.moveTo(965.60199f, 1489.98499f);
pathB.cubicTo(965.60199f, 1489.98499f, 964.879028f, 1487.19202f, 969.864014f, 1486.75f);
pathB.cubicTo(969.864014f, 1486.75f, 968.749023f, 1490.672f, 965.60199f, 1489.98499f);
pathB.close();
pathB.moveTo(970.666992f, 1492.81604f);
pathB.cubicTo(970.666992f, 1492.81604f, 967.327026f, 1494.49695f, 964.999023f, 1491.56299f);
pathB.cubicTo(964.999023f, 1491.56299f, 967.304016f, 1489.43896f, 970.666992f, 1492.81604f);
pathB.close();
pathB.moveTo(968.343994f, 1481.53796f);
pathB.cubicTo(971.573975f, 1479.94995f, 971.687988f, 1476.78601f, 971.687988f, 1476.78601f);
pathB.lineTo(971.393982f, 1466.83398f);
pathB.lineTo(954.960999f, 1466.83398f);
pathB.lineTo(954.666016f, 1476.78601f);
pathB.cubicTo(954.666016f, 1476.78601f, 954.780029f, 1479.94995f, 958.008972f, 1481.53796f);
pathB.cubicTo(960.781006f, 1482.90295f, 962.166992f, 1484.77698f, 962.166992f, 1484.77698f);
pathB.cubicTo(962.166992f, 1484.77698f, 962.747986f, 1485.70105f, 963.177979f, 1485.70105f);
pathB.cubicTo(963.606995f, 1485.70105f, 964.185974f, 1484.77698f, 964.185974f, 1484.77698f);
pathB.cubicTo(964.185974f, 1484.77698f, 965.573975f, 1482.90295f, 968.343994f, 1481.53796f);
pathB.close();
pathB.moveTo(963.215027f, 1486.67004f);
pathB.cubicTo(962.744995f, 1486.67004f, 962.106995f, 1485.65405f, 962.106995f, 1485.65405f);
pathB.cubicTo(962.106995f, 1485.65405f, 960.585022f, 1483.59595f, 957.539001f, 1482.09705f);
pathB.cubicTo(953.991028f, 1480.35205f, 953.867004f, 1476.87598f, 953.867004f, 1476.87598f);
pathB.lineTo(954.190002f, 1465.94397f);
pathB.lineTo(972.23999f, 1465.94397f);
pathB.lineTo(972.565002f, 1476.87695f);
pathB.cubicTo(972.565002f, 1476.87695f, 972.440979f, 1480.35303f, 968.891968f, 1482.09802f);
pathB.cubicTo(965.846008f, 1483.59705f, 964.325012f, 1485.65503f, 964.325012f, 1485.65503f);
pathB.cubicTo(964.325012f, 1485.65503f, 963.687012f, 1486.67004f, 963.215027f, 1486.67004f);
pathB.close();
pathB.moveTo(960.68103f, 1489.98499f);
pathB.cubicTo(957.533997f, 1490.672f, 956.417969f, 1486.75f, 956.417969f, 1486.75f);
pathB.cubicTo(961.403015f, 1487.19202f, 960.68103f, 1489.98499f, 960.68103f, 1489.98499f);
pathB.close();
pathB.moveTo(963.143005f, 1489.59802f);
pathB.cubicTo(963.763f, 1489.59802f, 964.265015f, 1490.09998f, 964.265015f, 1490.72095f);
pathB.cubicTo(964.265015f, 1491.34204f, 963.763f, 1491.84399f, 963.143005f, 1491.84399f);
pathB.cubicTo(962.521973f, 1491.84399f, 962.02002f, 1491.34204f, 962.02002f, 1490.72095f);
pathB.cubicTo(962.02002f, 1490.09998f, 962.521973f, 1489.59802f, 963.143005f, 1489.59802f);
pathB.close();
pathB.moveTo(961.283997f, 1491.56299f);
pathB.cubicTo(958.953979f, 1494.49695f, 955.61499f, 1492.81604f, 955.61499f, 1492.81604f);
pathB.cubicTo(958.97699f, 1489.43896f, 961.283997f, 1491.56299f, 961.283997f, 1491.56299f);
pathB.close();
pathB.moveTo(957.127014f, 1490.40002f);
pathB.cubicTo(955.455017f, 1493.026f, 951.414001f, 1491.21399f, 951.414001f, 1491.21399f);
pathB.cubicTo(954.802979f, 1488.23596f, 957.127014f, 1490.40002f, 957.127014f, 1490.40002f);
pathB.close();
pathB.moveTo(949.890991f, 1486.86804f);
pathB.cubicTo(947.112976f, 1488.40796f, 944.604004f, 1485.37f, 944.604004f, 1485.37f);
pathB.cubicTo(949.562012f, 1484.06494f, 949.890991f, 1486.86804f, 949.890991f, 1486.86804f);
pathB.close();
pathB.moveTo(947.070984f, 1480.45496f);
pathB.cubicTo(945.211975f, 1477.88501f, 948.786011f, 1475.59497f, 948.786011f, 1475.59497f);
pathB.cubicTo(949.934021f, 1479.68506f, 947.070984f, 1480.45496f, 947.070984f, 1480.45496f);
pathB.close();
pathB.moveTo(946.054016f, 1476.229f);
pathB.cubicTo(945.61499f, 1473.12903f, 949.046997f, 1471.97095f, 949.046997f, 1471.97095f);
pathB.cubicTo(949.205994f, 1476.349f, 946.054016f, 1476.229f, 946.054016f, 1476.229f);
pathB.close();
pathB.moveTo(948.427002f, 1484.453f);
pathB.cubicTo(946.440002f, 1482.23499f, 949.567993f, 1479.35205f, 949.567993f, 1479.35205f);
pathB.cubicTo(951.015991f, 1483.26099f, 948.427002f, 1484.453f, 948.427002f, 1484.453f);
pathB.close();
pathB.moveTo(947.294006f, 1484.198f);
pathB.cubicTo(944.210999f, 1485.49805f, 942.495972f, 1481.823f, 942.495972f, 1481.823f);
pathB.cubicTo(947.187988f, 1481.33496f, 947.294006f, 1484.198f, 947.294006f, 1484.198f);
pathB.close();
pathB.moveTo(946.255005f, 1481.276f);
pathB.cubicTo(943.094971f, 1481.93396f, 941.736023f, 1478.31494f, 941.736023f, 1478.31494f);
pathB.cubicTo(946.728027f, 1478.38904f, 946.255005f, 1481.276f, 946.255005f, 1481.276f);
pathB.close();
pathB.moveTo(945.312988f, 1478.18005f);
pathB.cubicTo(942.052979f, 1477.80103f, 942.651001f, 1473.87805f, 942.651001f, 1473.87805f);
pathB.cubicTo(946.562988f, 1475.66199f, 945.312988f, 1478.18005f, 945.312988f, 1478.18005f);
pathB.close();
pathB.moveTo(945.382019f, 1474.328f);
pathB.cubicTo(942.924011f, 1472.729f, 944.492004f, 1469.48706f, 944.492004f, 1469.48706f);
pathB.cubicTo(947.388977f, 1471.95703f, 945.382019f, 1474.328f, 945.382019f, 1474.328f);
pathB.close();
pathB.moveTo(946.797974f, 1470.27405f);
pathB.cubicTo(944.664978f, 1467.90198f, 947.083984f, 1465.50598f, 947.083984f, 1465.50598f);
pathB.cubicTo(949.145996f, 1468.82605f, 946.797974f, 1470.27405f, 946.797974f, 1470.27405f);
pathB.close();
pathB.moveTo(947.392029f, 1471.64197f);
pathB.cubicTo(947.624023f, 1468.56299f, 951.361023f, 1468.29199f, 951.361023f, 1468.29199f);
pathB.cubicTo(950.554016f, 1471.98499f, 947.392029f, 1471.64197f, 947.392029f, 1471.64197f);
pathB.close();
pathB.moveTo(948.64801f, 1468.15002f);
pathB.cubicTo(948.638977f, 1465.22095f, 952.265991f, 1464.46399f, 952.265991f, 1464.46399f);
pathB.cubicTo(951.672974f, 1468.53101f, 948.64801f, 1468.15002f, 948.64801f, 1468.15002f);
pathB.close();
pathB.moveTo(951.176025f, 1486.97803f);
pathB.cubicTo(948.963013f, 1484.62f, 951.361023f, 1481.77698f, 951.361023f, 1481.77698f);
pathB.cubicTo(953.734985f, 1485.48596f, 951.176025f, 1486.97803f, 951.176025f, 1486.97803f);
pathB.close();
pathB.moveTo(947.51001f, 1488.53101f);
pathB.cubicTo(947.51001f, 1488.53101f, 951.596985f, 1486.32202f, 953.234009f, 1489.08997f);
pathB.cubicTo(953.234009f, 1489.08997f, 951.158997f, 1491.03601f, 947.51001f, 1488.53101f);
pathB.close();
pathB.moveTo(955.120972f, 1488.94495f);
pathB.cubicTo(952.309021f, 1487.98303f, 953.458984f, 1483.93604f, 953.458984f, 1483.93604f);
pathB.cubicTo(957.004028f, 1486.37097f, 955.120972f, 1488.94495f, 955.120972f, 1488.94495f);
pathB.close();
pathB.moveTo(978.770996f, 1488.53101f);
pathB.cubicTo(975.122986f, 1491.03601f, 973.047974f, 1489.08997f, 973.047974f, 1489.08997f);
pathB.cubicTo(974.684998f, 1486.32202f, 978.770996f, 1488.53101f, 978.770996f, 1488.53101f);
pathB.close();
pathB.moveTo(975.106995f, 1486.97803f);
pathB.cubicTo(975.106995f, 1486.97803f, 972.546997f, 1485.48706f, 974.919983f, 1481.77698f);
pathB.cubicTo(974.919983f, 1481.776f, 977.31897f, 1484.61902f, 975.106995f, 1486.97803f);
pathB.close();
pathB.moveTo(974.016968f, 1464.46399f);
pathB.cubicTo(974.016968f, 1464.46399f, 977.643982f, 1465.22095f, 977.633972f, 1468.15002f);
pathB.cubicTo(977.633972f, 1468.15002f, 974.611023f, 1468.53101f, 974.016968f, 1464.46399f);
pathB.close();
pathB.moveTo(974.919983f, 1468.29199f);
pathB.cubicTo(974.919983f, 1468.29199f, 978.658997f, 1468.56299f, 978.890015f, 1471.64197f);
pathB.cubicTo(978.890015f, 1471.64197f, 975.72699f, 1471.98499f, 974.919983f, 1468.29199f);
pathB.close();
pathB.moveTo(979.197998f, 1465.50598f);
pathB.cubicTo(979.197998f, 1465.50598f, 981.619019f, 1467.90198f, 979.481995f, 1470.27405f);
pathB.cubicTo(979.481995f, 1470.27405f, 977.138f, 1468.82605f, 979.197998f, 1465.50598f);
pathB.close();
pathB.moveTo(980.900024f, 1474.328f);
pathB.cubicTo(980.900024f, 1474.328f, 978.893005f, 1471.95703f, 981.791016f, 1469.48706f);
pathB.cubicTo(981.791016f, 1469.48596f, 983.358032f, 1472.729f, 980.900024f, 1474.328f);
pathB.close();
pathB.moveTo(980.968994f, 1478.18005f);
pathB.cubicTo(980.968994f, 1478.18005f, 979.718018f, 1475.66199f, 983.632019f, 1473.87805f);
pathB.cubicTo(983.632019f, 1473.87805f, 984.229004f, 1477.80103f, 980.968994f, 1478.18005f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_narayana_publishers_com_194(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(1083.34314f, 445.65686f);
path.quadTo(1081, 443.313721f, 1081, 440);
path.lineTo(1257, 440);
path.quadTo(1257, 443.313721f, 1254.65686f, 445.65686f);
path.quadTo(1252.31372f, 448, 1249, 448);
path.lineTo(1089, 448);
path.quadTo(1085.68628f, 448, 1083.34314f, 445.65686f);
path.close();
path.moveTo(1083, 441);
path.lineTo(1255, 441);
path.quadTo(1255, 443.071075f, 1253.53552f, 444.535522f);
path.quadTo(1252.07104f, 446, 1250, 446);
path.lineTo(1088, 446);
path.quadTo(1085.92896f, 446, 1084.46448f, 444.535522f);
path.quadTo(1083, 443.071075f, 1083, 441);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(1081, 440);
pathB.lineTo(1082, 440);
pathB.lineTo(1090.01001f, 448);
pathB.lineTo(1081, 448);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_cooksnaps_com_17(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(170.340179f, 176);
path.lineTo(166, 176);
path.quadTo(161.964188f, 176, 158.299957f, 176.896912f);
path.quadTo(154.678162f, 177.952271f, 151.183014f, 179.9702f);
path.lineTo(150.316986f, 180.4702f);
path.quadTo(146.175812f, 182.861099f, 143.115921f, 186.081696f);
path.quadTo(140.693939f, 188.70134f, 138.99472f, 191.620407f);
path.quadTo(137.316833f, 194.550888f, 136.259338f, 197.957367f);
path.quadTo(135, 202.217865f, 135, 207);
path.lineTo(135, 208);
path.quadTo(135, 212.035751f, 135.896912f, 215.699997f);
path.quadTo(136.952286f, 219.321869f, 138.9702f, 222.816986f);
path.lineTo(139.4702f, 223.683014f);
path.quadTo(141.861099f, 227.824188f, 145.081696f, 230.884079f);
path.quadTo(147.70134f, 233.306061f, 150.620407f, 235.00528f);
path.quadTo(153.550888f, 236.683167f, 156.957367f, 237.740662f);
path.quadTo(161.217865f, 239, 166, 239);
path.lineTo(170.482162f, 239);
path.quadTo(176.307037f, 238.210968f, 181.816986f, 235.0298f);
path.lineTo(182.683014f, 234.5298f);
path.quadTo(182.686462f, 234.527817f, 182.689896f, 234.525818f);
path.quadTo(193.804352f, 228.105652f, 197.126709f, 215.70639f);
path.quadTo(200.450104f, 203.303314f, 194.0298f, 192.183014f);
path.lineTo(193.5298f, 191.316986f);
path.quadTo(187.109497f, 180.196686f, 174.706406f, 176.873276f);
path.quadTo(172.503067f, 176.282898f, 170.340179f, 176);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(139.4702f, 223.683014f);
pathB.lineTo(138.9702f, 222.816986f);
pathB.quadTo(132.549896f, 211.696686f, 135.873291f, 199.293594f);
pathB.quadTo(139.196686f, 186.890503f, 150.316986f, 180.4702f);
pathB.lineTo(151.183014f, 179.9702f);
pathB.quadTo(162.303314f, 173.549896f, 174.706406f, 176.873276f);
pathB.quadTo(187.109497f, 180.196686f, 193.5298f, 191.316986f);
pathB.lineTo(194.0298f, 192.183014f);
pathB.quadTo(200.450104f, 203.303314f, 197.126709f, 215.70639f);
pathB.quadTo(193.803314f, 228.109497f, 182.683014f, 234.5298f);
pathB.lineTo(181.816986f, 235.0298f);
pathB.quadTo(170.696686f, 241.450104f, 158.293594f, 238.126709f);
pathB.quadTo(145.890503f, 234.803314f, 139.4702f, 223.683014f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_swapspacesystems_com_5(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(819.050781f, 5539.72412f);
path.quadTo(819.651672f, 5539.1543f, 820.479858f, 5539.17578f);
path.lineTo(1191.35278f, 5548.8877f);
path.quadTo(1192.18091f, 5548.90918f, 1192.7511f, 5549.50977f);
path.quadTo(1193.32141f, 5550.11133f, 1193.29968f, 5550.93945f);
path.lineTo(1186.57214f, 5807.85107f);
path.quadTo(1186.55054f, 5808.6792f, 1185.94958f, 5809.24951f);
path.quadTo(1185.34863f, 5809.81982f, 1184.52051f, 5809.79834f);
path.lineTo(813.647705f, 5800.08643f);
path.quadTo(812.819519f, 5800.06494f, 812.249268f, 5799.46387f);
path.quadTo(811.679016f, 5798.86279f, 811.700684f, 5798.03467f);
path.lineTo(818.428162f, 5541.12305f);
path.quadTo(818.44989f, 5540.29492f, 819.050781f, 5539.72412f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(818.48053f, 5539.12354f);
pathB.lineTo(1193.35205f, 5548.93994f);
pathB.lineTo(1186.5199f, 5809.85059f);
pathB.lineTo(811.648376f, 5800.03418f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_kitcheninspirations_wordpress_com_66(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(47.1666679f, 27820.668f);
path.lineTo(60.8333359f, 27820.668f);
path.lineTo(60.8333359f, 27820.498f);
path.lineTo(47.1666679f, 27820.5f);
path.lineTo(47.1666679f, 27820.668f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(47.1666679f, 27820.668f);
pathB.lineTo(47.1666679f, 27820.498f);
pathB.lineTo(60.8333359f, 27820.5f);
pathB.lineTo(60.8333359f, 27820.668f);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_etiqadd_com_2464(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(630.378662f, 1293.42896f);
path.quadTo(631.257385f, 1292.55029f, 632.5f, 1292.55029f);
path.quadTo(633.742615f, 1292.55029f, 634.621338f, 1293.42896f);
path.lineTo(639.571045f, 1298.37866f);
path.quadTo(640.449768f, 1299.25732f, 640.449707f, 1300.5f);
path.quadTo(640.449768f, 1301.74268f, 639.571045f, 1302.62134f);
path.lineTo(634.621338f, 1307.57104f);
path.quadTo(633.742615f, 1308.44971f, 632.5f, 1308.44971f);
path.quadTo(631.257385f, 1308.44971f, 630.378662f, 1307.57104f);
path.lineTo(625.428955f, 1302.62134f);
path.quadTo(624.550232f, 1301.74268f, 624.550293f, 1300.5f);
path.quadTo(624.550232f, 1299.25732f, 625.428955f, 1298.37866f);
path.lineTo(630.378662f, 1293.42896f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(632.5f, 1291.30762f);
pathB.lineTo(641.692383f, 1300.5f);
pathB.lineTo(632.5f, 1309.69238f);
pathB.lineTo(623.307617f, 1300.5f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_narayana_verlag_de_194(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(1083.34314f, 513.65686f);
path.quadTo(1081, 511.313721f, 1081, 508);
path.lineTo(1257, 508);
path.quadTo(1257, 511.313721f, 1254.65686f, 513.65686f);
path.quadTo(1252.31372f, 516, 1249, 516);
path.lineTo(1089, 516);
path.quadTo(1085.68628f, 516, 1083.34314f, 513.65686f);
path.close();
path.moveTo(1083, 509);
path.lineTo(1255, 509);
path.quadTo(1255, 511.071075f, 1253.53552f, 512.535522f);
path.quadTo(1252.07104f, 514, 1250, 514);
path.lineTo(1088, 514);
path.quadTo(1085.92896f, 514, 1084.46448f, 512.535522f);
path.quadTo(1083, 511.071075f, 1083, 509);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(1081, 508);
pathB.lineTo(1082, 508);
pathB.lineTo(1090.01001f, 516);
pathB.lineTo(1081, 516);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_americascup_com_108(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(999.454102f, 689.17157f);
path.quadTo(1001.172f, 688, 1002.82886f, 688);
path.lineTo(1013.82886f, 688);
path.lineTo(1002.17114f, 713);
path.lineTo(991.171143f, 713);
path.quadTo(989.514282f, 713, 988.889038f, 711.82843f);
path.quadTo(988.263794f, 710.65686f, 989.036377f, 709);
path.lineTo(996.963623f, 692);
path.quadTo(997.736206f, 690.34314f, 999.454102f, 689.17157f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(998.828857f, 688);
pathB.lineTo(1013.82886f, 688);
pathB.lineTo(1002.17114f, 713);
pathB.lineTo(987.171143f, 713);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_vantageproduction_com_109(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(794.514709f, 759.485291f);
path.quadTo(791, 755.970581f, 791, 751);
path.lineTo(1133, 751);
path.quadTo(1133, 755.970581f, 1129.48523f, 759.485291f);
path.quadTo(1125.97058f, 763, 1121, 763);
path.lineTo(803, 763);
path.quadTo(798.029419f, 763, 794.514709f, 759.485291f);
path.close();
path.moveTo(793, 752);
path.lineTo(1131, 752);
path.quadTo(1131, 755.727905f, 1128.36401f, 758.363953f);
path.quadTo(1125.72791f, 761, 1122, 761);
path.lineTo(802, 761);
path.quadTo(798.272095f, 761, 795.636047f, 758.363953f);
path.quadTo(793, 755.727905f, 793, 752);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(791, 751);
pathB.lineTo(792, 751);
pathB.lineTo(804.01001f, 763);
pathB.lineTo(791, 763);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_aceinfographics_com_106(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(166.878677f, 7638.87891f);
path.quadTo(166, 7639.75732f, 166, 7641);
path.lineTo(166, 11577);
path.quadTo(166, 11578.2422f, 166.878677f, 11579.1211f);
path.quadTo(167.388f, 11579.6309f, 168.019989f, 11579.8447f);
path.lineTo(168.019974f, 11576.2979f);
path.quadTo(168, 11576.1533f, 168, 11576);
path.lineTo(168, 7642);
path.lineTo(168.000015f, 7641.99316f);
path.lineTo(168, 7640);
path.lineTo(166.878677f, 7638.87891f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(166, 7638);
pathB.lineTo(168.020004f, 7635.97998f);
pathB.lineTo(168, 11578);
pathB.lineTo(166, 11580);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_tcmevents_org_13(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(465.951904f, 547.960144f);
path.quadTo(465.66571f, 546.867371f, 465.404938f, 546);
path.lineTo(465.504089f, 546);
path.quadTo(465.670349f, 546.601257f, 465.84668f, 547.288391f);
path.quadTo(467.274506f, 552.852356f, 468.506836f, 560.718567f);
path.quadTo(467.336121f, 553.24585f, 465.951904f, 547.960144f);
path.close();
path.moveTo(470.591064f, 574.024353f);
path.quadTo(474.844055f, 601.176025f, 471.728271f, 620.364502f);
path.quadTo(470.567017f, 627.515991f, 468.635742f, 632);
path.lineTo(469.106812f, 632);
path.quadTo(470.791504f, 627.638672f, 471.833496f, 621.036255f);
path.quadTo(474.905701f, 601.569519f, 470.591064f, 574.024353f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(322.992462f, 541.475708f);
pathB.lineTo(465.531616f, 541.724426f);
pathB.lineTo(468.507751f, 560.724426f);
pathB.lineTo(325.968597f, 560.475708f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_paseoitaigara_com_br_56(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(633.147217f, 1247);
path.lineTo(718, 1162.14722f);
path.lineTo(802.852783f, 1247);
path.lineTo(718, 1331.85278f);
path.lineTo(633.147217f, 1247);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(635.268494f, 1244.87866f);
pathB.lineTo(715.878662f, 1164.26855f);
pathB.quadTo(716.757385f, 1163.38989f, 718, 1163.38989f);
pathB.quadTo(719.242615f, 1163.38989f, 720.121338f, 1164.26855f);
pathB.lineTo(800.731506f, 1244.87866f);
pathB.quadTo(801.610168f, 1245.75732f, 801.610168f, 1247);
pathB.quadTo(801.610229f, 1248.24268f, 800.731445f, 1249.12134f);
pathB.lineTo(720.121338f, 1329.73145f);
pathB.quadTo(719.242676f, 1330.61011f, 718, 1330.61011f);
pathB.quadTo(716.757385f, 1330.61011f, 715.878723f, 1329.73145f);
pathB.lineTo(635.268555f, 1249.12134f);
pathB.quadTo(634.389832f, 1248.24268f, 634.389832f, 1247);
pathB.quadTo(634.389832f, 1245.75732f, 635.268494f, 1244.87866f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
static void skpwww_mortgagemarketguide_com_109(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
path.moveTo(816.514709f, 781.485291f);
path.quadTo(813, 777.970581f, 813, 773);
path.lineTo(1133, 773);
path.quadTo(1133, 777.970581f, 1129.48523f, 781.485291f);
path.quadTo(1125.97058f, 785, 1121, 785);
path.lineTo(825, 785);
path.quadTo(820.029419f, 785, 816.514709f, 781.485291f);
path.close();
path.moveTo(815, 774);
path.lineTo(1131, 774);
path.quadTo(1131, 777.727905f, 1128.36401f, 780.363953f);
path.quadTo(1125.72791f, 783, 1122, 783);
path.lineTo(824, 783);
path.quadTo(820.272095f, 783, 817.636047f, 780.363953f);
path.quadTo(815, 777.727905f, 815, 774);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
pathB.moveTo(813, 773);
pathB.lineTo(814, 773);
pathB.lineTo(826.01001f, 785);
pathB.lineTo(813, 785);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
}
fix cubic/line intersection; add skp tests BUG=skia:2488 TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/252243003 git-svn-id: http://skia.googlecode.com/svn/trunk@14458 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-30 13:37:48 +00:00
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
static void skpwww_9to5mac_com_64(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
path.moveTo(365, 5101);
path.lineTo(365, 5082);
path.lineTo(366, 5083);
path.lineTo(367, 5092.96631f);
path.lineTo(367, 5100);
path.quadTo(367, 5101.50537f, 367.967712f, 5102.61084f);
path.lineTo(368.278717f, 5105.71045f);
path.quadTo(367.277618f, 5105.34863f, 366.464478f, 5104.53564f);
path.quadTo(365, 5103.07129f, 365, 5101);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
pathB.moveTo(365, 5082);
pathB.lineTo(365.848175f, 5081.15186f);
pathB.lineTo(368, 5103);
pathB.lineTo(365, 5106);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
}
static void skpwww_googleventures_com_32(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
path.moveTo(725.911682f, 898.767456f);
path.lineTo(741.232544f, 885.911682f);
path.lineTo(754.088318f, 901.232544f);
path.lineTo(738.767456f, 914.088318f);
path.lineTo(725.911682f, 898.767456f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
pathB.moveTo(728.37677f, 870.59082f);
pathB.lineTo(754.088257f, 901.232605f);
pathB.lineTo(738.767395f, 914.088379f);
pathB.lineTo(713.055908f, 883.446594f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
}
static void skpwww_devbridge_com_22(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
path.moveTo(4915, 1523);
path.quadTo(4887.24756f, 1523, 4867.62402f, 1542.6239f);
path.quadTo(4848, 1562.24768f, 4848, 1590);
path.quadTo(4848, 1617.75232f, 4867.62402f, 1637.3761f);
path.quadTo(4887.24756f, 1657, 4915, 1657);
path.quadTo(4942.75244f, 1657, 4962.37598f, 1637.3761f);
path.quadTo(4982, 1617.75232f, 4982, 1590);
path.quadTo(4982, 1562.24768f, 4962.37598f, 1542.6239f);
path.quadTo(4942.75244f, 1523, 4915, 1523);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
pathB.moveTo(4981.99902f, 1590);
pathB.quadTo(4981.99902f, 1617.75232f, 4962.375f, 1637.3761f);
pathB.quadTo(4942.75146f, 1657, 4914.99902f, 1657);
pathB.quadTo(4887.24658f, 1657, 4867.62305f, 1637.3761f);
pathB.quadTo(4847.99902f, 1617.75232f, 4847.99902f, 1590);
pathB.quadTo(4847.99902f, 1562.24768f, 4867.62305f, 1542.6239f);
pathB.quadTo(4887.24658f, 1523, 4914.99902f, 1523);
pathB.quadTo(4942.75146f, 1523, 4962.375f, 1542.6239f);
pathB.quadTo(4981.99902f, 1562.24768f, 4981.99902f, 1590);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
}
static void skpwww_alamdi_com_3(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
path.moveTo(10210.8789f, 5315.87891f);
path.quadTo(10211.7578f, 5315, 10213, 5315);
path.lineTo(10230, 5315);
path.quadTo(10231.2422f, 5315, 10232.1211f, 5315.87891f);
path.quadTo(10233, 5316.75732f, 10233, 5318);
path.lineTo(10233, 5338);
path.quadTo(10233, 5339.24268f, 10232.1211f, 5340.12109f);
path.quadTo(10231.2422f, 5341, 10230, 5341);
path.lineTo(10213, 5341);
path.quadTo(10211.7578f, 5341, 10210.8789f, 5340.12109f);
path.quadTo(10210, 5339.24268f, 10210, 5338);
path.lineTo(10210, 5318);
path.quadTo(10210, 5316.75732f, 10210.8789f, 5315.87891f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kEvenOdd);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
pathB.moveTo(10213, 5315);
pathB.lineTo(10230, 5315);
pathB.cubicTo(10231.6572f, 5315, 10233, 5316.34326f, 10233, 5318);
pathB.lineTo(10233, 5338);
pathB.cubicTo(10233, 5339.10449f, 10231.6572f, 5340, 10230, 5340);
pathB.lineTo(10213, 5340);
pathB.cubicTo(10211.3428f, 5340, 10210, 5339.10449f, 10210, 5338);
pathB.lineTo(10210, 5318);
pathB.cubicTo(10210, 5316.34326f, 10211.3428f, 5315, 10213, 5315);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
}
static void skpwww_familysurvivalprotocol_wordpress_com_61(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
path.moveTo(143, 14557);
path.lineTo(165, 14557);
path.lineTo(165, 14555.9902f);
path.lineTo(143, 14556);
path.lineTo(143, 14557);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
pathB.moveTo(143, 14557);
pathB.lineTo(143, 14555.9902f);
pathB.lineTo(165, 14556);
pathB.lineTo(165, 14557);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
}
static void skpwww_firstunitedbank_com_19(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
path.moveTo(808.585815f, 11673.5859f);
path.quadTo(809.17157f, 11673, 810, 11673);
path.lineTo(1032, 11673);
path.quadTo(1038.21326f, 11673, 1042.60657f, 11677.3936f);
path.quadTo(1047, 11681.7871f, 1047, 11688);
path.quadTo(1047, 11682.2012f, 1042.60657f, 11678.1006f);
path.quadTo(1038.21326f, 11674, 1032, 11674);
path.lineTo(810, 11674);
path.quadTo(809.585815f, 11674, 809.292908f, 11674.293f);
path.quadTo(809, 11674.5859f, 809, 11675);
path.lineTo(809, 11701);
path.quadTo(809, 11701.4141f, 809.292908f, 11701.707f);
path.quadTo(809.585815f, 11702, 810, 11702);
path.lineTo(1032, 11702);
path.quadTo(1038.21326f, 11702, 1042.60657f, 11697.8994f);
path.quadTo(1047, 11693.7988f, 1047, 11688);
path.quadTo(1047, 11694.2129f, 1042.60657f, 11698.6064f);
path.quadTo(1038.21326f, 11703, 1032, 11703);
path.lineTo(810, 11703);
path.quadTo(809.17157f, 11703, 808.585815f, 11702.4141f);
path.quadTo(808, 11701.8281f, 808, 11701);
path.lineTo(808, 11675);
path.quadTo(808, 11674.1719f, 808.585815f, 11673.5859f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
pathB.moveTo(808, 11703);
pathB.lineTo(809.5f, 11701.5f);
pathB.lineTo(1062.91907f, 11687.0811f);
pathB.lineTo(1047, 11703);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
}
static void skpwww_shinydemos_com_5(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
path.moveTo(205.884888f, 648.203857f);
path.lineTo(771.570374f, 82.5183716f);
path.lineTo(1110.98169f, 421.929626f);
path.lineTo(545.296143f, 987.615112f);
path.lineTo(205.884888f, 648.203857f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
pathB.moveTo(771.570374f, 82.5183716f);
pathB.lineTo(1110.98169f, 421.929626f);
pathB.lineTo(545.296204f, 987.615051f);
pathB.lineTo(205.884949f, 648.203796f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
}
static void skpwww_lptemp_com_3(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
path.moveTo(78.6429825f, 1394.30969f);
path.quadTo(79.6192932f, 1393.33337f, 81.0000076f, 1393.33337f);
path.lineTo(341, 1393.33337f);
path.quadTo(342.380707f, 1393.33337f, 343.357025f, 1394.30969f);
path.quadTo(344.333344f, 1395.28601f, 344.333344f, 1396.66675f);
path.lineTo(344.333344f, 1465.66663f);
path.quadTo(344.333344f, 1467.04736f, 343.357025f, 1468.02368f);
path.quadTo(342.380707f, 1469, 341, 1469);
path.lineTo(81.0000076f, 1469);
path.quadTo(79.6192932f, 1469, 78.6429825f, 1468.02368f);
path.quadTo(77.6666718f, 1467.04736f, 77.6666718f, 1465.66663f);
path.lineTo(77.6666718f, 1396.66675f);
path.quadTo(77.6666718f, 1395.28601f, 78.6429825f, 1394.30969f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kEvenOdd);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
pathB.moveTo(81, 1393.33337f);
pathB.lineTo(341, 1393.33337f);
pathB.cubicTo(342.840942f, 1393.33337f, 344.333344f, 1394.82568f, 344.333344f, 1396.66675f);
pathB.lineTo(344.333344f, 1465.66675f);
pathB.cubicTo(344.333344f, 1467.32361f, 342.840942f, 1468.66675f, 341, 1468.66675f);
pathB.lineTo(81, 1468.66675f);
pathB.cubicTo(79.15905f, 1468.66675f, 77.6666718f, 1467.32361f, 77.6666718f, 1465.66675f);
pathB.lineTo(77.6666718f, 1396.66675f);
pathB.cubicTo(77.6666718f, 1394.82568f, 79.15905f, 1393.33337f, 81, 1393.33337f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
}
fix pathops skp-derived bugs; add more failing tests TBR= Author: caryclark@google.com Review URL: https://codereview.chromium.org/463883002
2014-08-12 14:46:33 +00:00
static void skpwww_shinydemos_com_15(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix pathops skp-derived bugs; add more failing tests TBR= Author: caryclark@google.com Review URL: https://codereview.chromium.org/463883002
2014-08-12 14:46:33 +00:00
path.moveTo(1000, 310.947968f);
path.lineTo(771.570374f, 82.5183716f);
path.lineTo(205.884888f, 648.203857f);
path.lineTo(448.68103f, 891);
path.lineTo(641.911255f, 891);
path.lineTo(1000, 532.911316f);
path.lineTo(1000, 310.947968f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
fix pathops skp-derived bugs; add more failing tests TBR= Author: caryclark@google.com Review URL: https://codereview.chromium.org/463883002
2014-08-12 14:46:33 +00:00
pathB.moveTo(771.570374f, 82.5183716f);
pathB.lineTo(1110.98169f, 421.929626f);
pathB.lineTo(545.296204f, 987.615051f);
pathB.lineTo(205.884949f, 648.203796f);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix pathops skp-derived bugs; add more failing tests TBR= Author: caryclark@google.com Review URL: https://codereview.chromium.org/463883002
2014-08-12 14:46:33 +00:00
}
static void skpwww_lptemp_com_5(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
fix pathops skp-derived bugs; add more failing tests TBR= Author: caryclark@google.com Review URL: https://codereview.chromium.org/463883002
2014-08-12 14:46:33 +00:00
path.moveTo(78.6429825f, 3150.97632f);
path.quadTo(79.6192932f, 3150, 81.0000076f, 3150);
path.lineTo(341, 3150);
path.quadTo(342.380707f, 3150, 343.357025f, 3150.97632f);
path.quadTo(344.333344f, 3151.95264f, 344.333344f, 3153.33325f);
path.lineTo(344.333344f, 5205.3335f);
path.quadTo(344.333344f, 5206.71436f, 343.357025f, 5207.69092f);
path.quadTo(342.380707f, 5208.66699f, 341, 5208.66699f);
path.lineTo(81.0000076f, 5208.66699f);
path.quadTo(79.6192932f, 5208.66699f, 78.6429825f, 5207.69092f);
path.quadTo(77.6666718f, 5206.71436f, 77.6666718f, 5205.3335f);
path.lineTo(77.6666718f, 3153.33325f);
path.quadTo(77.6666718f, 3151.95264f, 78.6429825f, 3150.97632f);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kEvenOdd);
fix pathops skp-derived bugs; add more failing tests TBR= Author: caryclark@google.com Review URL: https://codereview.chromium.org/463883002
2014-08-12 14:46:33 +00:00
pathB.moveTo(81, 3150);
pathB.lineTo(341, 3150);
pathB.cubicTo(342.840942f, 3150, 344.333344f, 3151.49268f, 344.333344f, 3153.3335f);
pathB.lineTo(344.333344f, 5205.3335f);
pathB.cubicTo(344.333344f, 5206.99023f, 342.840942f, 5208.3335f, 341, 5208.3335f);
pathB.lineTo(81, 5208.3335f);
pathB.cubicTo(79.15905f, 5208.3335f, 77.6666718f, 5206.99023f, 77.6666718f, 5205.3335f);
pathB.lineTo(77.6666718f, 3153.3335f);
pathB.cubicTo(77.6666718f, 3151.49268f, 79.15905f, 3150, 81, 3150);
pathB.close();
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
}
static void skpwww_educationalcraft_com_4a(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
path.moveTo(941, 1494);
path.lineTo(941, 1464);
path.lineTo(985, 1464);
path.lineTo(985, 1494);
path.lineTo(941, 1494);
path.close();
SkPath pathB;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
pathB.setFillType(SkPathFillType::kWinding);
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
pathB.moveTo(984.546021f, 1478.31494f);
pathB.cubicTo(984.546021f, 1478.31494f, 984.543213f, 1478.32239f, 984.537598f, 1478.33655f);
pathB.cubicTo(984.419006f, 1478.63477f, 983.044373f, 1481.90405f, 980.026001f, 1481.276f);
pathB.cubicTo(980.026001f, 1481.276f, 980.02594f, 1481.27576f, 980.025879f, 1481.27527f);
pathB.cubicTo(980.018494f, 1481.22131f, 979.602478f, 1478.38831f, 984.546021f, 1478.31494f);
testPathOp(reporter, path, pathB, kIntersect_SkPathOp, filename);
fix pathops skp-derived bugs; add more failing tests TBR= Author: caryclark@google.com Review URL: https://codereview.chromium.org/463883002
2014-08-12 14:46:33 +00:00
}
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
static void skpwww_woothemes_com_1(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 1);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(SkBits2Float(0x44472795), SkBits2Float(0x455cdb8d)); // 796.618f, 3533.72f
path.lineTo(SkBits2Float(0x44467c27), SkBits2Float(0x455cdb8d)); // 793.94f, 3533.72f
path.lineTo(SkBits2Float(0x44467c27), SkBits2Float(0x455d055d)); // 793.94f, 3536.34f
path.lineTo(SkBits2Float(0x44472795), SkBits2Float(0x455d055d)); // 796.618f, 3536.34f
path.lineTo(SkBits2Float(0x44472795), SkBits2Float(0x455cdb8d)); // 796.618f, 3533.72f
SkPath path1(path);
path.reset();
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 0);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(SkBits2Float(0x4446861c), SkBits2Float(0x455cdb8d)); // 794.095f, 3533.72f
path.cubicTo(SkBits2Float(0x4446a0d8), SkBits2Float(0x455cefbb), SkBits2Float(0x444727a5), SkBits2Float(0x455d055d), SkBits2Float(0x444727a5), SkBits2Float(0x455d055d)); // 794.513f, 3534.98f, 796.619f, 3536.34f, 796.619f, 3536.34f
path.cubicTo(SkBits2Float(0x4446c5b0), SkBits2Float(0x455cf8a4), SkBits2Float(0x444693af), SkBits2Float(0x455cedad), SkBits2Float(0x44467c1b), SkBits2Float(0x455ce4b8)); // 795.089f, 3535.54f, 794.308f, 3534.85f, 793.939f, 3534.29f
path.lineTo(SkBits2Float(0x44467d70), SkBits2Float(0x455ce016)); // 793.96f, 3534.01f
path.cubicTo(SkBits2Float(0x44467fa9), SkBits2Float(0x455cde82), SkBits2Float(0x444682b5), SkBits2Float(0x455cdd03), SkBits2Float(0x4446861c), SkBits2Float(0x455cdb8d)); // 793.995f, 3533.91f, 794.042f, 3533.81f, 794.095f, 3533.72f
SkPath path2(path);
testPathOp(reporter, path1, path2, (SkPathOp) 1, filename);
}
static void skpwww_gorcraft_ru_1(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 1);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(SkBits2Float(0x44924000), SkBits2Float(0x458e7800)); // 1170, 4559
path.conicTo(SkBits2Float(0x44930000), SkBits2Float(0x458e7800), SkBits2Float(0x44930000), SkBits2Float(0x458ea800), SkBits2Float(0x3f3504f3)); // 1176, 4559, 1176, 4565, 0.707107f
path.lineTo(SkBits2Float(0x44930000), SkBits2Float(0x458f7000)); // 1176, 4590
path.conicTo(SkBits2Float(0x44930000), SkBits2Float(0x458f9800), SkBits2Float(0x44926000), SkBits2Float(0x458f9800), SkBits2Float(0x3f3504f3)); // 1176, 4595, 1171, 4595, 0.707107f
path.lineTo(SkBits2Float(0x42a60000), SkBits2Float(0x458f9800)); // 83, 4595
path.conicTo(SkBits2Float(0x429c0471), SkBits2Float(0x458f9800), SkBits2Float(0x429c0000), SkBits2Float(0x458f700c), SkBits2Float(0x3f352d2d)); // 78.0087f, 4595, 78, 4590.01f, 0.707721f
path.lineTo(SkBits2Float(0x429c0000), SkBits2Float(0x458ea800)); // 78, 4565
path.conicTo(SkBits2Float(0x429c0000), SkBits2Float(0x458e7800), SkBits2Float(0x42a80000), SkBits2Float(0x458e7800), SkBits2Float(0x3f3504f3)); // 78, 4559, 84, 4559, 0.707107f
path.lineTo(SkBits2Float(0x44924000), SkBits2Float(0x458e7800)); // 1170, 4559
path.close();
SkPath path1(path);
path.reset();
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 0);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(SkBits2Float(0x429c0000), SkBits2Float(0x458f7000)); // 78, 4590
path.lineTo(SkBits2Float(0x429c0000), SkBits2Float(0x458ea800)); // 78, 4565
path.conicTo(SkBits2Float(0x429c0000), SkBits2Float(0x458e7800), SkBits2Float(0x42a80000), SkBits2Float(0x458e7800), SkBits2Float(0x3f3504f3)); // 78, 4559, 84, 4559, 0.707107f
path.lineTo(SkBits2Float(0x431e0000), SkBits2Float(0x458e7800)); // 158, 4559
path.conicTo(SkBits2Float(0x431e0000), SkBits2Float(0x458e7800), SkBits2Float(0x431e0000), SkBits2Float(0x458e7800), SkBits2Float(0x3f3504f3)); // 158, 4559, 158, 4559, 0.707107f
path.lineTo(SkBits2Float(0x431e0000), SkBits2Float(0x458fa000)); // 158, 4596
path.conicTo(SkBits2Float(0x431e0000), SkBits2Float(0x458fa000), SkBits2Float(0x431e0000), SkBits2Float(0x458fa000), SkBits2Float(0x3f3504f3)); // 158, 4596, 158, 4596, 0.707107f
path.lineTo(SkBits2Float(0x42a80000), SkBits2Float(0x458fa000)); // 84, 4596
path.conicTo(SkBits2Float(0x429c0000), SkBits2Float(0x458fa000), SkBits2Float(0x429c0000), SkBits2Float(0x458f7000), SkBits2Float(0x3f3504f3)); // 78, 4596, 78, 4590, 0.707107f
path.close();
SkPath path2(path);
testPathOp(reporter, path1, path2, (SkPathOp) 2, filename);
}
static void skpwww_neda_net_1(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 1);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(SkBits2Float(0x447a0000), SkBits2Float(0x00000000)); // 1000, 0
path.lineTo(SkBits2Float(0x00000000), SkBits2Float(0x00000000)); // 0, 0
path.lineTo(SkBits2Float(0x00000000), SkBits2Float(0x44b6e000)); // 0, 1463
path.lineTo(SkBits2Float(0x447a0000), SkBits2Float(0x44b6e000)); // 1000, 1463
path.lineTo(SkBits2Float(0x447a0000), SkBits2Float(0x00000000)); // 1000, 0
path.close();
SkPath path1(path);
path.reset();
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 0);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(SkBits2Float(0x366a410f), SkBits2Float(0x43a38000)); // 3.49066e-06f, 327
path.lineTo(SkBits2Float(0x447a0000), SkBits2Float(0x43a38001)); // 1000, 327
path.lineTo(SkBits2Float(0x447a0000), SkBits2Float(0x4435c000)); // 1000, 727
path.lineTo(SkBits2Float(0xb66a410d), SkBits2Float(0x4435c000)); // -3.49066e-06f, 727
path.lineTo(SkBits2Float(0x366a410f), SkBits2Float(0x43a38000)); // 3.49066e-06f, 327
path.close();
SkPath path2(path);
testPathOp(reporter, path1, path2, (SkPathOp) 1, filename);
}
// "http___www_neda_net.skp" dir=87
static void skpwww_neda_net_2(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 1);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(SkBits2Float(0x442fc000), SkBits2Float(0x4546a000)); // 703, 3178
path.lineTo(SkBits2Float(0x441f4000), SkBits2Float(0x4546a000)); // 637, 3178
path.lineTo(SkBits2Float(0x441f4000), SkBits2Float(0x454ab000)); // 637, 3243
path.lineTo(SkBits2Float(0x442fc000), SkBits2Float(0x454ab000)); // 703, 3243
path.lineTo(SkBits2Float(0x442fc000), SkBits2Float(0x4546a000)); // 703, 3178
path.close();
SkPath path1(path);
path.reset();
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 0);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(SkBits2Float(0x44220e6e), SkBits2Float(0x45469c4c)); // 648.225f, 3177.77f
path.lineTo(SkBits2Float(0x442fc01c), SkBits2Float(0x45475696)); // 703.002f, 3189.41f
path.lineTo(SkBits2Float(0x442cf191), SkBits2Float(0x454aa3b5)); // 691.774f, 3242.23f
path.lineTo(SkBits2Float(0x441f3fe3), SkBits2Float(0x4549e96b)); // 636.998f, 3230.59f
path.lineTo(SkBits2Float(0x44220e6e), SkBits2Float(0x45469c4c)); // 648.225f, 3177.77f
path.close();
SkPath path2(path);
testPathOp(reporter, path1, path2, (SkPathOp) 1, filename);
}
static void skpwww_mybuilder_com_1(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(1000, 659);
path.lineTo(1000, 377);
path.lineTo(455, 377);
path.lineTo(455, 659);
path.lineTo(1000, 659);
path.close();
SkPath path1(path);
path.reset();
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(921.472f, 414.086f);
path.lineTo(968.815f, 386.754f);
path.lineTo(993.069f, 428.761f);
path.lineTo(945.726f, 456.096f);
path.lineTo(921.471f, 414.086f);
path.lineTo(921.472f, 414.086f);
path.close();
path.moveTo(971.151f, 422.889f);
path.cubicTo(966.509f, 414.848f, 957.649f, 411.727f, 950.181f, 416.038f);
path.lineTo(947.224f, 417.746f);
path.lineTo(946.979f, 417.887f);
path.lineTo(947.838f, 419.371f);
path.lineTo(947.844f, 419.367f);
path.lineTo(947.868f, 419.353f);
path.lineTo(947.945f, 419.309f);
path.cubicTo(947.988f, 419.285f, 947.988f, 419.285f, 948.023f, 419.263f);
path.cubicTo(948.039f, 419.255f, 948.039f, 419.255f, 948.047f, 419.25f);
path.lineTo(948.052f, 419.247f);
path.lineTo(947.196f, 417.762f);
path.lineTo(947.195f, 417.762f);
path.lineTo(946.888f, 417.939f);
path.lineTo(943.39f, 419.959f);
path.lineTo(944.249f, 421.443f);
path.lineTo(947.745f, 419.424f);
path.lineTo(948.05f, 419.247f);
path.lineTo(948.052f, 419.247f);
path.lineTo(947.195f, 417.763f);
path.cubicTo(947.193f, 417.763f, 947.193f, 417.763f, 947.19f, 417.766f);
path.lineTo(947.166f, 417.779f);
path.lineTo(947.087f, 417.825f);
path.lineTo(947.011f, 417.868f);
path.lineTo(946.987f, 417.883f);
path.lineTo(946.982f, 417.886f);
path.lineTo(946.98f, 417.886f);
path.lineTo(947.839f, 419.37f);
path.lineTo(948.083f, 419.229f);
path.lineTo(951.039f, 417.522f);
path.cubicTo(957.631f, 413.716f, 965.471f, 416.477f, 969.669f, 423.746f);
path.lineTo(971.153f, 422.889f);
path.lineTo(971.151f, 422.889f);
path.close();
SkPath path2(path);
testPathOp(reporter, path1, path2, kIntersect_SkPathOp, filename);
}
static void skpwww_nimble_com_au_1(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(188.6f, 1988.8f);
path.lineTo(188.6f, 2041.6f);
path.cubicTo(188.6f, 2065.4f, 208, 2084.8f, 231.8f, 2084.8f);
path.cubicTo(255.6f, 2084.8f, 275, 2065.4f, 275, 2041.6f);
path.lineTo(275.2f, 2041.6f);
path.lineTo(275.2f, 1988.8f);
path.lineTo(188.6f, 1988.8f);
path.close();
SkPath path1(path);
path.reset();
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kWinding);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(275, 2041.6f);
path.conicTo(275, 2084.8f, 231.8f, 2084.8f, 0.707107f);
path.conicTo(188.6f, 2084.8f, 188.6f, 2041.6f, 0.707107f);
path.conicTo(188.6f, 1998.4f, 231.8f, 1998.4f, 0.707107f);
path.conicTo(275, 1998.4f, 275, 2041.6f, 0.707107f);
path.close();
SkPath path2(path);
testPathOp(reporter, path1, path2, kIntersect_SkPathOp, filename);
}
static void skpwww_tinytots_com_1(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(75.96f, 26.318f);
path.lineTo(70.337f, 26.318f);
path.lineTo(70.337f, 32.376f);
path.lineTo(75.96f, 32.376f);
path.lineTo(75.96f, 26.318f);
path.close();
SkPath path1(path);
path.reset();
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kWinding);
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
path.moveTo(75.88f, 27.873f);
path.cubicTo(75.929f, 28.138f, 75.956f, 29.196f, 75.96f, 31.046f);
path.lineTo(72.766f, 32.376f);
path.cubicTo(72.763f, 30.525f, 72.735f, 29.468f, 72.686f, 29.203f);
path.cubicTo(72.636f, 28.94f, 72.519f, 28.722f, 72.335f, 28.552f);
path.cubicTo(72.248f, 28.472f, 72.058f, 28.364f, 71.763f, 28.228f);
path.cubicTo(72.425f, 27.933f, 72.425f, 27.933f, 73.395f, 27.498f);
path.cubicTo(72.425f, 27.933f, 72.425f, 27.933f, 71.763f, 28.228f);
path.cubicTo(71.425f, 28.072f, 70.95f, 27.878f, 70.337f, 27.647f);
path.lineTo(73.531f, 26.317f);
path.cubicTo(74.144f, 26.547f, 74.619f, 26.741f, 74.957f, 26.898f);
path.cubicTo(74.475f, 27.113f, 73.993f, 27.329f, 73.511f, 27.544f);
path.cubicTo(73.993f, 27.329f, 74.475f, 27.114f, 74.957f, 26.898f);
path.cubicTo(75.252f, 27.034f, 75.442f, 27.142f, 75.529f, 27.222f);
path.cubicTo(75.713f, 27.393f, 75.83f, 27.61f, 75.88f, 27.873f);
SkPath path2(path);
testPathOp(reporter, path1, path2, kIntersect_SkPathOp, filename);
}
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
static void http___www_emuleteca_cl_26(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 1);
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
path.moveTo(SkBits2Float(0x3f800000), SkBits2Float(0x44370000)); // 1, 732
path.conicTo(SkBits2Float(0x3f800000), SkBits2Float(0x4428c000), SkBits2Float(0x42680000), SkBits2Float(0x4428c000), SkBits2Float(0x3f3504f3)); // 1, 675, 58, 675, 0.707107f
path.conicTo(SkBits2Float(0x40400000), SkBits2Float(0x4428c000), SkBits2Float(0x40400000), SkBits2Float(0x44370000), SkBits2Float(0x3f3504f3)); // 3, 675, 3, 732, 0.707107f
path.conicTo(SkBits2Float(0x40400000), SkBits2Float(0x44444000), SkBits2Float(0x42680000), SkBits2Float(0x44444000), SkBits2Float(0x3f3504f3)); // 3, 785, 58, 785, 0.707107f
path.lineTo(SkBits2Float(0x446d0000), SkBits2Float(0x44444000)); // 948, 785
path.conicTo(SkBits2Float(0x447ac000), SkBits2Float(0x44444000), SkBits2Float(0x447ac000), SkBits2Float(0x44370000), SkBits2Float(0x3f3504f3)); // 1003, 785, 1003, 732, 0.707107f
path.conicTo(SkBits2Float(0x447ac000), SkBits2Float(0x4428c000), SkBits2Float(0x446d0000), SkBits2Float(0x4428c000), SkBits2Float(0x3f3504f3)); // 1003, 675, 948, 675, 0.707107f
path.conicTo(SkBits2Float(0x447b4000), SkBits2Float(0x4428c000), SkBits2Float(0x447b4000), SkBits2Float(0x44370000), SkBits2Float(0x3f3504f3)); // 1005, 675, 1005, 732, 0.707107f
path.conicTo(SkBits2Float(0x447b4000), SkBits2Float(0x44454000), SkBits2Float(0x446d0000), SkBits2Float(0x44454000), SkBits2Float(0x3f3504f3)); // 1005, 789, 948, 789, 0.707107f
path.lineTo(SkBits2Float(0x42680000), SkBits2Float(0x44454000)); // 58, 789
path.conicTo(SkBits2Float(0x3f800000), SkBits2Float(0x44454000), SkBits2Float(0x3f800000), SkBits2Float(0x44370000), SkBits2Float(0x3f3504f3)); // 1, 789, 1, 732, 0.707107f
path.close();
SkPath path1(path);
path.reset();
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 0);
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
path.moveTo(SkBits2Float(0x3f800000), SkBits2Float(0x4428c000)); // 1, 675
path.lineTo(SkBits2Float(0x42680000), SkBits2Float(0x4428c000)); // 58, 675
path.lineTo(SkBits2Float(0x3fc8f676), SkBits2Float(0x44454000)); // 1.57002f, 789
path.lineTo(SkBits2Float(0x3f800000), SkBits2Float(0x44454000)); // 1, 789
SkPath path2(path);
testPathOp(reporter, path1, path2, (SkPathOp) 1, filename);
}
static void http___www_emuleteca_cl_27(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 1);
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
path.moveTo(SkBits2Float(0x42680000), SkBits2Float(0x4428c000)); // 58, 675
path.conicTo(SkBits2Float(0x3f800000), SkBits2Float(0x4428c000), SkBits2Float(0x3f800000), SkBits2Float(0x44370000), SkBits2Float(0x3f3504f3)); // 1, 675, 1, 732, 0.707107f
path.conicTo(SkBits2Float(0x3f800000), SkBits2Float(0x443bd045), SkBits2Float(0x414acf56), SkBits2Float(0x443fa420), SkBits2Float(0x3f778612)); // 1, 751.254f, 12.6756f, 766.564f, 0.96689f
path.lineTo(SkBits2Float(0x41606e3d), SkBits2Float(0x443ef569)); // 14.0269f, 763.835f
path.conicTo(SkBits2Float(0x40400000), SkBits2Float(0x443b6c34), SkBits2Float(0x40400000), SkBits2Float(0x44370000), SkBits2Float(0x3f77ac46)); // 3, 749.691f, 3, 732, 0.967472f
path.conicTo(SkBits2Float(0x40400000), SkBits2Float(0x4428c000), SkBits2Float(0x42680000), SkBits2Float(0x4428c000), SkBits2Float(0x3f3504f3)); // 3, 675, 58, 675, 0.707107f
path.close();
SkPath path1(path);
path.reset();
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType((SkPathFillType) 0);
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
path.moveTo(SkBits2Float(0x3f800000), SkBits2Float(0x4428c000)); // 1, 675
path.lineTo(SkBits2Float(0x4c4a3de6), SkBits2Float(0xccca3d89)); // 5.30165e+07f, -1.06032e+08f
path.lineTo(SkBits2Float(0x41a71147), SkBits2Float(0x443b4eec)); // 20.8834f, 749.233f
path.lineTo(SkBits2Float(0x3f800000), SkBits2Float(0x44454000)); // 1, 789
SkPath path2(path);
testPathOp(reporter, path1, path2, (SkPathOp) 1, filename);
}
static void http___www_emuleteca_cl_28(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
SkPath path1(path);
path.reset();
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kWinding);
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
path.moveTo(SkBits2Float(0x3f800000), SkBits2Float(0x4428c000)); // 1, 675
path.lineTo(SkBits2Float(0x4c4a3de6), SkBits2Float(0xccca3d89)); // 5.30165e+07f, -1.06032e+08f
path.lineTo(SkBits2Float(0x41a71147), SkBits2Float(0x443b4eec)); // 20.8834f, 749.233f
path.lineTo(SkBits2Float(0x3f800000), SkBits2Float(0x44454000)); // 1, 789
SkPath path2(path);
testPathOp(reporter, path1, path2, kIntersect_SkPathOp, filename);
}
static void http___www_project2061_org(skiatest::Reporter* reporter, const char* filename) {
SkPath path;
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kEvenOdd);
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
path.moveTo(751, 62);
path.lineTo(497, 62);
path.lineTo(497, 138);
path.lineTo(751, 138);
path.lineTo(751, 62);
path.close();
SkPath path1(path);
path.reset();
Revert "Revert "switch to new filltype for SkPath"" This reverts commit 3e7af41224eecf8393f3cdc08c4c9800b2321c20. Change-Id: Id4f66b3956f4bdbe690db20fc478b7365ee89717 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/256676 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com> Auto-Submit: Mike Reed <reed@google.com>
2019-11-26 17:17:17 +00:00
path.setFillType(SkPathFillType::kWinding);
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
path.moveTo(699.889f, 115.29f);
path.cubicTo(699.889f, 114.913f, 699.627f, 114.723f, 699.111f, 114.723f);
path.cubicTo(698.604f, 114.723f, 697.843f, 114.671f, 697.096f, 114.423f);
path.cubicTo(695.94f, 114.094f, 694.53f, 113.168f, 692.746f, 108.775f);
path.cubicTo(690.469f, 103.12f, 685.644f, 90.563f, 682.761f, 83.061f);
path.lineTo(680.541f, 77.301f);
path.cubicTo(679.927f, 75.721f, 679.67f, 75.261f, 679.151f, 75.261f);
path.cubicTo(678.636f, 75.261f, 678.392f, 75.73f, 677.759f, 77.464f);
path.lineTo(665.868f, 109.881f);
path.cubicTo(664.861f, 112.669f, 663.852f, 114.339f, 661.675f, 114.622f);
path.cubicTo(661.675f, 114.622f, 661.551f, 114.631f, 661.378f, 114.642f);
path.cubicTo(661.072f, 114.597f, 660.75f, 114.531f, 660.427f, 114.424f);
path.cubicTo(659.272f, 114.095f, 657.861f, 113.169f, 656.078f, 108.776f);
path.cubicTo(653.797f, 103.112f, 648.967f, 90.541f, 646.08f, 83.029f);
path.lineTo(643.873f, 77.302f);
path.cubicTo(643.259f, 75.722f, 643.002f, 75.262f, 642.484f, 75.262f);
path.cubicTo(641.968f, 75.262f, 641.724f, 75.731f, 641.092f, 77.465f);
path.lineTo(629.2f, 109.879f);
path.cubicTo(628.193f, 112.667f, 627.184f, 114.337f, 625.007f, 114.62f);
path.cubicTo(625.007f, 114.62f, 624.884f, 114.629f, 624.71f, 114.64f);
path.cubicTo(624.404f, 114.595f, 624.081f, 114.529f, 623.759f, 114.422f);
path.cubicTo(622.603f, 114.093f, 621.193f, 113.167f, 619.409f, 108.774f);
path.cubicTo(617.132f, 103.119f, 612.307f, 90.562f, 609.424f, 83.06f);
path.lineTo(607.204f, 77.3f);
path.cubicTo(606.59f, 75.72f, 606.333f, 75.26f, 605.815f, 75.26f);
path.cubicTo(605.3f, 75.26f, 605.055f, 75.729f, 604.423f, 77.463f);
path.lineTo(592.531f, 109.879f);
path.cubicTo(591.524f, 112.667f, 590.514f, 114.337f, 588.338f, 114.62f);
path.cubicTo(588.337f, 114.62f, 586.852f, 114.722f, 586.852f, 114.722f);
path.cubicTo(586.418f, 114.722f, 586.126f, 114.95f, 586.126f, 115.289f);
path.cubicTo(586.126f, 115.96f, 586.97f, 115.96f, 587.376f, 115.96f);
path.lineTo(592.101f, 115.843f);
path.lineTo(593.295f, 115.804f);
path.lineTo(594.624f, 115.86f);
path.lineTo(597.381f, 115.961f);
path.cubicTo(598.241f, 115.961f, 598.422f, 115.596f, 598.422f, 115.29f);
path.cubicTo(598.422f, 115.078f, 598.321f, 114.723f, 597.643f, 114.723f);
path.lineTo(597.119f, 114.723f);
path.cubicTo(596.448f, 114.723f, 595.383f, 114.381f, 595.383f, 113.463f);
path.cubicTo(595.383f, 112.545f, 595.638f, 111.334f, 596.101f, 110.052f);
path.cubicTo(596.103f, 110.048f, 599.246f, 100.809f, 599.246f, 100.809f);
path.cubicTo(599.337f, 100.583f, 599.435f, 100.564f, 599.528f, 100.564f);
path.lineTo(610.791f, 100.564f);
path.cubicTo(610.929f, 100.564f, 610.958f, 100.585f, 611.035f, 100.737f);
path.cubicTo(611.035f, 100.741f, 615.896f, 113.554f, 615.896f, 113.554f);
path.cubicTo(615.972f, 113.754f, 616.01f, 113.942f, 616.01f, 114.105f);
path.cubicTo(616.01f, 114.228f, 615.988f, 114.338f, 615.945f, 114.429f);
path.cubicTo(615.876f, 114.573f, 615.779f, 114.619f, 615.741f, 114.631f);
path.cubicTo(615.272f, 114.748f, 615.2f, 115.061f, 615.2f, 115.237f);
path.cubicTo(615.2f, 115.766f, 615.792f, 115.796f, 616.772f, 115.844f);
path.lineTo(617.012f, 115.857f);
path.cubicTo(618.978f, 115.913f, 621.359f, 115.948f, 623.835f, 115.958f);
path.cubicTo(623.912f, 115.961f, 623.984f, 115.961f, 624.045f, 115.961f);
path.lineTo(624.118f, 115.959f);
path.cubicTo(624.494f, 115.961f, 624.871f, 115.961f, 625.25f, 115.961f);
path.cubicTo(625.47f, 115.961f, 625.662f, 115.946f, 625.827f, 115.917f);
path.lineTo(628.77f, 115.844f);
path.lineTo(629.965f, 115.805f);
path.lineTo(631.293f, 115.861f);
path.lineTo(634.051f, 115.962f);
path.cubicTo(634.911f, 115.962f, 635.092f, 115.597f, 635.092f, 115.291f);
path.cubicTo(635.092f, 115.079f, 634.991f, 114.724f, 634.313f, 114.724f);
path.lineTo(633.789f, 114.724f);
path.cubicTo(633.118f, 114.724f, 632.053f, 114.382f, 632.053f, 113.464f);
path.cubicTo(632.053f, 112.546f, 632.308f, 111.335f, 632.771f, 110.053f);
path.cubicTo(632.773f, 110.049f, 635.916f, 100.81f, 635.916f, 100.81f);
path.cubicTo(636.007f, 100.584f, 636.105f, 100.565f, 636.198f, 100.565f);
path.lineTo(647.46f, 100.565f);
path.cubicTo(647.598f, 100.565f, 647.627f, 100.586f, 647.704f, 100.737f);
path.cubicTo(647.704f, 100.741f, 652.565f, 113.554f, 652.565f, 113.554f);
path.cubicTo(652.641f, 113.754f, 652.679f, 113.942f, 652.679f, 114.105f);
path.cubicTo(652.679f, 114.228f, 652.657f, 114.338f, 652.613f, 114.43f);
path.cubicTo(652.544f, 114.574f, 652.447f, 114.619f, 652.409f, 114.632f);
path.cubicTo(651.94f, 114.749f, 651.868f, 115.062f, 651.868f, 115.238f);
path.cubicTo(651.868f, 115.767f, 652.46f, 115.797f, 653.44f, 115.845f);
path.lineTo(653.681f, 115.858f);
path.cubicTo(655.647f, 115.914f, 658.028f, 115.949f, 660.503f, 115.959f);
path.cubicTo(660.58f, 115.962f, 660.652f, 115.962f, 660.713f, 115.962f);
path.lineTo(660.787f, 115.96f);
path.cubicTo(661.162f, 115.962f, 661.54f, 115.962f, 661.918f, 115.962f);
path.cubicTo(662.139f, 115.962f, 662.33f, 115.947f, 662.496f, 115.918f);
path.lineTo(665.439f, 115.845f);
path.lineTo(666.633f, 115.806f);
path.lineTo(667.962f, 115.862f);
path.lineTo(670.719f, 115.963f);
path.cubicTo(671.579f, 115.963f, 671.76f, 115.598f, 671.76f, 115.292f);
path.cubicTo(671.76f, 115.08f, 671.659f, 114.725f, 670.981f, 114.725f);
path.lineTo(670.457f, 114.725f);
path.cubicTo(669.786f, 114.725f, 668.721f, 114.383f, 668.721f, 113.465f);
path.cubicTo(668.721f, 112.547f, 668.976f, 111.336f, 669.439f, 110.054f);
path.cubicTo(669.441f, 110.05f, 672.584f, 100.811f, 672.584f, 100.811f);
path.cubicTo(672.675f, 100.585f, 672.773f, 100.566f, 672.866f, 100.566f);
path.lineTo(684.128f, 100.566f);
path.cubicTo(684.266f, 100.566f, 684.295f, 100.587f, 684.372f, 100.739f);
path.cubicTo(684.372f, 100.743f, 689.233f, 113.556f, 689.233f, 113.556f);
path.cubicTo(689.309f, 113.756f, 689.347f, 113.944f, 689.347f, 114.107f);
path.cubicTo(689.347f, 114.23f, 689.325f, 114.34f, 689.281f, 114.431f);
path.cubicTo(689.212f, 114.575f, 689.115f, 114.621f, 689.077f, 114.633f);
path.cubicTo(688.608f, 114.75f, 688.536f, 115.063f, 688.536f, 115.239f);
path.cubicTo(688.536f, 115.768f, 689.128f, 115.798f, 690.108f, 115.846f);
path.lineTo(690.348f, 115.859f);
path.cubicTo(692.687f, 115.926f, 695.611f, 115.963f, 698.586f, 115.963f);
path.cubicTo(699.451f, 115.961f, 699.889f, 115.735f, 699.889f, 115.29f);
path.close();
path.moveTo(600.18f, 98.176f);
path.cubicTo(600.181f, 98.171f, 600.185f, 98.158f, 600.185f, 98.158f);
path.cubicTo(600.191f, 98.142f, 604.801f, 84.049f, 604.801f, 84.049f);
path.cubicTo(604.865f, 83.857f, 604.915f, 83.756f, 604.951f, 83.697f);
path.cubicTo(604.987f, 83.756f, 605.037f, 83.858f, 605.102f, 84.051f);
path.cubicTo(605.103f, 84.054f, 610.08f, 98.149f, 610.08f, 98.149f);
path.cubicTo(610.079f, 98.145f, 610.079f, 98.145f, 610.079f, 98.147f);
path.cubicTo(610.079f, 98.149f, 610.081f, 98.164f, 610.083f, 98.176f);
path.lineTo(600.18f, 98.176f);
path.close();
path.moveTo(636.849f, 98.176f);
path.cubicTo(636.851f, 98.171f, 636.854f, 98.158f, 636.854f, 98.158f);
path.cubicTo(636.859f, 98.142f, 641.469f, 84.049f, 641.469f, 84.049f);
path.cubicTo(641.533f, 83.857f, 641.584f, 83.756f, 641.62f, 83.697f);
path.cubicTo(641.656f, 83.756f, 641.706f, 83.857f, 641.771f, 84.051f);
path.cubicTo(641.773f, 84.054f, 646.749f, 98.149f, 646.749f, 98.149f);
path.cubicTo(646.749f, 98.145f, 646.748f, 98.145f, 646.748f, 98.147f);
path.cubicTo(646.748f, 98.15f, 646.75f, 98.165f, 646.751f, 98.176f);
path.lineTo(636.849f, 98.176f);
path.close();
path.moveTo(673.517f, 98.176f);
path.cubicTo(673.519f, 98.171f, 673.522f, 98.158f, 673.522f, 98.158f);
path.cubicTo(673.528f, 98.142f, 678.138f, 84.049f, 678.138f, 84.049f);
path.cubicTo(678.202f, 83.857f, 678.252f, 83.756f, 678.288f, 83.697f);
path.cubicTo(678.324f, 83.756f, 678.375f, 83.858f, 678.439f, 84.051f);
path.cubicTo(678.44f, 84.054f, 683.417f, 98.149f, 683.417f, 98.149f);
path.cubicTo(683.416f, 98.145f, 683.416f, 98.145f, 683.416f, 98.147f);
path.cubicTo(683.416f, 98.149f, 683.418f, 98.164f, 683.42f, 98.176f);
path.lineTo(673.517f, 98.176f);
path.close();
path.moveTo(702.086f, 115.256f);
path.lineTo(702.089f, 115.257f);
path.cubicTo(704.075f, 116.223f, 706.408f, 116.692f, 709.22f, 116.692f);
path.cubicTo(712.384f, 116.692f, 715.17f, 115.918f, 717.275f, 114.454f);
path.cubicTo(720.639f, 112.074f, 721.837f, 108.31f, 721.837f, 105.581f);
path.cubicTo(721.837f, 101.538f, 720.57f, 98.203f, 714.092f, 93.115f);
path.lineTo(712.625f, 91.966f);
path.cubicTo(707.567f, 87.838f, 706.331f, 86.155f, 706.331f, 83.396f);
path.cubicTo(706.331f, 79.927f, 708.819f, 77.595f, 712.519f, 77.595f);
path.cubicTo(716.925f, 77.595f, 718.41f, 79.437f, 718.843f, 80.229f);
path.cubicTo(719.131f, 80.754f, 719.442f, 82.122f, 719.494f, 82.739f);
path.cubicTo(719.543f, 83.13f, 719.608f, 83.65f, 720.167f, 83.65f);
path.cubicTo(720.789f, 83.65f, 720.789f, 82.704f, 720.789f, 82.195f);
path.cubicTo(720.789f, 79.413f, 720.918f, 77.758f, 720.973f, 77.052f);
path.lineTo(720.998f, 76.662f);
path.cubicTo(720.998f, 76.113f, 720.581f, 76.043f, 720.219f, 76.043f);
path.cubicTo(719.847f, 76.043f, 719.469f, 76.003f, 718.438f, 75.788f);
path.cubicTo(716.936f, 75.428f, 715.369f, 75.26f, 713.514f, 75.26f);
path.cubicTo(706.771f, 75.26f, 702.415f, 79.048f, 702.415f, 84.91f);
path.cubicTo(702.415f, 88.438f, 703.589f, 91.748f, 709.319f, 96.434f);
path.lineTo(711.676f, 98.365f);
path.cubicTo(716.332f, 102.196f, 717.607f, 104.172f, 717.607f, 107.564f);
path.cubicTo(717.607f, 110.888f, 715.121f, 114.253f, 710.371f, 114.253f);
path.cubicTo(708.428f, 114.253f, 703.808f, 113.754f, 702.868f, 109.132f);
path.cubicTo(702.716f, 108.424f, 702.716f, 107.718f, 702.716f, 107.25f);
path.cubicTo(702.716f, 106.94f, 702.716f, 106.422f, 702.094f, 106.422f);
path.cubicTo(701.519f, 106.422f, 701.455f, 107.116f, 701.421f, 107.488f);
path.cubicTo(701.421f, 107.495f, 701.375f, 108.117f, 701.375f, 108.117f);
path.cubicTo(701.293f, 109.171f, 701.158f, 110.933f, 701.158f, 113.044f);
path.cubicTo(701.16f, 114.432f, 701.197f, 114.813f, 702.086f, 115.256f);
path.close();
path.moveTo(501.274f, 129.973f);
path.lineTo(498.058f, 136.963f);
path.lineTo(498.62f, 136.963f);
path.lineTo(499.723f, 134.561f);
path.lineTo(503.124f, 134.561f);
path.lineTo(504.227f, 136.963f);
path.lineTo(504.788f, 136.963f);
path.lineTo(501.571f, 129.973f);
path.lineTo(501.274f, 129.973f);
path.close();
path.moveTo(499.926f, 134.114f);
path.lineTo(501.417f, 130.848f);
path.lineTo(502.918f, 134.114f);
path.lineTo(499.926f, 134.114f);
path.close();
path.moveTo(510.117f, 130.023f);
path.lineTo(507.677f, 130.023f);
path.lineTo(507.677f, 136.963f);
path.lineTo(510.209f, 136.963f);
path.cubicTo(512.966f, 136.963f, 513.916f, 135.101f, 513.916f, 133.493f);
path.cubicTo(513.916f, 131.967f, 513.078f, 130.023f, 510.117f, 130.023f);
path.close();
path.moveTo(510.178f, 136.515f);
path.lineTo(508.217f, 136.515f);
path.lineTo(508.217f, 130.471f);
path.lineTo(510.147f, 130.471f);
path.cubicTo(512.036f, 130.471f, 513.333f, 131.712f, 513.333f, 133.493f);
path.cubicTo(513.333f, 135.447f, 511.853f, 136.515f, 510.178f, 136.515f);
path.close();
path.moveTo(519.105f, 135.894f);
path.lineTo(516.634f, 130.023f);
path.lineTo(516.093f, 130.023f);
path.lineTo(519.024f, 137.034f);
path.lineTo(519.177f, 137.034f);
path.lineTo(522.098f, 130.023f);
path.lineTo(521.557f, 130.023f);
path.lineTo(519.105f, 135.894f);
path.close();
path.moveTo(526.817f, 129.973f);
path.lineTo(523.601f, 136.963f);
path.lineTo(524.162f, 136.963f);
path.lineTo(525.265f, 134.561f);
path.lineTo(528.665f, 134.561f);
path.lineTo(529.768f, 136.963f);
path.lineTo(530.33f, 136.963f);
path.lineTo(527.113f, 129.973f);
path.lineTo(526.817f, 129.973f);
path.close();
path.moveTo(525.469f, 134.114f);
path.lineTo(526.96f, 130.848f);
path.lineTo(528.461f, 134.114f);
path.lineTo(525.469f, 134.114f);
path.close();
path.moveTo(538.947f, 136.088f);
path.lineTo(533.668f, 130.024f);
path.lineTo(533.208f, 130.024f);
path.lineTo(533.208f, 136.964f);
path.lineTo(533.749f, 136.964f);
path.lineTo(533.749f, 130.899f);
path.lineTo(539.038f, 136.964f);
path.lineTo(539.487f, 136.964f);
path.lineTo(539.487f, 130.024f);
path.lineTo(538.946f, 130.024f);
path.lineTo(538.946f, 136.088f);
path.lineTo(538.947f, 136.088f);
path.close();
path.moveTo(543.41f, 133.503f);
path.cubicTo(543.41f, 131.743f, 544.717f, 130.43f, 546.453f, 130.43f);
path.cubicTo(547.28f, 130.43f, 548.067f, 130.644f, 548.934f, 131.102f);
path.lineTo(548.934f, 130.512f);
path.cubicTo(548.087f, 130.125f, 547.26f, 129.932f, 546.483f, 129.932f);
path.cubicTo(544.574f, 129.932f, 542.828f, 131.377f, 542.828f, 133.503f);
path.cubicTo(542.828f, 135.538f, 544.411f, 137.054f, 546.534f, 137.054f);
path.cubicTo(547.341f, 137.054f, 548.198f, 136.83f, 549.005f, 136.402f);
path.lineTo(549.005f, 135.843f);
path.cubicTo(548.055f, 136.341f, 547.31f, 136.555f, 546.523f, 136.555f);
path.cubicTo(544.707f, 136.556f, 543.41f, 135.294f, 543.41f, 133.503f);
path.close();
path.moveTo(552.079f, 136.963f);
path.lineTo(552.62f, 136.963f);
path.lineTo(552.62f, 130.023f);
path.lineTo(552.079f, 130.023f);
path.lineTo(552.079f, 136.963f);
path.close();
path.moveTo(561.984f, 136.088f);
path.lineTo(556.705f, 130.024f);
path.lineTo(556.245f, 130.024f);
path.lineTo(556.245f, 136.964f);
path.lineTo(556.786f, 136.964f);
path.lineTo(556.786f, 130.899f);
path.lineTo(562.075f, 136.964f);
path.lineTo(562.524f, 136.964f);
path.lineTo(562.524f, 130.024f);
path.lineTo(561.983f, 130.024f);
path.lineTo(561.983f, 136.088f);
path.lineTo(561.984f, 136.088f);
path.close();
path.moveTo(570.122f, 134.257f);
path.lineTo(571.592f, 134.257f);
path.lineTo(571.592f, 136.129f);
path.cubicTo(571.041f, 136.403f, 570.489f, 136.556f, 569.683f, 136.556f);
path.cubicTo(567.488f, 136.556f, 566.456f, 135.05f, 566.456f, 133.493f);
path.cubicTo(566.456f, 131.733f, 567.763f, 130.43f, 569.529f, 130.43f);
path.cubicTo(570.366f, 130.43f, 571.153f, 130.654f, 572, 131.132f);
path.lineTo(572, 130.522f);
path.cubicTo(571.02f, 130.084f, 570.397f, 129.932f, 569.61f, 129.932f);
path.cubicTo(567.455f, 129.932f, 565.873f, 131.438f, 565.873f, 133.473f);
path.cubicTo(565.873f, 135.529f, 567.496f, 137.055f, 569.661f, 137.055f);
path.cubicTo(570.468f, 137.055f, 571.305f, 136.862f, 572.132f, 136.465f);
path.lineTo(572.132f, 133.809f);
path.lineTo(570.121f, 133.809f);
path.lineTo(570.121f, 134.257f);
path.lineTo(570.122f, 134.257f);
path.close();
path.moveTo(580.681f, 131.753f);
path.cubicTo(580.681f, 131.001f, 581.314f, 130.43f, 582.151f, 130.43f);
path.cubicTo(582.631f, 130.43f, 582.998f, 130.552f, 583.713f, 130.959f);
path.lineTo(583.713f, 130.37f);
path.cubicTo(583.172f, 130.075f, 582.641f, 129.933f, 582.099f, 129.933f);
path.cubicTo(580.996f, 129.933f, 580.098f, 130.747f, 580.098f, 131.775f);
path.cubicTo(580.098f, 134.065f, 583.417f, 133.2f, 583.417f, 135.173f);
path.cubicTo(583.417f, 135.956f, 582.774f, 136.557f, 581.947f, 136.557f);
path.cubicTo(581.406f, 136.557f, 580.966f, 136.374f, 580.191f, 135.814f);
path.lineTo(580.191f, 136.445f);
path.cubicTo(580.804f, 136.852f, 581.406f, 137.055f, 581.978f, 137.055f);
path.cubicTo(583.111f, 137.055f, 584, 136.221f, 584, 135.152f);
path.cubicTo(584, 132.72f, 580.681f, 133.605f, 580.681f, 131.753f);
path.close();
path.moveTo(587.543f, 133.503f);
path.cubicTo(587.543f, 131.743f, 588.85f, 130.43f, 590.586f, 130.43f);
path.cubicTo(591.413f, 130.43f, 592.199f, 130.644f, 593.067f, 131.102f);
path.lineTo(593.067f, 130.512f);
path.cubicTo(592.22f, 130.125f, 591.393f, 129.932f, 590.616f, 129.932f);
path.cubicTo(588.707f, 129.932f, 586.96f, 131.377f, 586.96f, 133.503f);
path.cubicTo(586.96f, 135.538f, 588.543f, 137.054f, 590.667f, 137.054f);
path.cubicTo(591.473f, 137.054f, 592.331f, 136.83f, 593.138f, 136.402f);
path.lineTo(593.138f, 135.843f);
path.cubicTo(592.188f, 136.341f, 591.443f, 136.555f, 590.657f, 136.555f);
path.cubicTo(588.84f, 136.556f, 587.543f, 135.294f, 587.543f, 133.503f);
path.close();
path.moveTo(596.212f, 136.963f);
path.lineTo(596.753f, 136.963f);
path.lineTo(596.753f, 130.023f);
path.lineTo(596.212f, 130.023f);
path.lineTo(596.212f, 136.963f);
path.close();
path.moveTo(600.94f, 133.697f);
path.lineTo(604.116f, 133.697f);
path.lineTo(604.116f, 133.25f);
path.lineTo(600.94f, 133.25f);
path.lineTo(600.94f, 130.472f);
path.lineTo(604.198f, 130.472f);
path.lineTo(604.198f, 130.024f);
path.lineTo(600.399f, 130.024f);
path.lineTo(600.399f, 136.964f);
path.lineTo(604.279f, 136.964f);
path.lineTo(604.279f, 136.516f);
path.lineTo(600.94f, 136.516f);
path.lineTo(600.94f, 133.697f);
path.close();
path.moveTo(613.163f, 136.088f);
path.lineTo(607.884f, 130.024f);
path.lineTo(607.424f, 130.024f);
path.lineTo(607.424f, 136.964f);
path.lineTo(607.965f, 136.964f);
path.lineTo(607.965f, 130.899f);
path.lineTo(613.254f, 136.964f);
path.lineTo(613.703f, 136.964f);
path.lineTo(613.703f, 130.024f);
path.lineTo(613.162f, 130.024f);
path.lineTo(613.162f, 136.088f);
path.lineTo(613.163f, 136.088f);
path.close();
path.moveTo(617.625f, 133.503f);
path.cubicTo(617.625f, 131.743f, 618.932f, 130.43f, 620.668f, 130.43f);
path.cubicTo(621.495f, 130.43f, 622.281f, 130.644f, 623.149f, 131.102f);
path.lineTo(623.149f, 130.512f);
path.cubicTo(622.301f, 130.125f, 621.474f, 129.932f, 620.698f, 129.932f);
path.cubicTo(618.789f, 129.932f, 617.042f, 131.377f, 617.042f, 133.503f);
path.cubicTo(617.042f, 135.538f, 618.625f, 137.054f, 620.749f, 137.054f);
path.cubicTo(621.556f, 137.054f, 622.413f, 136.83f, 623.22f, 136.402f);
path.lineTo(623.22f, 135.843f);
path.cubicTo(622.27f, 136.341f, 621.525f, 136.555f, 620.739f, 136.555f);
path.cubicTo(618.922f, 136.556f, 617.625f, 135.294f, 617.625f, 133.503f);
path.close();
path.moveTo(626.856f, 133.697f);
path.lineTo(630.032f, 133.697f);
path.lineTo(630.032f, 133.25f);
path.lineTo(626.856f, 133.25f);
path.lineTo(626.856f, 130.472f);
path.lineTo(630.114f, 130.472f);
path.lineTo(630.114f, 130.024f);
path.lineTo(626.316f, 130.024f);
path.lineTo(626.316f, 136.964f);
path.lineTo(630.196f, 136.964f);
path.lineTo(630.196f, 136.516f);
path.lineTo(626.857f, 136.516f);
path.lineTo(626.857f, 133.697f);
path.lineTo(626.856f, 133.697f);
path.close();
path.moveTo(633.115f, 136.963f);
path.lineTo(633.33f, 136.963f);
path.cubicTo(633.33f, 137.309f, 633.299f, 137.523f, 633.115f, 137.797f);
path.lineTo(633.115f, 138.154f);
path.cubicTo(633.565f, 137.828f, 633.779f, 137.269f, 633.779f, 136.75f);
path.lineTo(633.779f, 135.976f);
path.lineTo(633.115f, 135.976f);
path.lineTo(633.115f, 136.963f);
path.close();
path.moveTo(641.511f, 131.753f);
path.cubicTo(641.511f, 131.001f, 642.144f, 130.43f, 642.981f, 130.43f);
path.cubicTo(643.461f, 130.43f, 643.828f, 130.552f, 644.544f, 130.959f);
path.lineTo(644.544f, 130.37f);
path.cubicTo(644.003f, 130.075f, 643.472f, 129.933f, 642.931f, 129.933f);
path.cubicTo(641.828f, 129.933f, 640.929f, 130.747f, 640.929f, 131.775f);
path.cubicTo(640.929f, 134.065f, 644.247f, 133.2f, 644.247f, 135.173f);
path.cubicTo(644.247f, 135.956f, 643.604f, 136.557f, 642.777f, 136.557f);
path.cubicTo(642.236f, 136.557f, 641.797f, 136.374f, 641.02f, 135.814f);
path.lineTo(641.02f, 136.445f);
path.cubicTo(641.633f, 136.852f, 642.235f, 137.055f, 642.807f, 137.055f);
path.cubicTo(643.941f, 137.055f, 644.829f, 136.221f, 644.829f, 135.152f);
path.cubicTo(644.829f, 132.72f, 641.511f, 133.605f, 641.511f, 131.753f);
path.close();
path.moveTo(648.649f, 133.697f);
path.lineTo(651.824f, 133.697f);
path.lineTo(651.824f, 133.25f);
path.lineTo(648.649f, 133.25f);
path.lineTo(648.649f, 130.472f);
path.lineTo(651.906f, 130.472f);
path.lineTo(651.906f, 130.024f);
path.lineTo(648.107f, 130.024f);
path.lineTo(648.107f, 136.964f);
path.lineTo(651.988f, 136.964f);
path.lineTo(651.988f, 136.516f);
path.lineTo(648.649f, 136.516f);
path.lineTo(648.649f, 133.697f);
path.close();
path.moveTo(657.91f, 133.503f);
path.cubicTo(658.799f, 133.177f, 659.237f, 132.618f, 659.237f, 131.814f);
path.cubicTo(659.237f, 130.888f, 658.635f, 130.023f, 656.929f, 130.023f);
path.lineTo(655.142f, 130.023f);
path.lineTo(655.142f, 136.963f);
path.lineTo(655.683f, 136.963f);
path.lineTo(655.683f, 133.635f);
path.lineTo(656.623f, 133.635f);
path.cubicTo(657.46f, 133.635f, 657.726f, 133.747f, 658.563f, 135.06f);
path.lineTo(659.778f, 136.963f);
path.lineTo(660.411f, 136.963f);
path.lineTo(658.838f, 134.561f);
path.cubicTo(658.513f, 134.073f, 658.237f, 133.676f, 657.91f, 133.503f);
path.close();
path.moveTo(656.838f, 133.188f);
path.lineTo(655.684f, 133.188f);
path.lineTo(655.684f, 130.472f);
path.lineTo(656.817f, 130.472f);
path.cubicTo(658.094f, 130.472f, 658.655f, 131.041f, 658.655f, 131.825f);
path.cubicTo(658.655f, 132.659f, 658.063f, 133.188f, 656.838f, 133.188f);
path.close();
path.moveTo(664.968f, 135.894f);
path.lineTo(662.497f, 130.023f);
path.lineTo(661.956f, 130.023f);
path.lineTo(664.887f, 137.034f);
path.lineTo(665.04f, 137.034f);
path.lineTo(667.961f, 130.023f);
path.lineTo(667.42f, 130.023f);
path.lineTo(664.968f, 135.894f);
path.close();
path.moveTo(670.829f, 136.963f);
path.lineTo(671.37f, 136.963f);
path.lineTo(671.37f, 130.023f);
path.lineTo(670.829f, 130.023f);
path.lineTo(670.829f, 136.963f);
path.close();
path.moveTo(680.734f, 136.088f);
path.lineTo(675.455f, 130.024f);
path.lineTo(674.995f, 130.024f);
path.lineTo(674.995f, 136.964f);
path.lineTo(675.536f, 136.964f);
path.lineTo(675.536f, 130.899f);
path.lineTo(680.826f, 136.964f);
path.lineTo(681.275f, 136.964f);
path.lineTo(681.275f, 130.024f);
path.lineTo(680.733f, 130.024f);
path.lineTo(680.733f, 136.088f);
path.lineTo(680.734f, 136.088f);
path.close();
path.moveTo(688.873f, 134.257f);
path.lineTo(690.343f, 134.257f);
path.lineTo(690.343f, 136.129f);
path.cubicTo(689.791f, 136.403f, 689.24f, 136.556f, 688.433f, 136.556f);
path.cubicTo(686.238f, 136.556f, 685.206f, 135.05f, 685.206f, 133.493f);
path.cubicTo(685.206f, 131.733f, 686.514f, 130.43f, 688.28f, 130.43f);
path.cubicTo(689.117f, 130.43f, 689.903f, 130.654f, 690.751f, 131.132f);
path.lineTo(690.751f, 130.522f);
path.cubicTo(689.771f, 130.084f, 689.148f, 129.932f, 688.362f, 129.932f);
path.cubicTo(686.207f, 129.932f, 684.624f, 131.438f, 684.624f, 133.473f);
path.cubicTo(684.624f, 135.529f, 686.247f, 137.055f, 688.413f, 137.055f);
path.cubicTo(689.219f, 137.055f, 690.057f, 136.862f, 690.884f, 136.465f);
path.lineTo(690.884f, 133.809f);
path.lineTo(688.873f, 133.809f);
path.lineTo(688.873f, 134.257f);
path.close();
path.moveTo(699.432f, 131.753f);
path.cubicTo(699.432f, 131.001f, 700.065f, 130.43f, 700.902f, 130.43f);
path.cubicTo(701.381f, 130.43f, 701.749f, 130.552f, 702.464f, 130.959f);
path.lineTo(702.464f, 130.37f);
path.cubicTo(701.923f, 130.075f, 701.391f, 129.933f, 700.85f, 129.933f);
path.cubicTo(699.747f, 129.933f, 698.849f, 130.747f, 698.849f, 131.775f);
path.cubicTo(698.849f, 134.065f, 702.168f, 133.2f, 702.168f, 135.173f);
path.cubicTo(702.168f, 135.956f, 701.525f, 136.557f, 700.698f, 136.557f);
path.cubicTo(700.156f, 136.557f, 699.718f, 136.374f, 698.941f, 135.814f);
path.lineTo(698.941f, 136.445f);
path.cubicTo(699.553f, 136.852f, 700.156f, 137.055f, 700.728f, 137.055f);
path.cubicTo(701.862f, 137.055f, 702.75f, 136.221f, 702.75f, 135.152f);
path.cubicTo(702.75f, 132.72f, 699.432f, 133.605f, 699.432f, 131.753f);
path.close();
path.moveTo(709.407f, 129.932f);
path.cubicTo(707.345f, 129.932f, 705.731f, 131.499f, 705.731f, 133.493f);
path.cubicTo(705.731f, 135.488f, 707.345f, 137.054f, 709.407f, 137.054f);
path.cubicTo(711.48f, 137.054f, 713.104f, 135.487f, 713.104f, 133.493f);
path.cubicTo(713.104f, 131.499f, 711.481f, 129.932f, 709.407f, 129.932f);
path.close();
path.moveTo(709.428f, 136.556f);
path.cubicTo(707.683f, 136.556f, 706.314f, 135.223f, 706.314f, 133.493f);
path.cubicTo(706.314f, 131.763f, 707.682f, 130.43f, 709.428f, 130.43f);
path.cubicTo(711.175f, 130.43f, 712.523f, 131.763f, 712.523f, 133.493f);
path.cubicTo(712.522f, 135.223f, 711.175f, 136.556f, 709.428f, 136.556f);
path.close();
path.moveTo(716.739f, 133.503f);
path.cubicTo(716.739f, 131.743f, 718.046f, 130.43f, 719.782f, 130.43f);
path.cubicTo(720.609f, 130.43f, 721.395f, 130.644f, 722.264f, 131.102f);
path.lineTo(722.264f, 130.512f);
path.cubicTo(721.416f, 130.125f, 720.589f, 129.932f, 719.813f, 129.932f);
path.cubicTo(717.903f, 129.932f, 716.157f, 131.377f, 716.157f, 133.503f);
path.cubicTo(716.157f, 135.538f, 717.74f, 137.054f, 719.863f, 137.054f);
path.cubicTo(720.67f, 137.054f, 721.528f, 136.83f, 722.335f, 136.402f);
path.lineTo(722.335f, 135.843f);
path.cubicTo(721.385f, 136.341f, 720.64f, 136.555f, 719.854f, 136.555f);
path.cubicTo(718.037f, 136.556f, 716.739f, 135.294f, 716.739f, 133.503f);
path.close();
path.moveTo(725.409f, 136.963f);
path.lineTo(725.95f, 136.963f);
path.lineTo(725.95f, 130.023f);
path.lineTo(725.409f, 130.023f);
path.lineTo(725.409f, 136.963f);
path.close();
path.moveTo(730.136f, 133.697f);
path.lineTo(733.312f, 133.697f);
path.lineTo(733.312f, 133.25f);
path.lineTo(730.136f, 133.25f);
path.lineTo(730.136f, 130.472f);
path.lineTo(733.393f, 130.472f);
path.lineTo(733.393f, 130.024f);
path.lineTo(729.595f, 130.024f);
path.lineTo(729.595f, 136.964f);
path.lineTo(733.475f, 136.964f);
path.lineTo(733.475f, 136.516f);
path.lineTo(730.136f, 136.516f);
path.lineTo(730.136f, 133.697f);
path.close();
path.moveTo(735.865f, 130.472f);
path.lineTo(738.479f, 130.472f);
path.lineTo(738.479f, 136.964f);
path.lineTo(739.02f, 136.964f);
path.lineTo(739.02f, 130.472f);
path.lineTo(741.634f, 130.472f);
path.lineTo(741.634f, 130.024f);
path.lineTo(735.865f, 130.024f);
path.lineTo(735.865f, 130.472f);
path.close();
path.moveTo(749.017f, 130.023f);
path.lineTo(746.699f, 132.771f);
path.lineTo(744.371f, 130.023f);
path.lineTo(743.717f, 130.023f);
path.lineTo(746.423f, 133.239f);
path.lineTo(746.423f, 136.963f);
path.lineTo(746.964f, 136.963f);
path.lineTo(746.964f, 133.239f);
path.lineTo(749.67f, 130.023f);
path.lineTo(749.017f, 130.023f);
path.close();
path.moveTo(578.773f, 115.549f);
path.lineTo(578.773f, 62.773f);
path.lineTo(557.571f, 62.773f);
path.cubicTo(562.752f, 67.658f, 575.797f, 106.652f, 578.773f, 115.549f);
path.close();
path.moveTo(544.754f, 62.861f);
path.lineTo(524.496f, 62.861f);
path.lineTo(524.496f, 115.55f);
path.lineTo(524.934f, 115.55f);
path.lineTo(544.544f, 64.169f);
path.lineTo(528.157f, 115.549f);
path.lineTo(528.682f, 115.549f);
path.lineTo(546.557f, 67.658f);
path.lineTo(531.675f, 115.548f);
path.lineTo(532.376f, 115.548f);
path.lineTo(548.57f, 71.846f);
path.lineTo(535.293f, 115.507f);
path.lineTo(535.906f, 115.507f);
path.lineTo(544.281f, 97.58f);
path.lineTo(542.267f, 97.58f);
path.lineTo(550.409f, 76.033f);
path.lineTo(544.281f, 97.578f);
path.lineTo(546.119f, 97.578f);
path.lineTo(538.678f, 115.551f);
path.lineTo(565.029f, 115.551f);
path.cubicTo(562.052f, 106.477f, 550.759f, 67.92f, 544.754f, 62.861f);
path.close();
SkPath path2(path);
testPathOp(reporter, path1, path2, kIntersect_SkPathOp, filename);
}
Convert NULL and 0 to nullptr. This was created by looking at warnings produced by clang's -Wzero-as-null-pointer-constant. This updates most issues in Skia code. However, there are places where GL and Vulkan want pointer values which are explicitly 0, external headers which use NULL directly, and possibly more uses in un-compiled sources (for other platforms). Change-Id: Id22fbac04d5c53497a53d734f0896b4f06fe8345 Reviewed-on: https://skia-review.googlesource.com/39521 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Ben Wagner <bungeman@google.com>
2017-08-28 14:34:05 +00:00
static void (*skipTest)(skiatest::Reporter* , const char* filename) = nullptr;
static void (*firstTest)(skiatest::Reporter* , const char* filename) = nullptr;
static void (*stopTest)(skiatest::Reporter* , const char* filename) = nullptr;
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
static struct TestDesc tests[] = {
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
TEST(skpwww_gorcraft_ru_1),
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
TEST(http___www_project2061_org),
TEST(http___www_emuleteca_cl_27),
TEST(http___www_emuleteca_cl_26),
TEST(http___www_emuleteca_cl_28),
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
TEST(skpwww_nimble_com_au_1),
TEST(skpwww_mybuilder_com_1),
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
TEST(skpwww_neda_net_2),
Enabling clip stack flattening exercises path ops. Iterating through the 903K skps that represent the imagable 1M top web pages triggers a number of bugs, some of which are addressed here. Some web pages trigger intersecting cubic representations of arc with their conic counterparts. This exposed a flaw in coincident detection that caused an infinite loop. The loop alternatively extended the coincident section and, determining the that the bounds of the curve pairs did not overlap, deleted the extension. Track the number of times the coincident detection is called, and if it exceeds an empirically found limit, assume that the curves are coincident and force it to be so. The loop count limit can be determined by enabling DEBUG_T_SECT_LOOP_COUNT and running all tests. The largest count is reported on completion. Another class of bugs was caused by concident detection duplicating nearly identical points that had been merged earlier. To track these bugs, the 'handle coincidence' code was duplicated as a const debug variety that reported if one of a dozen or so irregularities are present; then it is easier to see when a block of code that fixes one irregularity regresses another. Creating the debug const code version exposed some non-debug code that could be const, and some that was experimental and could be removed. Set DEBUG_COINCIDENCE to track coincidence health and handling. For running on Chrome, DEBUG_VERIFY checks the result of pathops against the same operation using SkRegion to verify that the results are nearly the same. When visualizing the pathops work using tools/pathops_visualizer.htm, set DEBUG_DUMP_ALIGNMENT to see the curves after they've been aligned for coincidence. Other bugs fixed include detecting when a section of a pair of curves have devolved into lines and are coincident. TBR=reed@google.com Review URL: https://codereview.chromium.org/1394503003
2015-10-16 16:03:38 +00:00
TEST(skpwww_woothemes_com_1),
TEST(skpwww_neda_net_1),
Reland of path ops: fix conic weight and partial coincidence (patchset #1 id:1 of https://codereview.chromium.org/1408923003/ ) Reason for revert: suppressions have landed in chrome Original issue's description: > Revert of path ops: fix conic weight and partial coincidence (patchset #5 id:80001 of https://codereview.chromium.org/1413763002/ ) > > Reason for revert: > path ops change breaks svg clipping layout tests -- conic is now more accurate, changing edge of circle in clip > > These need to be rebaselined > > svg/clip-path/clip-path-child-clipped.svg > svg/clip-path/clip-path-nonzero.svg > svg/clip-path/clip-path-evenodd-nonzero.svg > svg/clip-path/clip-path-nonzero-evenodd.svg > > Original issue's description: > > The remaining 1m skp bugs are asserts that can be harmlessly > > suppressed and bugs around conics. > > > > The conic calculation for a subdivided w was just wrong. > > > > Also added debugging to template intersection to initialize > > reused structures and dump additional data. > > > > TBR=reed@google.com > > > > Committed: https://skia.googlesource.com/skia/+/ef33b1e739b23a1201100ff17a572da85b03d9af > > TBR= > NOPRESUBMIT=true > NOTREECHECKS=true > NOTRY=true > > Committed: https://skia.googlesource.com/skia/+/f428df1be3e96d3f8970d0f7f415b862f7da5404 TBR= NOPRESUBMIT=true NOTREECHECKS=true NOTRY=true Review URL: https://codereview.chromium.org/1407003016
2015-10-22 14:23:52 +00:00
TEST(skpwww_tinytots_com_1),
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
TEST(skpwww_educationalcraft_com_4a),
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
TEST(skpwww_lptemp_com_3),
TEST(skpwww_shinydemos_com_5),
fix pathops skp-derived bugs; add more failing tests TBR= Author: caryclark@google.com Review URL: https://codereview.chromium.org/463883002
2014-08-12 14:46:33 +00:00
TEST(skpwww_lptemp_com_5),
TEST(skpwww_shinydemos_com_15),
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
TEST(skpwww_familysurvivalprotocol_wordpress_com_61),
TEST(skpwww_alamdi_com_3),
TEST(skpwww_devbridge_com_22),
fix pathops skp-derived bugs; add more failing tests TBR= Author: caryclark@google.com Review URL: https://codereview.chromium.org/463883002
2014-08-12 14:46:33 +00:00
TEST(skpwww_firstunitedbank_com_19),
TEST(skpwww_googleventures_com_32),
update pathops core and tests split out skpclip (the test of 1M pictures) into its own project TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/400033002
2014-07-18 12:08:14 +00:00
TEST(skpwww_9to5mac_com_64),
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
TEST(skpwww_wartepop_blogspot_com_br_6),
TEST(skpwww_wartepop_blogspot_com_br_6a),
TEST(skpwww_cooksnaps_com_32a),
TEST(skpwww_argus_presse_fr_41),
TEST(skpwww_cooksnaps_com_17),
TEST(skpwww_cooksnaps_com_32),
TEST(skpwww_kitcheninspirations_wordpress_com_66),
TEST(skpwww_tcmevents_org_13),
TEST(skpwww_narayana_publishers_com_194),
TEST(skpwww_swapspacesystems_com_5),
TEST(skpwww_vantageproduction_com_109),
TEST(skpwww_americascup_com_108),
TEST(skpwww_narayana_verlag_de_194),
TEST(skpwww_etiqadd_com_2464),
TEST(skpwww_paseoitaigara_com_br_56),
TEST(skpwww_mortgagemarketguide_com_109),
TEST(skpwww_aceinfographics_com_106),
TEST(skpwww_educationalcraft_com_4),
TEST(skpwww_kitcheninspirations_wordpress_com_32),
TEST(skpwww_artblart_com_8),
TEST(skpwww_docgelo_com_66),
TEST(skpwww_uniquefx_net_442),
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
TEST(skpwww_defense_studies_blogspot_com_64),
TEST(skpwww_kenlevine_blogspot_com_28),
TEST(skpwww_fashionscandal_com_94),
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
TEST(skpwww_thaienews_blogspot_com_36),
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
TEST(skpwww_galaxystwo_com_4),
TEST(skpwww_catingueiraonline_com_352),
TEST(skpwww_evolvehq_com_210),
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
TEST(skpwww_odia_com_br_26),
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
TEST(skpwww_lokado_de_173),
TEST(skpwww_seopack_blogspot_com_2153),
TEST(skpwww_partsdata_de_53),
TEST(skpwww_simplysaru_com_40),
TEST(skpwww_jessicaslens_wordpress_com_222),
TEST(skpwww_kpopexplorer_net_22),
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
TEST(skpwww_tunero_de_24),
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
TEST(skpwww_karnivool_com_au_11),
TEST(skpwww_pindosiya_com_99),
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
TEST(skpwww_contextualnewsfeeds_com_346),
TEST(skpwww_helha_be_109),
TEST(skpwww_phototransferapp_com_24),
TEST(skpwww_phototransferapp_com_24x),
TEST(skpwww_gruposejaumdivulgador_com_br_4),
fix minor skp-found bugs remove globals from pathops_unittest BUG=skia:2460 TBR=mtklein Author: caryclark@google.com Review URL: https://codereview.chromium.org/239563004 git-svn-id: http://skia.googlecode.com/svn/trunk@14378 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-25 12:59:11 +00:00
TEST(skpwww_hubbyscook_com_22),
TEST(skpwww_maturesupertube_com_21),
TEST(skpwww_getgold_jp_731),
TEST(skpwww_trashness_com_36),
TEST(skpwww_exystence_net_61),
TEST(skpwww_320kbps_net_2231),
TEST(skpwww_heartiste_wordpress_com_86),
TEST(skpwww_hairjobsearch_com_31),
TEST(skpwww_alucinados_net_101),
Mike R: please sanity check SkPostConfig.h Mike K: please sanity check Test.cpp and skia_test.cpp Feel free to look at the rest, but I don't expect any in depth review of path ops innards. Path Ops first iteration used QuickSort to order segments radiating from an intersection to compute the winding rule. This revision uses a circular sort instead. Breaking out the circular sort into its own long-lived structure (SkOpAngle) allows doing less work and provides a home for caching additional sorting data. The circle sort is more stable than the former sort, has a robust ordering and fewer exceptions. It finds unsortable ordering less often. It is less reliant on the initial curve tangent, using convex hulls instead whenever it can. Additional debug validation makes sure that the computed structures are self-consistent. A new visualization tool helps verify that the angle ordering is correct. The 70+M tests pass with this change on Windows, Mac, Linux 32 and Linux 64 in debug and release. R=mtklein@google.com, reed@google.com Author: caryclark@google.com Review URL: https://codereview.chromium.org/131103009 git-svn-id: http://skia.googlecode.com/svn/trunk@14183 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-04-14 17:08:59 +00:00
TEST(skpnamecheap_com_405),
TEST(skpelpais_com_18),
TEST(skpwww_cityads_ru_249),
TEST(skpwww_alrakoba_net_62),
TEST(skpwww_dealnews_com_315),
TEST(skpwww_inmotionhosting_com_9),
fix line intersection git-svn-id: http://skia.googlecode.com/svn/trunk@12197 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-08 18:00:01 +00:00
TEST(skpskpicture14),
TEST(skpskpicture15),
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
TEST(skpwww_meb_gov_tr_6),
TEST(skpwww_sciality_com_101),
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
TEST(skpwww_booking_com_68),
TEST(skpwww_despegar_com_mx_272),
TEST(skpwww_lavoixdunord_fr_11),
TEST(skppptv_com_62),
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
TEST(skppchappy_com_au102),
TEST(skpsciality_com161),
TEST(skpi_gino_com16),
Enabling the canvas bit to turn the clip stack into a flat replace exposed around 100 failures when testing the 800K skp set generated from the top 1M web sites. This fixes all but one of those failures. Major changes include: - Replace angle indices with angle pointers. This was motivated by the need to add angles later but not renumber existing angles. - Aggressive segment chase. When the winding is known on a segment, more aggressively passing that winding to adjacent segments allows fragmented data sets to succeed. - Line segments with ends nearly the same are treated as coincident first. - Transfer partial coincidence by observing that if segment A is partially coincident to B and C then B and C may be partially coincident. TBR=reed Author: caryclark@google.com Review URL: https://codereview.chromium.org/272153002
2014-06-17 12:15:38 +00:00
TEST(skpnaoxrane_ru23),
TEST(skptcmevents_org23),
TEST(skpredbullskatearcade_es16),
TEST(skpfinanzasdigital_com9),
pathops work in progress BUG= Review URL: https://codereview.chromium.org/52653002 git-svn-id: http://skia.googlecode.com/svn/trunk@12089 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-11-01 17:36:03 +00:00
TEST(skpgithub_io_26),
TEST(skpgithub_io_25),
TEST(skpwww_meb_gov_tr_5),
TEST(skpwww_sciality_com_100),
TEST(skpwww_joomla_org_23),
TEST(skpwww_macrumors_com_131),
TEST(skpwww_briian_com_34),
TEST(skpwww_leadpages_net_84),
TEST(skpwww_fj_p_com_22),
TEST(skppartainasdemo250_org56),
TEST(skpsd_graphic_net104),
TEST(skpbreakmystyle_com10),
TEST(skpnational_com_au81),
TEST(skprentacheat_com30),
TEST(skptracksflow_com9),
TEST(skpautobutler_dk29),
TEST(skponlinecollege_org144),
TEST(skphostloco_com11),
TEST(skpsergeychunkevich_com8),
TEST(skpmlk_com326),
TEST(skpcyclist_friends_gr52),
TEST(skpfrauen_magazin_com83),
TEST(skpthesuburbanite_com213),
TEST(skpsudoestenegocios_com186),
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
TEST(skpmtrk_uz27),
TEST(skpilkoora_com37),
TEST(skpmm4everfriends_com43),
TEST(skpflite_com41),
TEST(skpcheeseandburger_com225),
TEST(skpeverytechpro_blogspot_com100),
};
static const size_t testCount = SK_ARRAY_COUNT(tests);
static bool runReverse = false;
Get rid of DEFINE_TESTCLASS_SHORT() macro. Instead tests should be written using DEF_TEST() macro, which is much nicer and simplifies the process of setting up an unit test. BUG=None TEST=skpskgr_test, pathops_unittest R=mtklein@google.com Review URL: https://codereview.chromium.org/117863005 git-svn-id: http://skia.googlecode.com/svn/trunk@12870 2bbb7eff-a529-9590-31e7-b0007b416f81
2014-01-02 21:45:03 +00:00
DEF_TEST(PathOpsSkp, reporter) {
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
#if DEBUG_SHOW_TEST_NAME
strncpy(DEBUG_FILENAME_STRING, "", DEBUG_FILENAME_STRING_LENGTH);
#endif
cumulative pathops patch Replace the implicit curve intersection with a geometric curve intersection. The implicit intersection proved mathematically unstable and took a long time to zero in on an answer. Use pointers instead of indices to refer to parts of curves. Indices required awkward renumbering. Unify t and point values so that small intervals can be eliminated in one pass. Break cubics up front to eliminate loops and cusps. Make the Simplify and Op code more regular and eliminate arbitrary differences. Add a builder that takes an array of paths and operators. Delete unused code. BUG=skia:3588 R=reed@google.com Review URL: https://codereview.chromium.org/1037573004
2015-03-26 14:52:43 +00:00
RunTestSet(reporter, tests, testCount, firstTest, skipTest, stopTest, runReverse);
path ops work in progress make more skps work remove edit files BUG= Review URL: https://codereview.chromium.org/23542056 git-svn-id: http://skia.googlecode.com/svn/trunk@11570 2bbb7eff-a529-9590-31e7-b0007b416f81
2013-10-02 14:49:34 +00:00
}
Reference in New Issue Copy Permalink