AuroraMiddleware/skia2
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
59a1df225d
skia2/tests/TessellatingPathRendererTests.cpp

536 lines
22 KiB
C++
Raw Normal View History

Tessellating GPU path renderer. This path renderer converts paths to linear contours, resolves intersections via Bentley-Ottman, implements a trapezoidal decomposition a la Fournier and Montuno to produce triangles, and renders those with a single draw call. It does not currently do antialiasing, so it must be used in conjunction with multisampling. A fair amount of the code is to handle floating point edge cases in intersections. Rather than perform exact computations (which would require arbitrary precision arithmetic), we reconnect the mesh to reflect the intersection points. For example, intersections can occur above the current vertex, and force edges to be merged into the current vertex, requiring a restart of the intersections. Splitting edges for intersections can also force them to merge with formerly-distinct edges in the same polygon, or to violate the ordering of the active edge list, or the active edge state of split edges. BUG=skia: Review URL: https://codereview.chromium.org/855513004
2015-02-26 14:58:17 +00:00
/*
* Copyright 2015 Google Inc.
*
* Use of this source code is governed by a BSD-style license that can be
* found in the LICENSE file.
*/
Remove GrTestTarget Change-Id: I4d2e1aa54732ff75b44a197255035c0e04548219 Reviewed-on: https://skia-review.googlesource.com/7327 Commit-Queue: Robert Phillips <robertphillips@google.com> Reviewed-by: Brian Salomon <bsalomon@google.com>
2017-01-20 14:11:37 +00:00
#include "Test.h"
Speculative fix for ASAN build. Check for NULL GrContext in TessellatingPathRendererTests. Also fix an upload nag (no #includes before GPU #ifdef). TBR=bsalomon@google.com Review URL: https://codereview.chromium.org/1097943004
2015-04-20 22:00:06 +00:00
#include "SkPath.h"
Tessellating GPU path renderer. This path renderer converts paths to linear contours, resolves intersections via Bentley-Ottman, implements a trapezoidal decomposition a la Fournier and Montuno to produce triangles, and renders those with a single draw call. It does not currently do antialiasing, so it must be used in conjunction with multisampling. A fair amount of the code is to handle floating point edge cases in intersections. Rather than perform exact computations (which would require arbitrary precision arithmetic), we reconnect the mesh to reflect the intersection points. For example, intersections can occur above the current vertex, and force edges to be merged into the current vertex, requiring a restart of the intersections. Splitting edges for intersections can also force them to merge with formerly-distinct edges in the same polygon, or to violate the ordering of the active edge list, or the active edge state of split edges. BUG=skia: Review URL: https://codereview.chromium.org/855513004
2015-02-26 14:58:17 +00:00
#if SK_SUPPORT_GPU
Make GrClip.h private BUG=skia: Change-Id: I4aa3c2707811ece3a63d161035e316c3bbc1cd15 Reviewed-on: https://skia-review.googlesource.com/9482 Commit-Queue: Brian Salomon <bsalomon@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2017-03-09 14:03:58 +00:00
#include "GrClip.h"
Generate list of GPU contexts outside tests Use DEF_GPUTEST_FOR_*_CONTEXT macros to obtain the test GPU context. Makes changing the context -related classes easier, since not all tests need to be changed. BUG=skia:2992 Review URL: https://codereview.chromium.org/1448873002
2015-12-01 12:35:26 +00:00
#include "GrContext.h"
GrTessellator: when sanitizing contours, remove non-finite points. NaNs can't be correctly sorted. Don't run the unit test on VK backend, since it requires large (>64K bytes) vertex buffer uploads. Bug: 757650 Change-Id: I667693f135a090a5d9076bb7a2ec6879fc06d645 Reviewed-on: https://skia-review.googlesource.com/37484 Commit-Queue: Stephen White <senorblanco@chromium.org> Reviewed-by: Greg Daniel <egdaniel@google.com>
2017-08-23 17:56:07 +00:00
#include "GrContextPriv.h"
GrTessellator (AA): null GrGeometryProcessor crash fix. Add a null-check on geometry processor creation. Add a test which exercises it: AA tessellator, with a non-invertible matrix and a fragment processor which needs local coords (e.g., linear gradient). BUG=691902 Change-Id: I005b893aed58d3ad2500c41501045ac94b0b4b95 Reviewed-on: https://skia-review.googlesource.com/8462 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-02-15 16:45:16 +00:00
#include "SkGradientShader.h"
Reland of SkShaderBase Introduce a private base class (SkShaderBase), to hide implementation details from the public interface (SkShader). Change-Id: Ib1d76cde880bd51868b97408710f8bb38128e536 Reviewed-on: https://skia-review.googlesource.com/17925 Reviewed-by: Florin Malita <fmalita@chromium.org> Commit-Queue: Florin Malita <fmalita@chromium.org>
2017-05-25 14:38:07 +00:00
#include "SkShaderBase.h"
Clean up GrResourceProvider usage The only substantive changes are the removal of GrProxy instantiation in: SkGpuBlurUtils::GaussianBlur GrSimpleTextureEffect::Make* Change-Id: I10970609693bd6ff5b3a3c21b41d82642bb277bc Reviewed-on: https://skia-review.googlesource.com/19965 Reviewed-by: Greg Daniel <egdaniel@google.com> Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Robert Phillips <robertphillips@google.com>
2017-06-15 16:07:18 +00:00
#include "effects/GrPorterDuffXferProcessor.h"
Make GrClip.h private BUG=skia: Change-Id: I4aa3c2707811ece3a63d161035e316c3bbc1cd15 Reviewed-on: https://skia-review.googlesource.com/9482 Commit-Queue: Brian Salomon <bsalomon@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2017-03-09 14:03:58 +00:00
#include "ops/GrTessellatingPathRenderer.h"
Tessellating GPU path renderer. This path renderer converts paths to linear contours, resolves intersections via Bentley-Ottman, implements a trapezoidal decomposition a la Fournier and Montuno to produce triangles, and renders those with a single draw call. It does not currently do antialiasing, so it must be used in conjunction with multisampling. A fair amount of the code is to handle floating point edge cases in intersections. Rather than perform exact computations (which would require arbitrary precision arithmetic), we reconnect the mesh to reflect the intersection points. For example, intersections can occur above the current vertex, and force edges to be merged into the current vertex, requiring a restart of the intersections. Splitting edges for intersections can also force them to merge with formerly-distinct edges in the same polygon, or to violate the ordering of the active edge list, or the active edge state of split edges. BUG=skia: Review URL: https://codereview.chromium.org/855513004
2015-02-26 14:58:17 +00:00
/*
* These tests pass by not crashing, hanging or asserting in Debug.
*/
// Tests active edges made inactive by splitting.
// Also tests active edge list forced into an invalid ordering by
// splitting (mopped up in cleanup_active_edges()).
static SkPath create_path_0() {
SkPath path;
path.moveTo(229.127044677734375f, 67.34100341796875f);
path.lineTo(187.8097381591796875f, -6.7729740142822265625f);
path.lineTo(171.411407470703125f, 50.94266510009765625f);
path.lineTo(245.5253753662109375f, 9.6253643035888671875f);
path.moveTo(208.4683990478515625f, 30.284009933471679688f);
path.lineTo(171.411407470703125f, 50.94266510009765625f);
path.lineTo(187.8097381591796875f, -6.7729740142822265625f);
return path;
}
// Intersections which fall exactly on the current vertex, and require
// a restart of the intersection checking.
static SkPath create_path_1() {
SkPath path;
path.moveTo(314.483551025390625f, 486.246002197265625f);
path.lineTo(385.41949462890625f, 532.8087158203125f);
path.lineTo(373.232879638671875f, 474.05938720703125f);
path.lineTo(326.670166015625f, 544.995361328125f);
path.moveTo(349.951507568359375f, 509.52734375f);
path.lineTo(373.232879638671875f, 474.05938720703125f);
path.lineTo(385.41949462890625f, 532.8087158203125f);
return path;
}
// Tests active edges which are removed by splitting.
static SkPath create_path_2() {
SkPath path;
path.moveTo(343.107391357421875f, 613.62176513671875f);
path.lineTo(426.632415771484375f, 628.5740966796875f);
path.lineTo(392.3460693359375f, 579.33544921875f);
path.lineTo(377.39373779296875f, 662.86041259765625f);
path.moveTo(384.869873046875f, 621.097900390625f);
path.lineTo(392.3460693359375f, 579.33544921875f);
path.lineTo(426.632415771484375f, 628.5740966796875f);
return path;
}
// Collinear edges merged in set_top().
// Also, an intersection between left and right enclosing edges which
// falls above the current vertex.
static SkPath create_path_3() {
SkPath path;
path.moveTo(545.95751953125f, 791.69854736328125f);
path.lineTo(612.05816650390625f, 738.494140625f);
path.lineTo(552.4056396484375f, 732.0460205078125f);
path.lineTo(605.61004638671875f, 798.14666748046875f);
path.moveTo(579.00787353515625f, 765.0963134765625f);
path.lineTo(552.4056396484375f, 732.0460205078125f);
path.lineTo(612.05816650390625f, 738.494140625f);
return path;
}
// Tests active edges which are made inactive by set_top().
static SkPath create_path_4() {
SkPath path;
path.moveTo(819.2725830078125f, 751.77447509765625f);
path.lineTo(820.70904541015625f, 666.933837890625f);
path.lineTo(777.57049560546875f, 708.63592529296875f);
path.lineTo(862.4111328125f, 710.0723876953125f);
path.moveTo(819.99078369140625f, 709.3541259765625f);
path.lineTo(777.57049560546875f, 708.63592529296875f);
path.lineTo(820.70904541015625f, 666.933837890625f);
return path;
}
static SkPath create_path_5() {
SkPath path;
path.moveTo(823.33209228515625f, 749.052734375f);
path.lineTo(823.494873046875f, 664.20013427734375f);
path.lineTo(780.9871826171875f, 706.5450439453125f);
path.lineTo(865.8397216796875f, 706.70782470703125f);
path.moveTo(823.4134521484375f, 706.6263427734375f);
path.lineTo(780.9871826171875f, 706.5450439453125f);
path.lineTo(823.494873046875f, 664.20013427734375f);
return path;
}
static SkPath create_path_6() {
SkPath path;
path.moveTo(954.862548828125f, 562.8349609375f);
path.lineTo(899.32818603515625f, 498.679443359375f);
path.lineTo(895.017578125f, 558.52435302734375f);
path.lineTo(959.17315673828125f, 502.990081787109375f);
path.moveTo(927.0953369140625f, 530.7572021484375f);
path.lineTo(895.017578125f, 558.52435302734375f);
path.lineTo(899.32818603515625f, 498.679443359375f);
return path;
}
static SkPath create_path_7() {
SkPath path;
path.moveTo(958.5330810546875f, 547.35516357421875f);
path.lineTo(899.93109130859375f, 485.989013671875f);
path.lineTo(898.54901123046875f, 545.97308349609375f);
path.lineTo(959.9151611328125f, 487.37109375f);
path.moveTo(929.2320556640625f, 516.67205810546875f);
path.lineTo(898.54901123046875f, 545.97308349609375f);
path.lineTo(899.93109130859375f, 485.989013671875f);
return path;
}
static SkPath create_path_8() {
SkPath path;
path.moveTo(389.8609619140625f, 369.326873779296875f);
path.lineTo(470.6290283203125f, 395.33697509765625f);
path.lineTo(443.250030517578125f, 341.9478759765625f);
path.lineTo(417.239959716796875f, 422.7159423828125f);
path.moveTo(430.244964599609375f, 382.3319091796875f);
path.lineTo(443.250030517578125f, 341.9478759765625f);
path.lineTo(470.6290283203125f, 395.33697509765625f);
return path;
}
static SkPath create_path_9() {
SkPath path;
path.moveTo(20, 20);
path.lineTo(50, 80);
path.lineTo(20, 80);
path.moveTo(80, 50);
path.lineTo(50, 50);
path.lineTo(20, 50);
return path;
}
static SkPath create_path_10() {
SkPath path;
path.moveTo(257.19439697265625f, 320.876617431640625f);
path.lineTo(190.113037109375f, 320.58978271484375f);
path.lineTo(203.64404296875f, 293.8145751953125f);
path.moveTo(203.357177734375f, 360.896026611328125f);
path.lineTo(216.88824462890625f, 334.120819091796875f);
path.lineTo(230.41925048828125f, 307.345611572265625f);
return path;
}
// A degenerate segments case, where both upper and lower segments of
// a split edge must remain active.
static SkPath create_path_11() {
SkPath path;
path.moveTo(231.9331207275390625f, 306.2012939453125f);
path.lineTo(191.4859161376953125f, 306.04547119140625f);
path.lineTo(231.0659332275390625f, 300.2642822265625f);
path.moveTo(189.946807861328125f, 302.072265625f);
path.lineTo(179.79705810546875f, 294.859771728515625f);
path.lineTo(191.0016021728515625f, 296.165679931640625f);
path.moveTo(150.8942108154296875f, 304.900146484375f);
path.lineTo(179.708892822265625f, 297.849029541015625f);
path.lineTo(190.4742279052734375f, 299.11895751953125f);
return path;
}
// Handle the case where edge.dist(edge.fTop) != 0.0.
static SkPath create_path_12() {
SkPath path;
path.moveTo( 0.0f, 400.0f);
path.lineTo( 138.0f, 202.0f);
path.lineTo( 0.0f, 202.0f);
path.moveTo( 12.62693023681640625f, 250.57464599609375f);
path.lineTo( 8.13896942138671875f, 254.556884765625f);
path.lineTo(-18.15641021728515625f, 220.40203857421875f);
path.lineTo(-15.986493110656738281f, 219.6513519287109375f);
path.moveTo( 36.931194305419921875f, 282.485504150390625f);
path.lineTo( 15.617521286010742188f, 261.2901611328125f);
path.lineTo( 10.3829498291015625f, 252.565765380859375f);
path.lineTo(-16.165292739868164062f, 222.646026611328125f);
return path;
}
// A degenerate segments case which exercises inactive edges being
// made active by splitting.
static SkPath create_path_13() {
SkPath path;
path.moveTo(690.62127685546875f, 509.25555419921875f);
path.lineTo(99.336181640625f, 511.71405029296875f);
path.lineTo(708.362548828125f, 512.4349365234375f);
path.lineTo(729.9940185546875f, 516.3114013671875f);
path.lineTo(738.708984375f, 518.76995849609375f);
path.lineTo(678.3463134765625f, 510.0819091796875f);
path.lineTo(681.21795654296875f, 504.81378173828125f);
path.moveTo(758.52764892578125f, 521.55963134765625f);
path.lineTo(719.1549072265625f, 514.50372314453125f);
path.lineTo(689.59063720703125f, 512.0628662109375f);
path.lineTo(679.78216552734375f, 507.447845458984375f);
return path;
}
// Tests vertices which become "orphaned" (ie., no connected edges)
// after simplification.
static SkPath create_path_14() {
SkPath path;
path.moveTo(217.326019287109375f, 166.4752960205078125f);
path.lineTo(226.279266357421875f, 170.929473876953125f);
path.lineTo(234.3973388671875f, 177.0623626708984375f);
path.lineTo(262.0921630859375f, 188.746124267578125f);
path.moveTo(196.23638916015625f, 174.0722198486328125f);
path.lineTo(416.15277099609375f, 180.138214111328125f);
path.lineTo(192.651947021484375f, 304.0228271484375f);
return path;
}
Fix for out-of-bounds intersection (found by fuzzer). Sometimes, the intersection returned by check_intersection() is out-of-bounds for both edges (above both tops or below both bottoms) due to floating-point inaccuracy. This causes split_edge() to create a tiny negative-length edge on one side (which would then assert). Although we could safely remove this assert and allow the negative length edge to be removed, it's faster/safer to simply avoid its creation in the first place by adjusting one edge to the other edge's endpoint. Added a new unit test to exercise this case. Review URL: https://codereview.chromium.org/968993002
2015-03-02 17:34:13 +00:00
static SkPath create_path_15() {
SkPath path;
path.moveTo( 0.0f, 0.0f);
path.lineTo(10000.0f, 0.0f);
path.lineTo( 0.0f, -1.0f);
path.lineTo(10000.0f, 0.000001f);
path.lineTo( 0.0f, -30.0f);
return path;
}
Fix assert caused by floating point error in tessellating path renderer. In rare cases, floating point error causes the tesselator to add the same Vertex to more than one Poly on the same side. This was not a big problem when we were allocating new vertices when constructing Polys, but after https://codereview.chromium.org/2029243002 it causes more serious issues, since each Edge can only belong to two Polys, and violating this condition messes up the linked list of Edges used for left & right Polys and the associated estimated vertex count. The fix is to simply let the first Poly win, and skip that vertex for subsequent Polys. Since this only occurs in cases where vertices are very close to each other, it should have little visual effect. This is also exercised by Nebraska-StateSeal.svg. BUG=skia:5636 GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2259493002 Review-Url: https://codereview.chromium.org/2259493002
2016-08-17 21:56:22 +00:00
// Reduction of Nebraska-StateSeal.svg. Floating point error causes the
// same edge to be added to more than one poly on the same side.
static SkPath create_path_16() {
SkPath path;
path.moveTo(170.8199920654296875, 491.86700439453125);
path.lineTo(173.7649993896484375, 489.7340087890625);
path.lineTo(174.1450958251953125, 498.545989990234375);
path.lineTo( 171.998992919921875, 500.88201904296875);
path.moveTo(168.2922515869140625, 498.66265869140625);
path.lineTo(169.8589935302734375, 497.94500732421875);
path.lineTo( 172, 500.88299560546875);
path.moveTo( 169.555267333984375, 490.70111083984375);
path.lineTo(173.7649993896484375, 489.7340087890625);
path.lineTo( 170.82000732421875, 491.86700439453125);
return path;
}
GrTessellator (AA): null GrGeometryProcessor crash fix. Add a null-check on geometry processor creation. Add a test which exercises it: AA tessellator, with a non-invertible matrix and a fragment processor which needs local coords (e.g., linear gradient). BUG=691902 Change-Id: I005b893aed58d3ad2500c41501045ac94b0b4b95 Reviewed-on: https://skia-review.googlesource.com/8462 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-02-15 16:45:16 +00:00
// A simple concave path. Test this with a non-invertible matrix.
static SkPath create_path_17() {
SkPath path;
path.moveTo(20, 20);
path.lineTo(80, 20);
path.lineTo(30, 30);
path.lineTo(20, 80);
return path;
}
GrTessellator: fix for vertex coincident with enclosing edge. If a previously-enclosing edge coincides exactly with the current vertex, there are no two adjacent edges which enclose the vertex. Since find_enclosing_edges() ensures that the left enclosing edge is to the left of the vertex, the fix is to split the right enclosing edge on the current vertex and restart intersection tests. Bug: 716720 Change-Id: Id26c5b92a6d6139f348e99554638cded37e81a8e Reviewed-on: https://skia-review.googlesource.com/15261 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-05-03 20:00:38 +00:00
// A shape with a vertex collinear to the right hand edge.
// This messes up find_enclosing_edges.
static SkPath create_path_18() {
SkPath path;
path.moveTo(80, 20);
path.lineTo(80, 60);
path.lineTo(20, 60);
path.moveTo(80, 50);
path.lineTo(80, 80);
path.lineTo(20, 80);
return path;
}
GrTessellator: fix for three edges becoming collinear after splitting. In rare cases, a single edge can become collinear with two adjacent neighbour edges after it's split. The solution is to continue to merge until no collinear edges are found. BUG=722000 Change-Id: Ia5dd212b7acfb40ed1d6c74ebfa9e4a4746fe40a Reviewed-on: https://skia-review.googlesource.com/17963 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-05-25 18:47:11 +00:00
// Exercises the case where an edge becomes collinear with *two* of its
// adjacent neighbour edges after splitting.
// This is a reduction from
// http://mooooo.ooo/chebyshev-sine-approximation/horner_ulp.svg
static SkPath create_path_19() {
SkPath path;
path.moveTo( 351.99298095703125, 348.23046875);
path.lineTo( 351.91876220703125, 347.33984375);
path.lineTo( 351.91876220703125, 346.1953125);
path.lineTo( 351.90313720703125, 347.734375);
path.lineTo( 351.90313720703125, 346.1328125);
path.lineTo( 351.87579345703125, 347.93359375);
path.lineTo( 351.87579345703125, 345.484375);
path.lineTo( 351.86407470703125, 347.7890625);
path.lineTo( 351.86407470703125, 346.2109375);
path.lineTo( 351.84844970703125, 347.63763427734375);
path.lineTo( 351.84454345703125, 344.19232177734375);
path.lineTo( 351.78204345703125, 346.9483642578125);
path.lineTo( 351.758636474609375, 347.18310546875);
path.lineTo( 351.75469970703125, 346.75);
path.lineTo( 351.75469970703125, 345.46875);
path.lineTo( 352.5546875, 345.46875);
path.lineTo( 352.55078125, 347.01953125);
path.lineTo( 351.75079345703125, 347.02313232421875);
path.lineTo( 351.74688720703125, 346.15203857421875);
path.lineTo( 351.74688720703125, 347.646148681640625);
path.lineTo( 352.5390625, 346.94140625);
path.lineTo( 351.73907470703125, 346.94268798828125);
path.lineTo( 351.73516845703125, 344.48565673828125);
path.lineTo( 352.484375, 346.73828125);
path.lineTo( 351.68438720703125, 346.7401123046875);
path.lineTo( 352.4765625, 346.546875);
path.lineTo( 351.67657470703125, 346.54937744140625);
path.lineTo( 352.47265625, 346.75390625);
path.lineTo( 351.67266845703125, 346.756622314453125);
path.lineTo( 351.66876220703125, 345.612091064453125);
return path;
}
GrTessellator: fix logging; fix test comment. Change-Id: Iaded9025a1518d8c69dbe366deb1035e0bd4295b Reviewed-on: https://skia-review.googlesource.com/19289 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-06-09 14:06:26 +00:00
// An intersection above the first vertex in the mesh.
// Reduction from http://crbug.com/730687
GrTessellator: fix intersection above the first vertex. Handle the case where the an intersection falls not only above both edge endpoints, but above the first vertex in the mesh. This requires passing the mesh into check_for_intersection(), in order to modify the head. We also need to rewind the mesh after insertion, since we need to rewind to the newly-inserted vertex. This also cleans up vertex ID computation a little (for logging), so that vertices before the first vertex or after the last have a reasonable ID. It also cleans up the intersection-on-endpoint special cases by refactoring the calls to split_edge(). BUG=730687 Change-Id: Idea736eca7b7c3c5d8a470b1373a16ad8e649e80 Reviewed-on: https://skia-review.googlesource.com/19069 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-06-08 18:41:01 +00:00
static SkPath create_path_20() {
SkPath path;
path.moveTo( 2822128.5, 235.026336669921875);
path.lineTo( 2819349.25, 235.3623504638671875);
path.lineTo( -340558688, 23.83478546142578125);
path.lineTo( -340558752, 25.510419845581054688);
path.lineTo( -340558720, 27.18605804443359375);
return path;
}
GrTessellator: fix two NaN issues. If a point in the path rounds to +inf/-inf, the intersection code can produce NaN, which is unsortable. Fix: ignore non-finite intersections. Quadratic interpolation can sometimes produce NaN, which will never satisfy the flatness criterion. Abort if any of the interpolated points are non-finite. Bug:732023 Change-Id: If5881796e589c75b8f74459f42d00918619713a2 Reviewed-on: https://skia-review.googlesource.com/19467 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-06-12 15:43:18 +00:00
// An intersection whose result is NaN (due to rounded-to-inf endpoint).
static SkPath create_path_21() {
SkPath path;
path.moveTo(1.7889142061167663539e+38, 39338463358011572224.0);
path.lineTo( 1647.4193115234375, -522.603515625);
path.lineTo( 1677.74560546875, -529.0028076171875);
path.lineTo( 1678.29541015625, -528.7847900390625);
path.lineTo( 1637.5167236328125, -519.79266357421875);
path.lineTo( 1647.4193115234375, -522.603515625);
return path;
}
// A quad which becomes NaN when interpolated.
static SkPath create_path_22() {
SkPath path;
path.moveTo(-5.71889e+13f, 1.36759e+09f);
path.quadTo(2.45472e+19f, -3.12406e+15f, -2.19589e+18f, 2.79462e+14f);
return path;
}
GrTessellator: fix for collinear out-of-range intersections. Change 3b5a3fa8b1c11d4bd4499b040311f4c3553ebf8c introduced support for out-of-range intersections, which is necessary when the top and bottom vertices of an edge differ by only one ULP in the primary sort order and can't be split in-order. However, some out-of-range intersections produce edges which cancel each other out on splitting, in particular when the intersection is collinear with the newly-computed edge. This undoes the effect of the split. The tessellator then rewinds, re-detects the intersection, resplits, an infinite loop. The fix is to check for out-of-range intersections which are also collinear, and ignore them. This is ok, because these are not the cases we care about it change 3b5a3f above, which are never collinear. Bug: 753867 Change-Id: I590231e0e6f19c98f1ccf46cb7acc8a63ba35a9d Reviewed-on: https://skia-review.googlesource.com/34925 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-08-15 21:01:15 +00:00
// A path which contains out-of-range colinear intersections.
static SkPath create_path_23() {
SkPath path;
path.moveTo( 0, 63.39080047607421875);
path.lineTo(-0.70804601907730102539, 63.14350128173828125);
path.lineTo(-7.8608899287380243391e-17, 64.14080047607421875);
path.moveTo( 0, 64.14080047607421875);
path.lineTo(44.285900115966796875, 64.14080047607421875);
path.lineTo( 0, 62.64080047607421875);
path.moveTo(21.434900283813476562, -0.24732701480388641357);
path.lineTo(-0.70804601907730102539, 63.14350128173828125);
path.lineTo(0.70804601907730102539, 63.6381988525390625);
return path;
}
GrTessellator: when sanitizing contours, remove non-finite points. NaNs can't be correctly sorted. Don't run the unit test on VK backend, since it requires large (>64K bytes) vertex buffer uploads. Bug: 757650 Change-Id: I667693f135a090a5d9076bb7a2ec6879fc06d645 Reviewed-on: https://skia-review.googlesource.com/37484 Commit-Queue: Stephen White <senorblanco@chromium.org> Reviewed-by: Greg Daniel <egdaniel@google.com>
2017-08-23 17:56:07 +00:00
// A path which results in infs and nans when conics are converted to quads.
static SkPath create_path_24() {
SkPath path;
path.moveTo(-2.20883e+37f, -1.02892e+37f);
path.conicTo(-2.00958e+38f, -9.36107e+37f, -1.7887e+38f, -8.33215e+37f, 0.707107f);
path.conicTo(-1.56782e+38f, -7.30323e+37f, 2.20883e+37f, 1.02892e+37f, 0.707107f);
path.conicTo(2.00958e+38f, 9.36107e+37f, 1.7887e+38f, 8.33215e+37f, 0.707107f);
path.conicTo(1.56782e+38f, 7.30323e+37f, -2.20883e+37f, -1.02892e+37f, 0.707107f);
return path;
}
GrTessellator: implement straight skeleton, phase 2. This CL implements two major changes to the AA tessellating path renderer: 1) Fix inverted edges after stroke and simplify. Instead of detecting and fixing edges which invert on stroking during the stroking pass, we run the full simplify pass on both inner and outer contours, then create edge collapse events for the overlap regions. We then process the edge events in a priority queue and process them in order of decreasing alpha (this is the "edge event" part of the straight skeleton algorithm). By doing it after simplification, we ensure that there's a full-alpha intersection vertex to join the collapse edge to (which may have <1 alpha), so no spurious gradients appear in the rendered path. 2) "Pointy" vertices (defined as those which meet at an acute angle less than 14 degrees) are now properly bevelled off during stroking. This removes antialiasing artifacts which extend beyond the path boundary. Some ancillary changes: The extracted boundaries which are input to stroking have their line equations pre-normalized, and multiplied by winding. This simplifies a lot of code which was performing this computation on the fly. The workaround for the "intruding vertex" problem was removed, since the straight skeleton now moves the intruding vertex before it can cause problems. Bug: 756823 Change-Id: I271ed32be6847da55273b387e8c04bbf9b512b70 Reviewed-on: https://skia-review.googlesource.com/87341 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-12-19 23:09:54 +00:00
// An edge collapse event which also collapses a neighbour, requiring
// its event to be removed.
static SkPath create_path_25() {
SkPath path;
path.moveTo( 43.44110107421875, 148.15106201171875);
path.lineTo( 44.64471435546875, 148.16748046875);
path.lineTo( 46.35009765625, 147.403076171875);
path.lineTo( 46.45404052734375, 148.34906005859375);
path.lineTo( 45.0400390625, 148.54205322265625);
path.lineTo( 44.624053955078125, 148.9810791015625);
path.lineTo( 44.59405517578125, 149.16107177734375);
path.lineTo( 44.877044677734375, 149.62005615234375);
path.lineTo(144.373016357421875, 68.8070068359375);
return path;
}
// An edge collapse event causes an edge to become collinear, requiring
// its event to be removed.
static SkPath create_path_26() {
SkPath path;
path.moveTo( 43.44110107421875, 148.15106201171875);
path.lineTo( 44.64471435546875, 148.16748046875);
path.lineTo( 46.35009765625, 147.403076171875);
path.lineTo( 46.45404052734375, 148.34906005859375);
path.lineTo( 45.0400390625, 148.54205322265625);
path.lineTo( 44.624053955078125, 148.9810791015625);
path.lineTo( 44.59405517578125, 149.16107177734375);
path.lineTo( 44.877044677734375, 149.62005615234375);
path.lineTo(144.373016357421875, 68.8070068359375);
return path;
}
GrTessellator: fix for points which become non-finite on AA stroking. If input points are near-infinite, they may become inf or NaN when stroked. Before converting the results of intersection from double to float, clamp them to the [-FLT_MAX/FLT_MAX] range. BUG=798679 Change-Id: I7d61130dd26147a9b7cfd38aa96567e3867b5c3e Reviewed-on: https://skia-review.googlesource.com/90983 Commit-Queue: Stephen White <senorblanco@chromium.org> Reviewed-by: Brian Osman <brianosman@google.com>
2018-01-04 19:01:10 +00:00
// A path which results in non-finite points when stroked and bevelled for AA.
static SkPath create_path_27() {
SkPath path;
path.moveTo(8.5027233009104409507e+37, 1.7503381025241130639e+37);
path.lineTo(7.0923661737711584874e+37, 1.4600074517285415699e+37);
path.lineTo(7.0848733446033294691e+37, 1.4584649744781838604e+37);
path.lineTo(-2.0473916115129349496e+37, -4.2146796450364162012e+36);
path.lineTo(2.0473912312177548811e+37, 4.2146815465123165435e+36);
return path;
}
GrTessellator: fix for assert on bevelling. With some large values, intersection for bevelling will fail. These should just skip the point, not assert. BUG=798912 Change-Id: Ie5c8cc3c9387055e1e31480321a231f0e6ff153b Reviewed-on: https://skia-review.googlesource.com/91141 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2018-01-04 21:45:51 +00:00
// AA stroking this path produces intersection failures on bevelling.
// This should skip the point, but not assert.
static SkPath create_path_28() {
SkPath path;
path.moveTo(-7.5952312625177475154e+21, -2.6819185100266674911e+24);
path.lineTo( 1260.3787841796875, 1727.7947998046875);
path.lineTo( 1260.5567626953125, 1728.0386962890625);
path.lineTo(1.1482511310557754163e+21, 4.054538502765980051e+23);
path.lineTo(-7.5952312625177475154e+21, -2.6819185100266674911e+24);
return path;
}
GrTessellator: set a cap on quadratic linearization. Some pathological cases don't converge to a reasonable number of points when using uniform linearization of quadratic points. Cap them to the maximum which GrPathUtils supports. Add reduced test case from crbug-762369. BUG=762369 Change-Id: Icc744018e5c01a0e0fe2ec00613bdb25e49614e9 Reviewed-on: https://skia-review.googlesource.com/92721 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2018-01-09 16:49:08 +00:00
// A quad which generates a huge number of points (>2B) when uniformly
// linearized. This should not hang or OOM.
static SkPath create_path_29() {
SkPath path;
path.moveTo(10, 0);
path.lineTo(0, 0);
path.quadTo(10, 0, 0, 8315084722602508288);
return path;
}
Make GrFragmentProcessor be non-refcounted and use std::unique_ptr. Change-Id: I985e54a071338e99292a5aa2f42c92bc115b4008 Reviewed-on: https://skia-review.googlesource.com/32760 Commit-Queue: Brian Salomon <bsalomon@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2017-08-11 13:40:37 +00:00
static std::unique_ptr<GrFragmentProcessor> create_linear_gradient_processor(GrContext* ctx) {
GrTessellator: implement straight skeleton, phase 2. This CL implements two major changes to the AA tessellating path renderer: 1) Fix inverted edges after stroke and simplify. Instead of detecting and fixing edges which invert on stroking during the stroking pass, we run the full simplify pass on both inner and outer contours, then create edge collapse events for the overlap regions. We then process the edge events in a priority queue and process them in order of decreasing alpha (this is the "edge event" part of the straight skeleton algorithm). By doing it after simplification, we ensure that there's a full-alpha intersection vertex to join the collapse edge to (which may have <1 alpha), so no spurious gradients appear in the rendered path. 2) "Pointy" vertices (defined as those which meet at an acute angle less than 14 degrees) are now properly bevelled off during stroking. This removes antialiasing artifacts which extend beyond the path boundary. Some ancillary changes: The extracted boundaries which are input to stroking have their line equations pre-normalized, and multiplied by winding. This simplifies a lot of code which was performing this computation on the fly. The workaround for the "intruding vertex" problem was removed, since the straight skeleton now moves the intruding vertex before it can cause problems. Bug: 756823 Change-Id: I271ed32be6847da55273b387e8c04bbf9b512b70 Reviewed-on: https://skia-review.googlesource.com/87341 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-12-19 23:09:54 +00:00
GrTessellator (AA): null GrGeometryProcessor crash fix. Add a null-check on geometry processor creation. Add a test which exercises it: AA tessellator, with a non-invertible matrix and a fragment processor which needs local coords (e.g., linear gradient). BUG=691902 Change-Id: I005b893aed58d3ad2500c41501045ac94b0b4b95 Reviewed-on: https://skia-review.googlesource.com/8462 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-02-15 16:45:16 +00:00
SkPoint pts[2] = { {0, 0}, {1, 1} };
SkColor colors[2] = { SK_ColorGREEN, SK_ColorBLUE };
sk_sp<SkShader> shader = SkGradientShader::MakeLinear(
pts, colors, nullptr, SK_ARRAY_COUNT(colors), SkShader::kClamp_TileMode);
Pass GrRenderTargetContext's GrColorSpaceInfo to SkShader and SkColorFilter. Also to SkColorTo(Premul|Unpremul)GrColor4f. This can avoid cache lookups to find GrColorSpaceXforms as the xform pointer is stored in GrColorSpaceInfo after the first lookup. Also uses GrColorSpaceInfo to construct GrTextUtils::Paint. Bug: skia: Change-Id: Idf19d512a60d2269e6921c7fb54d93aee499a70d Reviewed-on: https://skia-review.googlesource.com/63660 Commit-Queue: Brian Salomon <bsalomon@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2017-10-25 19:12:19 +00:00
GrColorSpaceInfo colorSpaceInfo(nullptr, kRGBA_8888_GrPixelConfig);
SkShaderBase::AsFPArgs args(ctx, &SkMatrix::I(), &SkMatrix::I(),
SkFilterQuality::kLow_SkFilterQuality, &colorSpaceInfo);
Reland of SkShaderBase Introduce a private base class (SkShaderBase), to hide implementation details from the public interface (SkShader). Change-Id: Ib1d76cde880bd51868b97408710f8bb38128e536 Reviewed-on: https://skia-review.googlesource.com/17925 Reviewed-by: Florin Malita <fmalita@chromium.org> Commit-Queue: Florin Malita <fmalita@chromium.org>
2017-05-25 14:38:07 +00:00
return as_SB(shader)->asFragmentProcessor(args);
GrTessellator (AA): null GrGeometryProcessor crash fix. Add a null-check on geometry processor creation. Add a test which exercises it: AA tessellator, with a non-invertible matrix and a fragment processor which needs local coords (e.g., linear gradient). BUG=691902 Change-Id: I005b893aed58d3ad2500c41501045ac94b0b4b95 Reviewed-on: https://skia-review.googlesource.com/8462 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-02-15 16:45:16 +00:00
}
Remove atlas creation from GrResourceProvider This is pulled out of: https://skia-review.googlesource.com/c/6680/ (Make SkImage_Gpu be deferred) and is only tangentially related to the goal of that CL. Change-Id: I6b6db4869597070f85ab3b9fea178fc88c104f87 Reviewed-on: https://skia-review.googlesource.com/9106 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Robert Phillips <robertphillips@google.com>
2017-03-01 19:32:46 +00:00
static void test_path(GrContext* ctx,
GrTessellator (AA): null GrGeometryProcessor crash fix. Add a null-check on geometry processor creation. Add a test which exercises it: AA tessellator, with a non-invertible matrix and a fragment processor which needs local coords (e.g., linear gradient). BUG=691902 Change-Id: I005b893aed58d3ad2500c41501045ac94b0b4b95 Reviewed-on: https://skia-review.googlesource.com/8462 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-02-15 16:45:16 +00:00
GrRenderTargetContext* renderTargetContext,
const SkPath& path,
const SkMatrix& matrix = SkMatrix::I(),
GrAAType aaType = GrAAType::kNone,
Make GrFragmentProcessor be non-refcounted and use std::unique_ptr. Change-Id: I985e54a071338e99292a5aa2f42c92bc115b4008 Reviewed-on: https://skia-review.googlesource.com/32760 Commit-Queue: Brian Salomon <bsalomon@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2017-08-11 13:40:37 +00:00
std::unique_ptr<GrFragmentProcessor> fp = nullptr) {
Tessellating GPU path renderer. This path renderer converts paths to linear contours, resolves intersections via Bentley-Ottman, implements a trapezoidal decomposition a la Fournier and Montuno to produce triangles, and renders those with a single draw call. It does not currently do antialiasing, so it must be used in conjunction with multisampling. A fair amount of the code is to handle floating point edge cases in intersections. Rather than perform exact computations (which would require arbitrary precision arithmetic), we reconnect the mesh to reflect the intersection points. For example, intersections can occur above the current vertex, and force edges to be merged into the current vertex, requiring a restart of the intersections. Splitting edges for intersections can also force them to merge with formerly-distinct edges in the same polygon, or to violate the ordering of the active edge list, or the active edge state of split edges. BUG=skia: Review URL: https://codereview.chromium.org/855513004
2015-02-26 14:58:17 +00:00
GrTessellatingPathRenderer tess;
Make GrClipMaskManager stateless and push GrPipelineBuilder construction downstack This will be followed up with a CL to remove the GrRenderTarget from the GrPipelineBuilder. Split out of: https://codereview.chromium.org/1988923002/ (Declassify GrClipMaskManager and Remove GrRenderTarget and GrDrawTarget from GrPipelineBuilder) GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2035823002 Review-Url: https://codereview.chromium.org/2035823002
2016-06-03 17:59:20 +00:00
GrPaint paint;
Revert "Revert "Removing ref counting from GrXPFactory."" This reverts commit 003312a211e65f35e402d6fe80a32e23d4c94ac4. Change-Id: Ib41065e5c356d1dd99e70fa10611ac6756c2b79d Reviewed-on: https://skia-review.googlesource.com/6803 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Brian Salomon <bsalomon@google.com>
2017-01-09 16:46:10 +00:00
paint.setXPFactory(GrPorterDuffXPFactory::Get(SkBlendMode::kSrc));
GrTessellator (AA): null GrGeometryProcessor crash fix. Add a null-check on geometry processor creation. Add a test which exercises it: AA tessellator, with a non-invertible matrix and a fragment processor which needs local coords (e.g., linear gradient). BUG=691902 Change-Id: I005b893aed58d3ad2500c41501045ac94b0b4b95 Reviewed-on: https://skia-review.googlesource.com/8462 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-02-15 16:45:16 +00:00
if (fp) {
Make GrFragmentProcessor be non-refcounted and use std::unique_ptr. Change-Id: I985e54a071338e99292a5aa2f42c92bc115b4008 Reviewed-on: https://skia-review.googlesource.com/32760 Commit-Queue: Brian Salomon <bsalomon@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2017-08-11 13:40:37 +00:00
paint.addColorFragmentProcessor(std::move(fp));
GrTessellator (AA): null GrGeometryProcessor crash fix. Add a null-check on geometry processor creation. Add a test which exercises it: AA tessellator, with a non-invertible matrix and a fragment processor which needs local coords (e.g., linear gradient). BUG=691902 Change-Id: I005b893aed58d3ad2500c41501045ac94b0b4b95 Reviewed-on: https://skia-review.googlesource.com/8462 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-02-15 16:45:16 +00:00
}
Make GrClipMaskManager stateless and push GrPipelineBuilder construction downstack This will be followed up with a CL to remove the GrRenderTarget from the GrPipelineBuilder. Split out of: https://codereview.chromium.org/1988923002/ (Declassify GrClipMaskManager and Remove GrRenderTarget and GrDrawTarget from GrPipelineBuilder) GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2035823002 Review-Url: https://codereview.chromium.org/2035823002
2016-06-03 17:59:20 +00:00
Convert GrClip to an abstract base class Converts GrClip to an abstract base class and adds a "GrFixedClip" implementation. GrFixedClip denotes a clip implemented with fixed- function hardware. GrFixedClip allows us to remove the stateful "fClipMode" member from GrClipMaskManager, and in the future will be able to nicely encapsulate window rectangles. After this change GrClipMaskManager is just a wrapper around GrDrawTarget. We may want to consider removing it altogether. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1971343002 Review-Url: https://codereview.chromium.org/1971343002
2016-05-13 17:25:00 +00:00
GrNoClip noClip;
Improve heuristics for when to use ccpr Gives the convex and tessellating renderers first claim on larger paths, and the distance field renderer first claim on complex, non-volatile paths. This also requires plumbing the clip bounds through GrPathRenderer::CanDrawPathArgs. Bug: skia: Change-Id: I16e1d35ad5ee63251e33f113b1579cbba60456da Reviewed-on: https://skia-review.googlesource.com/42224 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Chris Dalton <csmartdalton@google.com>
2017-09-08 22:25:08 +00:00
SkIRect clipConservativeBounds = SkIRect::MakeWH(renderTargetContext->width(),
renderTargetContext->height());
Replace GrStrokeInfo with GrStyle. A side effect is that arbitrary path effects can no be pushed deeper into the Ganesh flow for paths. They may be applied by path renderers. GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1957363002 Committed: https://skia.googlesource.com/skia/+/33595bdf4b64a745f6340338d307e806e96c587f Review-Url: https://codereview.chromium.org/1957363002
2016-05-10 16:14:17 +00:00
GrStyle style(SkStrokeRec::kFill_InitStyle);
Use GrShape in GrPathRenderer. GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2081383006 Review-Url: https://codereview.chromium.org/2081383006
2016-06-24 17:42:16 +00:00
GrShape shape(path, style);
Remove atlas creation from GrResourceProvider This is pulled out of: https://skia-review.googlesource.com/c/6680/ (Make SkImage_Gpu be deferred) and is only tangentially related to the goal of that CL. Change-Id: I6b6db4869597070f85ab3b9fea178fc88c104f87 Reviewed-on: https://skia-review.googlesource.com/9106 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Robert Phillips <robertphillips@google.com>
2017-03-01 19:32:46 +00:00
GrPathRenderer::DrawPathArgs args{ctx,
Make GrPaints move their GrProcessor ownership into GrPipelineBuilder. This makes GrPaints usable only once. In some places we must make copies in order to issue draws with the same paint state. Change-Id: Ie816e5185ce93a064111cad64c6880e1e21184c2 Reviewed-on: https://skia-review.googlesource.com/6844 Reviewed-by: Robert Phillips <robertphillips@google.com> Commit-Queue: Brian Salomon <bsalomon@google.com>
2017-01-11 18:42:54 +00:00
std::move(paint),
&GrUserStencilSettings::kUnused,
renderTargetContext,
&noClip,
Improve heuristics for when to use ccpr Gives the convex and tessellating renderers first claim on larger paths, and the distance field renderer first claim on complex, non-volatile paths. This also requires plumbing the clip bounds through GrPathRenderer::CanDrawPathArgs. Bug: skia: Change-Id: I16e1d35ad5ee63251e33f113b1579cbba60456da Reviewed-on: https://skia-review.googlesource.com/42224 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Chris Dalton <csmartdalton@google.com>
2017-09-08 22:25:08 +00:00
&clipConservativeBounds,
GrTessellator (AA): null GrGeometryProcessor crash fix. Add a null-check on geometry processor creation. Add a test which exercises it: AA tessellator, with a non-invertible matrix and a fragment processor which needs local coords (e.g., linear gradient). BUG=691902 Change-Id: I005b893aed58d3ad2500c41501045ac94b0b4b95 Reviewed-on: https://skia-review.googlesource.com/8462 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-02-15 16:45:16 +00:00
&matrix,
Make GrPaints move their GrProcessor ownership into GrPipelineBuilder. This makes GrPaints usable only once. In some places we must make copies in order to issue draws with the same paint state. Change-Id: Ie816e5185ce93a064111cad64c6880e1e21184c2 Reviewed-on: https://skia-review.googlesource.com/6844 Reviewed-by: Robert Phillips <robertphillips@google.com> Commit-Queue: Brian Salomon <bsalomon@google.com>
2017-01-11 18:42:54 +00:00
&shape,
GrTessellator (AA): null GrGeometryProcessor crash fix. Add a null-check on geometry processor creation. Add a test which exercises it: AA tessellator, with a non-invertible matrix and a fragment processor which needs local coords (e.g., linear gradient). BUG=691902 Change-Id: I005b893aed58d3ad2500c41501045ac94b0b4b95 Reviewed-on: https://skia-review.googlesource.com/8462 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-02-15 16:45:16 +00:00
aaType,
Make GrPaints move their GrProcessor ownership into GrPipelineBuilder. This makes GrPaints usable only once. In some places we must make copies in order to issue draws with the same paint state. Change-Id: Ie816e5185ce93a064111cad64c6880e1e21184c2 Reviewed-on: https://skia-review.googlesource.com/6844 Reviewed-by: Robert Phillips <robertphillips@google.com> Commit-Queue: Brian Salomon <bsalomon@google.com>
2017-01-11 18:42:54 +00:00
false};
Also provides the ResourceProvider to onDrawPath which allows the DF PR to no longer require access to GrContext. Review URL: https://codereview.chromium.org/1265763002
2015-07-31 13:48:27 +00:00
tess.drawPath(args);
Tessellating GPU path renderer. This path renderer converts paths to linear contours, resolves intersections via Bentley-Ottman, implements a trapezoidal decomposition a la Fournier and Montuno to produce triangles, and renders those with a single draw call. It does not currently do antialiasing, so it must be used in conjunction with multisampling. A fair amount of the code is to handle floating point edge cases in intersections. Rather than perform exact computations (which would require arbitrary precision arithmetic), we reconnect the mesh to reflect the intersection points. For example, intersections can occur above the current vertex, and force edges to be merged into the current vertex, requiring a restart of the intersections. Splitting edges for intersections can also force them to merge with formerly-distinct edges in the same polygon, or to violate the ordering of the active edge list, or the active edge state of split edges. BUG=skia: Review URL: https://codereview.chromium.org/855513004
2015-02-26 14:58:17 +00:00
}
Convert some unit tests from running on GL contexts to running on all contexts GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1873313002 Review URL: https://codereview.chromium.org/1873313002
2016-04-11 21:40:50 +00:00
DEF_GPUTEST_FOR_ALL_CONTEXTS(TessellatingPathRendererTests, reporter, ctxInfo) {
Remove GrTestTarget Change-Id: I4d2e1aa54732ff75b44a197255035c0e04548219 Reviewed-on: https://skia-review.googlesource.com/7327 Commit-Queue: Robert Phillips <robertphillips@google.com> Reviewed-by: Brian Salomon <bsalomon@google.com>
2017-01-20 14:11:37 +00:00
GrContext* ctx = ctxInfo.grContext();
Rm makeRenderTargetContext in favor of deferred version (take 3) This is a reland of: https://skia-review.googlesource.com/c/13001/ (Rm makeRenderTargetContext in favor of deferred version) TBR=bsalomon@google.com Change-Id: If81f4d9fb889c091cd37ffde133d906fb3e37773 Reviewed-on: https://skia-review.googlesource.com/14027 Commit-Queue: Robert Phillips <robertphillips@google.com> Reviewed-by: Greg Daniel <egdaniel@google.com>
2017-04-24 14:57:28 +00:00
sk_sp<GrRenderTargetContext> rtc(ctx->makeDeferredRenderTargetContext(
SkBackingFit::kApprox,
Remove GrTestTarget Change-Id: I4d2e1aa54732ff75b44a197255035c0e04548219 Reviewed-on: https://skia-review.googlesource.com/7327 Commit-Queue: Robert Phillips <robertphillips@google.com> Reviewed-by: Brian Salomon <bsalomon@google.com>
2017-01-20 14:11:37 +00:00
800, 800,
kRGBA_8888_GrPixelConfig,
nullptr,
0,
Have mip status match surface when snapping image from wrapped object Also fixes some bugs involved with creating mipped SkSurfaces. Bug: skia: Change-Id: I6e0109000eadd2bdee4a907d3ee2231104528165 Reviewed-on: https://skia-review.googlesource.com/65063 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Greg Daniel <egdaniel@google.com>
2017-10-30 17:41:26 +00:00
GrMipMapped::kNo,
Remove GrTestTarget Change-Id: I4d2e1aa54732ff75b44a197255035c0e04548219 Reviewed-on: https://skia-review.googlesource.com/7327 Commit-Queue: Robert Phillips <robertphillips@google.com> Reviewed-by: Brian Salomon <bsalomon@google.com>
2017-01-20 14:11:37 +00:00
kTopLeft_GrSurfaceOrigin));
Rename GrDrawContext to GrRenderTargetContext This is in preparation for GrTextureContext and GrSurfaceContext BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4030 Change-Id: Ie58c93052e68f3f1f5fe8d15d63760de274a6fbd Reviewed-on: https://skia-review.googlesource.com/4030 Commit-Queue: Brian Osman <brianosman@google.com> Reviewed-by: Brian Salomon <bsalomon@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com>
2016-10-27 18:47:55 +00:00
if (!rtc) {
Retract GrRenderTarget from GrTestTarget Split out of: https://codereview.chromium.org/1988923002/ (Declassify GrClipMaskManager and Remove GrRenderTarget and GrDrawTarget from GrPipelineBuilder) GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=2004433002 Review-Url: https://codereview.chromium.org/2004433002
2016-05-20 18:14:33 +00:00
return;
}
Remove GrTestTarget Change-Id: I4d2e1aa54732ff75b44a197255035c0e04548219 Reviewed-on: https://skia-review.googlesource.com/7327 Commit-Queue: Robert Phillips <robertphillips@google.com> Reviewed-by: Brian Salomon <bsalomon@google.com>
2017-01-20 14:11:37 +00:00
ctx->flush();
Remove atlas creation from GrResourceProvider This is pulled out of: https://skia-review.googlesource.com/c/6680/ (Make SkImage_Gpu be deferred) and is only tangentially related to the goal of that CL. Change-Id: I6b6db4869597070f85ab3b9fea178fc88c104f87 Reviewed-on: https://skia-review.googlesource.com/9106 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Robert Phillips <robertphillips@google.com>
2017-03-01 19:32:46 +00:00
test_path(ctx, rtc.get(), create_path_0());
test_path(ctx, rtc.get(), create_path_1());
test_path(ctx, rtc.get(), create_path_2());
test_path(ctx, rtc.get(), create_path_3());
test_path(ctx, rtc.get(), create_path_4());
test_path(ctx, rtc.get(), create_path_5());
test_path(ctx, rtc.get(), create_path_6());
test_path(ctx, rtc.get(), create_path_7());
test_path(ctx, rtc.get(), create_path_8());
test_path(ctx, rtc.get(), create_path_9());
test_path(ctx, rtc.get(), create_path_10());
test_path(ctx, rtc.get(), create_path_11());
test_path(ctx, rtc.get(), create_path_12());
test_path(ctx, rtc.get(), create_path_13());
test_path(ctx, rtc.get(), create_path_14());
test_path(ctx, rtc.get(), create_path_15());
test_path(ctx, rtc.get(), create_path_16());
GrTessellator (AA): null GrGeometryProcessor crash fix. Add a null-check on geometry processor creation. Add a test which exercises it: AA tessellator, with a non-invertible matrix and a fragment processor which needs local coords (e.g., linear gradient). BUG=691902 Change-Id: I005b893aed58d3ad2500c41501045ac94b0b4b95 Reviewed-on: https://skia-review.googlesource.com/8462 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-02-15 16:45:16 +00:00
SkMatrix nonInvertibleMatrix = SkMatrix::MakeScale(0, 0);
Make GrFragmentProcessor be non-refcounted and use std::unique_ptr. Change-Id: I985e54a071338e99292a5aa2f42c92bc115b4008 Reviewed-on: https://skia-review.googlesource.com/32760 Commit-Queue: Brian Salomon <bsalomon@google.com> Reviewed-by: Brian Osman <brianosman@google.com>
2017-08-11 13:40:37 +00:00
std::unique_ptr<GrFragmentProcessor> fp(create_linear_gradient_processor(ctx));
test_path(ctx, rtc.get(), create_path_17(), nonInvertibleMatrix, GrAAType::kCoverage,
std::move(fp));
GrTessellator: fix for vertex coincident with enclosing edge. If a previously-enclosing edge coincides exactly with the current vertex, there are no two adjacent edges which enclose the vertex. Since find_enclosing_edges() ensures that the left enclosing edge is to the left of the vertex, the fix is to split the right enclosing edge on the current vertex and restart intersection tests. Bug: 716720 Change-Id: Id26c5b92a6d6139f348e99554638cded37e81a8e Reviewed-on: https://skia-review.googlesource.com/15261 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-05-03 20:00:38 +00:00
test_path(ctx, rtc.get(), create_path_18());
GrTessellator: fix for three edges becoming collinear after splitting. In rare cases, a single edge can become collinear with two adjacent neighbour edges after it's split. The solution is to continue to merge until no collinear edges are found. BUG=722000 Change-Id: Ia5dd212b7acfb40ed1d6c74ebfa9e4a4746fe40a Reviewed-on: https://skia-review.googlesource.com/17963 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-05-25 18:47:11 +00:00
test_path(ctx, rtc.get(), create_path_19());
GrTessellator: fix intersection above the first vertex. Handle the case where the an intersection falls not only above both edge endpoints, but above the first vertex in the mesh. This requires passing the mesh into check_for_intersection(), in order to modify the head. We also need to rewind the mesh after insertion, since we need to rewind to the newly-inserted vertex. This also cleans up vertex ID computation a little (for logging), so that vertices before the first vertex or after the last have a reasonable ID. It also cleans up the intersection-on-endpoint special cases by refactoring the calls to split_edge(). BUG=730687 Change-Id: Idea736eca7b7c3c5d8a470b1373a16ad8e649e80 Reviewed-on: https://skia-review.googlesource.com/19069 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-06-08 18:41:01 +00:00
test_path(ctx, rtc.get(), create_path_20(), SkMatrix(), GrAAType::kCoverage);
GrTessellator: fix two NaN issues. If a point in the path rounds to +inf/-inf, the intersection code can produce NaN, which is unsortable. Fix: ignore non-finite intersections. Quadratic interpolation can sometimes produce NaN, which will never satisfy the flatness criterion. Abort if any of the interpolated points are non-finite. Bug:732023 Change-Id: If5881796e589c75b8f74459f42d00918619713a2 Reviewed-on: https://skia-review.googlesource.com/19467 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-06-12 15:43:18 +00:00
test_path(ctx, rtc.get(), create_path_21(), SkMatrix(), GrAAType::kCoverage);
test_path(ctx, rtc.get(), create_path_22());
GrTessellator: fix for collinear out-of-range intersections. Change 3b5a3fa8b1c11d4bd4499b040311f4c3553ebf8c introduced support for out-of-range intersections, which is necessary when the top and bottom vertices of an edge differ by only one ULP in the primary sort order and can't be split in-order. However, some out-of-range intersections produce edges which cancel each other out on splitting, in particular when the intersection is collinear with the newly-computed edge. This undoes the effect of the split. The tessellator then rewinds, re-detects the intersection, resplits, an infinite loop. The fix is to check for out-of-range intersections which are also collinear, and ignore them. This is ok, because these are not the cases we care about it change 3b5a3f above, which are never collinear. Bug: 753867 Change-Id: I590231e0e6f19c98f1ccf46cb7acc8a63ba35a9d Reviewed-on: https://skia-review.googlesource.com/34925 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-08-15 21:01:15 +00:00
test_path(ctx, rtc.get(), create_path_23());
Add support for vkCmdCopyBuffer This is then used when need to update GPU only buffers with data of size greater than 65536. We create a temporary transfer buffer and then copy that buffer into our GPU buffer. Bug: skia: Change-Id: I4bb9cb660f2ac1ccbbd1b508bb4ca6876342136f Reviewed-on: https://skia-review.googlesource.com/38725 Reviewed-by: Jim Van Verth <jvanverth@google.com> Commit-Queue: Greg Daniel <egdaniel@google.com>
2017-08-25 15:55:50 +00:00
test_path(ctx, rtc.get(), create_path_24());
GrTessellator: implement straight skeleton, phase 2. This CL implements two major changes to the AA tessellating path renderer: 1) Fix inverted edges after stroke and simplify. Instead of detecting and fixing edges which invert on stroking during the stroking pass, we run the full simplify pass on both inner and outer contours, then create edge collapse events for the overlap regions. We then process the edge events in a priority queue and process them in order of decreasing alpha (this is the "edge event" part of the straight skeleton algorithm). By doing it after simplification, we ensure that there's a full-alpha intersection vertex to join the collapse edge to (which may have <1 alpha), so no spurious gradients appear in the rendered path. 2) "Pointy" vertices (defined as those which meet at an acute angle less than 14 degrees) are now properly bevelled off during stroking. This removes antialiasing artifacts which extend beyond the path boundary. Some ancillary changes: The extracted boundaries which are input to stroking have their line equations pre-normalized, and multiplied by winding. This simplifies a lot of code which was performing this computation on the fly. The workaround for the "intruding vertex" problem was removed, since the straight skeleton now moves the intruding vertex before it can cause problems. Bug: 756823 Change-Id: I271ed32be6847da55273b387e8c04bbf9b512b70 Reviewed-on: https://skia-review.googlesource.com/87341 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2017-12-19 23:09:54 +00:00
test_path(ctx, rtc.get(), create_path_25(), SkMatrix(), GrAAType::kCoverage);
test_path(ctx, rtc.get(), create_path_26(), SkMatrix(), GrAAType::kCoverage);
GrTessellator: fix for points which become non-finite on AA stroking. If input points are near-infinite, they may become inf or NaN when stroked. Before converting the results of intersection from double to float, clamp them to the [-FLT_MAX/FLT_MAX] range. BUG=798679 Change-Id: I7d61130dd26147a9b7cfd38aa96567e3867b5c3e Reviewed-on: https://skia-review.googlesource.com/90983 Commit-Queue: Stephen White <senorblanco@chromium.org> Reviewed-by: Brian Osman <brianosman@google.com>
2018-01-04 19:01:10 +00:00
test_path(ctx, rtc.get(), create_path_27(), SkMatrix(), GrAAType::kCoverage);
GrTessellator: fix for assert on bevelling. With some large values, intersection for bevelling will fail. These should just skip the point, not assert. BUG=798912 Change-Id: Ie5c8cc3c9387055e1e31480321a231f0e6ff153b Reviewed-on: https://skia-review.googlesource.com/91141 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2018-01-04 21:45:51 +00:00
test_path(ctx, rtc.get(), create_path_28(), SkMatrix(), GrAAType::kCoverage);
GrTessellator: set a cap on quadratic linearization. Some pathological cases don't converge to a reasonable number of points when using uniform linearization of quadratic points. Cap them to the maximum which GrPathUtils supports. Add reduced test case from crbug-762369. BUG=762369 Change-Id: Icc744018e5c01a0e0fe2ec00613bdb25e49614e9 Reviewed-on: https://skia-review.googlesource.com/92721 Reviewed-by: Brian Salomon <bsalomon@google.com> Commit-Queue: Stephen White <senorblanco@chromium.org>
2018-01-09 16:49:08 +00:00
test_path(ctx, rtc.get(), create_path_29());
Tessellating GPU path renderer. This path renderer converts paths to linear contours, resolves intersections via Bentley-Ottman, implements a trapezoidal decomposition a la Fournier and Montuno to produce triangles, and renders those with a single draw call. It does not currently do antialiasing, so it must be used in conjunction with multisampling. A fair amount of the code is to handle floating point edge cases in intersections. Rather than perform exact computations (which would require arbitrary precision arithmetic), we reconnect the mesh to reflect the intersection points. For example, intersections can occur above the current vertex, and force edges to be merged into the current vertex, requiring a restart of the intersections. Splitting edges for intersections can also force them to merge with formerly-distinct edges in the same polygon, or to violate the ordering of the active edge list, or the active edge state of split edges. BUG=skia: Review URL: https://codereview.chromium.org/855513004
2015-02-26 14:58:17 +00:00
}
#endif
Reference in New Issue Copy Permalink