AuroraMiddleware/skia2
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
4b3d2b48c6
skia2/gm/aaxfermodes.cpp

274 lines
9.8 KiB
C++
Raw Normal View History

Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +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.
*/
make SkXfermode.h go away This is step one: - make SkXfermode useless to public clients - everything they should need is in SkBlendMode.h Step two: - remove SkXfermode.h entirely (since skia core will already be using SkXfermodePriv.h) BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4534 Change-Id: If2cea9f71df92430ed6644edb98dd306c5572cbc Reviewed-on: https://skia-review.googlesource.com/4534 Commit-Queue: Mike Reed <reed@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2016-11-15 21:44:34 +00:00
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
#include "gm.h"
GM: some header cleanup gm.h includes sk_tool_utils.h but does not use it. The bulk of this CL makes each gm that uses sk_tool_utils include it. sk_tool_utils.h also provided SkRandom and SkTDArray, so a couple GMs add those headers too. Change-Id: Ieb2a7c542f0ca89c3223f744fc11b0ff37af36c1 Reviewed-on: https://skia-review.googlesource.com/10014 Commit-Queue: Mike Klein <mtklein@chromium.org> Reviewed-by: Herb Derby <herb@google.com>
2017-03-22 17:47:51 +00:00
#include "sk_tool_utils.h"
make SkXfermode.h go away This is step one: - make SkXfermode useless to public clients - everything they should need is in SkBlendMode.h Step two: - remove SkXfermode.h entirely (since skia core will already be using SkXfermodePriv.h) BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4534 Change-Id: If2cea9f71df92430ed6644edb98dd306c5572cbc Reviewed-on: https://skia-review.googlesource.com/4534 Commit-Queue: Mike Reed <reed@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2016-11-15 21:44:34 +00:00
#include "SkBlendModePriv.h"
IWYU: 'core' target, files starting A-C. TBR=reed@google.com Verbal lgtm, does not change API. Committed: https://skia.googlesource.com/skia/+/7403d87db8e43d4c2b5b25ac22a0ebc22bd09d69 Review URL: https://codereview.chromium.org/1265033002
2015-08-05 20:57:49 +00:00
#include "SkPath.h"
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
#include "SkShader.h"
enum {
make SkXfermode.h go away This is step one: - make SkXfermode useless to public clients - everything they should need is in SkBlendMode.h Step two: - remove SkXfermode.h entirely (since skia core will already be using SkXfermodePriv.h) BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4534 Change-Id: If2cea9f71df92430ed6644edb98dd306c5572cbc Reviewed-on: https://skia-review.googlesource.com/4534 Commit-Queue: Mike Reed <reed@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2016-11-15 21:44:34 +00:00
kXfermodeCount = (int)SkBlendMode::kLastMode + 1 + 1, // extra for arith
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
kShapeSize = 22,
kShapeSpacing = 36,
kShapeTypeSpacing = 4 * kShapeSpacing / 3,
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
kPaintSpacing = 4 * kShapeTypeSpacing,
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
kLabelSpacing = 3 * kShapeSize,
kMargin = kShapeSpacing / 2,
kXfermodeTypeSpacing = kLabelSpacing + 2 * kPaintSpacing + kShapeTypeSpacing,
kTitleSpacing = 3 * kShapeSpacing / 4,
kSubtitleSpacing = 5 * kShapeSpacing / 8
};
Replace a lot of 'static const' with 'constexpr' or 'const'. 'static const' means, there must be at most one of these, and initialize it at compile time if possible or runtime if necessary. This leads to unexpected code execution, and TSAN* will complain about races on the guard variables. Generally 'constexpr' or 'const' are better choices. Neither can cause races: they're either intialized at compile time (constexpr) or intialized each time independently (const). This CL prefers constexpr where possible, and uses const where not. It even prefers constexpr over const where they don't make a difference... I want to have lots of examples of constexpr for people to see and mimic. The scoped-to-class static has nothing to do with any of this, and is not changed. * Not yet on the bots, which use an older TSAN. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2300623005 Review-Url: https://codereview.chromium.org/2300623005
2016-09-01 18:24:54 +00:00
constexpr SkColor kBGColor = 0xc8d2b887;
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
Replace a lot of 'static const' with 'constexpr' or 'const'. 'static const' means, there must be at most one of these, and initialize it at compile time if possible or runtime if necessary. This leads to unexpected code execution, and TSAN* will complain about races on the guard variables. Generally 'constexpr' or 'const' are better choices. Neither can cause races: they're either intialized at compile time (constexpr) or intialized each time independently (const). This CL prefers constexpr where possible, and uses const where not. It even prefers constexpr over const where they don't make a difference... I want to have lots of examples of constexpr for people to see and mimic. The scoped-to-class static has nothing to do with any of this, and is not changed. * Not yet on the bots, which use an older TSAN. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2300623005 Review-Url: https://codereview.chromium.org/2300623005
2016-09-01 18:24:54 +00:00
constexpr SkColor kShapeColors[2] = {
0x82ff0080, // input color unknown
0xff00ffff, // input color opaque
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
};
enum Shape {
kSquare_Shape,
kDiamond_Shape,
kOval_Shape,
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
kConcave_Shape,
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
kLast_Shape = kConcave_Shape
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
};
/**
* Verifies AA works properly on all Xfermodes, including arithmetic, with both opaque and unknown
* src colors.
*/
Revert "Revert "replace Arithmetic xfermode with imagefilter"" This reverts commit 10ff5bfa789b6b602464e8511fdf676c6f5b1bd4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2680 TBR= Change-Id: Iac2f4d48d227ff426ccf02b7eae280f382ad3580 Reviewed-on: https://skia-review.googlesource.com/2680 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com>
2016-09-26 22:51:24 +00:00
class AAXfermodesGM : public skiagm::GM {
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
public:
AAXfermodesGM() {}
protected:
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
enum DrawingPass {
kCheckerboard_Pass,
kBackground_Pass,
kShape_Pass
};
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
SkString onShortName() override {
return SkString("aaxfermodes");
}
SkISize onISize() override {
return SkISize::Make(2 * kMargin + 2 * kXfermodeTypeSpacing -
(kXfermodeTypeSpacing - (kLabelSpacing + 2 * kPaintSpacing)),
2 * kMargin + kTitleSpacing + kSubtitleSpacing +
make SkXfermode.h go away This is step one: - make SkXfermode useless to public clients - everything they should need is in SkBlendMode.h Step two: - remove SkXfermode.h entirely (since skia core will already be using SkXfermodePriv.h) BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4534 Change-Id: If2cea9f71df92430ed6644edb98dd306c5572cbc Reviewed-on: https://skia-review.googlesource.com/4534 Commit-Queue: Mike Reed <reed@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2016-11-15 21:44:34 +00:00
(1 + (int)SkBlendMode::kLastCoeffMode) * kShapeSpacing);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
}
void onOnceBeforeDraw() override {
fLabelPaint.setAntiAlias(true);
rename portable_typeface_always to portable_typeface TBR=reed@google.com Review URL: https://codereview.chromium.org/1257773002
2015-07-24 19:09:25 +00:00
sk_tool_utils::set_portable_typeface(&fLabelPaint);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
fLabelPaint.setTextSize(5 * kShapeSize/8);
fLabelPaint.setSubpixelText(true);
Replace a lot of 'static const' with 'constexpr' or 'const'. 'static const' means, there must be at most one of these, and initialize it at compile time if possible or runtime if necessary. This leads to unexpected code execution, and TSAN* will complain about races on the guard variables. Generally 'constexpr' or 'const' are better choices. Neither can cause races: they're either intialized at compile time (constexpr) or intialized each time independently (const). This CL prefers constexpr where possible, and uses const where not. It even prefers constexpr over const where they don't make a difference... I want to have lots of examples of constexpr for people to see and mimic. The scoped-to-class static has nothing to do with any of this, and is not changed. * Not yet on the bots, which use an older TSAN. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2300623005 Review-Url: https://codereview.chromium.org/2300623005
2016-09-01 18:24:54 +00:00
constexpr SkScalar radius = -1.4f * kShapeSize/2;
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
SkPoint pts[4] = {
{-radius, 0},
{0, -1.33f * radius},
{radius, 0},
{0, 1.33f * radius}
};
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
fOval.moveTo(pts[0]);
fOval.quadTo(pts[1], pts[2]);
fOval.quadTo(pts[3], pts[0]);
fConcave.moveTo(-radius, 0);
fConcave.quadTo(0, 0, 0, -radius);
fConcave.quadTo(0, 0, radius, 0);
fConcave.quadTo(0, 0, 0, radius);
fConcave.quadTo(0, 0, -radius, 0);
fConcave.close();
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
}
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
void draw_pass(SkCanvas* canvas, DrawingPass drawingPass) {
SkRect clipRect =
{ -kShapeSize*11/16, -kShapeSize*11/16, kShapeSize*11/16, kShapeSize*11/16 };
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
canvas->save();
if (kCheckerboard_Pass == drawingPass) {
canvas->translate(kMargin, kMargin);
}
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
canvas->translate(0, kTitleSpacing);
for (size_t xfermodeSet = 0; xfermodeSet < 2; xfermodeSet++) {
make SkXfermode.h go away This is step one: - make SkXfermode useless to public clients - everything they should need is in SkBlendMode.h Step two: - remove SkXfermode.h entirely (since skia core will already be using SkXfermodePriv.h) BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4534 Change-Id: If2cea9f71df92430ed6644edb98dd306c5572cbc Reviewed-on: https://skia-review.googlesource.com/4534 Commit-Queue: Mike Reed <reed@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2016-11-15 21:44:34 +00:00
size_t firstMode = ((size_t)SkBlendMode::kLastCoeffMode + 1) * xfermodeSet;
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
canvas->save();
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
if (kShape_Pass == drawingPass) {
fLabelPaint.setTextAlign(SkPaint::kCenter_Align);
add drawString helper to canvas Many tests and examples use drawText with a guess of how long the text is in bytes, or a call to strlen(). Add a helper to SkCanvas to simplify these examples. Add another helper for SkString. R=reed@google.com Change-Id: I0204a31e938f065606f08ee7cd9a6b36db791ee2 Reviewed-on: https://skia-review.googlesource.com/13642 Commit-Queue: Cary Clark <caryclark@google.com> Reviewed-by: Cary Clark <caryclark@google.com> Reviewed-by: Mike Reed <reed@google.com> Reviewed-by: Cary Clark <caryclark@skia.org>
2017-04-28 19:35:12 +00:00
canvas->drawString("Src Unknown",
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
kLabelSpacing + kShapeTypeSpacing * 1.5f + kShapeSpacing / 2,
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
kSubtitleSpacing / 2 + fLabelPaint.getTextSize() / 3, fLabelPaint);
add drawString helper to canvas Many tests and examples use drawText with a guess of how long the text is in bytes, or a call to strlen(). Add a helper to SkCanvas to simplify these examples. Add another helper for SkString. R=reed@google.com Change-Id: I0204a31e938f065606f08ee7cd9a6b36db791ee2 Reviewed-on: https://skia-review.googlesource.com/13642 Commit-Queue: Cary Clark <caryclark@google.com> Reviewed-by: Cary Clark <caryclark@google.com> Reviewed-by: Mike Reed <reed@google.com> Reviewed-by: Cary Clark <caryclark@skia.org>
2017-04-28 19:35:12 +00:00
canvas->drawString("Src Opaque",
Style bikeshed - remove extraneous whitespace GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1842753002 Review URL: https://codereview.chromium.org/1842753002
2016-03-29 16:03:52 +00:00
kLabelSpacing + kShapeTypeSpacing * 1.5f + kShapeSpacing / 2 +
kPaintSpacing, kSubtitleSpacing / 2 + fLabelPaint.getTextSize() / 3,
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
fLabelPaint);
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
}
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
canvas->translate(0, kSubtitleSpacing + kShapeSpacing/2);
make SkXfermode.h go away This is step one: - make SkXfermode useless to public clients - everything they should need is in SkBlendMode.h Step two: - remove SkXfermode.h entirely (since skia core will already be using SkXfermodePriv.h) BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4534 Change-Id: If2cea9f71df92430ed6644edb98dd306c5572cbc Reviewed-on: https://skia-review.googlesource.com/4534 Commit-Queue: Mike Reed <reed@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2016-11-15 21:44:34 +00:00
for (size_t m = 0; m <= (size_t)SkBlendMode::kLastCoeffMode; m++) {
if (firstMode + m > (size_t)SkBlendMode::kLastMode) {
Revert "Revert "replace Arithmetic xfermode with imagefilter"" This reverts commit 10ff5bfa789b6b602464e8511fdf676c6f5b1bd4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2680 TBR= Change-Id: Iac2f4d48d227ff426ccf02b7eae280f382ad3580 Reviewed-on: https://skia-review.googlesource.com/2680 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com>
2016-09-26 22:51:24 +00:00
break;
}
Revert[8] "replace SkXfermode obj with SkBlendMode enum in paints" This reverts commit c245574ba3d0e2ade6c94b2812de3baa383bf4c4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2396953002 TBR= Review-Url: https://codereview.chromium.org/2396953002
2016-10-06 00:33:02 +00:00
SkBlendMode mode = static_cast<SkBlendMode>(firstMode + m);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
canvas->save();
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
if (kShape_Pass == drawingPass) {
this->drawModeName(canvas, mode);
}
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
canvas->translate(kLabelSpacing + kShapeSpacing/2, 0);
for (size_t colorIdx = 0; colorIdx < SK_ARRAY_COUNT(kShapeColors); colorIdx++) {
SkPaint paint;
this->setupShapePaint(canvas, kShapeColors[colorIdx], mode, &paint);
SkASSERT(colorIdx == 0 || 255 == paint.getAlpha());
canvas->save();
for (size_t shapeIdx = 0; shapeIdx <= kLast_Shape; shapeIdx++) {
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
if (kShape_Pass != drawingPass) {
canvas->save();
canvas->clipRect(clipRect);
if (kCheckerboard_Pass == drawingPass) {
sk_tool_utils::draw_checkerboard(canvas, 0xffffffff, 0xffc6c3c6,
10);
} else {
SkASSERT(kBackground_Pass == drawingPass);
Revert[8] "replace SkXfermode obj with SkBlendMode enum in paints" This reverts commit c245574ba3d0e2ade6c94b2812de3baa383bf4c4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2396953002 TBR= Review-Url: https://codereview.chromium.org/2396953002
2016-10-06 00:33:02 +00:00
canvas->drawColor(kBGColor, SkBlendMode::kSrc);
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
}
canvas->restore();
} else {
this->drawShape(canvas, static_cast<Shape>(shapeIdx), paint, mode);
}
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
canvas->translate(kShapeTypeSpacing, 0);
}
canvas->restore();
canvas->translate(kPaintSpacing, 0);
}
canvas->restore();
canvas->translate(0, kShapeSpacing);
}
canvas->restore();
canvas->translate(kXfermodeTypeSpacing, 0);
}
canvas->restore();
}
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
void onDraw(SkCanvas* canvas) override {
draw_pass(canvas, kCheckerboard_Pass);
Style Change: NULL->nullptr DOCS_PREVIEW= https://skia.org/?cl=1316233002 Review URL: https://codereview.chromium.org/1316233002
2015-08-27 14:41:13 +00:00
canvas->saveLayer(nullptr, nullptr);
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
canvas->translate(kMargin, kMargin);
draw_pass(canvas, kBackground_Pass);
SkPaint titlePaint(fLabelPaint);
titlePaint.setTextSize(9 * titlePaint.getTextSize() / 8);
titlePaint.setFakeBoldText(true);
titlePaint.setTextAlign(SkPaint::kCenter_Align);
add drawString helper to canvas Many tests and examples use drawText with a guess of how long the text is in bytes, or a call to strlen(). Add a helper to SkCanvas to simplify these examples. Add another helper for SkString. R=reed@google.com Change-Id: I0204a31e938f065606f08ee7cd9a6b36db791ee2 Reviewed-on: https://skia-review.googlesource.com/13642 Commit-Queue: Cary Clark <caryclark@google.com> Reviewed-by: Cary Clark <caryclark@google.com> Reviewed-by: Mike Reed <reed@google.com> Reviewed-by: Cary Clark <caryclark@skia.org>
2017-04-28 19:35:12 +00:00
canvas->drawString("Porter Duff",
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
kLabelSpacing + 4 * kShapeTypeSpacing,
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
kTitleSpacing / 2 + titlePaint.getTextSize() / 3, titlePaint);
add drawString helper to canvas Many tests and examples use drawText with a guess of how long the text is in bytes, or a call to strlen(). Add a helper to SkCanvas to simplify these examples. Add another helper for SkString. R=reed@google.com Change-Id: I0204a31e938f065606f08ee7cd9a6b36db791ee2 Reviewed-on: https://skia-review.googlesource.com/13642 Commit-Queue: Cary Clark <caryclark@google.com> Reviewed-by: Cary Clark <caryclark@google.com> Reviewed-by: Mike Reed <reed@google.com> Reviewed-by: Cary Clark <caryclark@skia.org>
2017-04-28 19:35:12 +00:00
canvas->drawString("Advanced",
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
kXfermodeTypeSpacing + kLabelSpacing + 4 * kShapeTypeSpacing,
Today's gm include many differences caused by platform font implementations. This experiment replaces the label used in the aaxfermodes gm with aliased text generated from paths common to all platforms. Since there is no way today to generate all dm output from trybots, this will be checked in to confirm that this strategy provides simpler output across devices. This does not introduce a new public interface; instead, dm uses a extern backdoor to install the SkTypeface::CreateFromName handler. Review URL: https://codereview.chromium.org/1163283002
2015-06-10 16:31:09 +00:00
kTitleSpacing / 2 + titlePaint.getTextSize() / 3, titlePaint);
draw_pass(canvas, kShape_Pass);
canvas->restore();
}
Revert[8] "replace SkXfermode obj with SkBlendMode enum in paints" This reverts commit c245574ba3d0e2ade6c94b2812de3baa383bf4c4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2396953002 TBR= Review-Url: https://codereview.chromium.org/2396953002
2016-10-06 00:33:02 +00:00
void drawModeName(SkCanvas* canvas, SkBlendMode mode) {
make SkXfermode.h go away This is step one: - make SkXfermode useless to public clients - everything they should need is in SkBlendMode.h Step two: - remove SkXfermode.h entirely (since skia core will already be using SkXfermodePriv.h) BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=4534 Change-Id: If2cea9f71df92430ed6644edb98dd306c5572cbc Reviewed-on: https://skia-review.googlesource.com/4534 Commit-Queue: Mike Reed <reed@google.com> Reviewed-by: Florin Malita <fmalita@chromium.org>
2016-11-15 21:44:34 +00:00
const char* modeName = SkBlendMode_Name(mode);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
fLabelPaint.setTextAlign(SkPaint::kRight_Align);
add drawString helper to canvas Many tests and examples use drawText with a guess of how long the text is in bytes, or a call to strlen(). Add a helper to SkCanvas to simplify these examples. Add another helper for SkString. R=reed@google.com Change-Id: I0204a31e938f065606f08ee7cd9a6b36db791ee2 Reviewed-on: https://skia-review.googlesource.com/13642 Commit-Queue: Cary Clark <caryclark@google.com> Reviewed-by: Cary Clark <caryclark@google.com> Reviewed-by: Mike Reed <reed@google.com> Reviewed-by: Cary Clark <caryclark@skia.org>
2017-04-28 19:35:12 +00:00
canvas->drawString(modeName, kLabelSpacing - kShapeSize / 4,
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
fLabelPaint.getTextSize() / 4, fLabelPaint);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
}
Revert[8] "replace SkXfermode obj with SkBlendMode enum in paints" This reverts commit c245574ba3d0e2ade6c94b2812de3baa383bf4c4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2396953002 TBR= Review-Url: https://codereview.chromium.org/2396953002
2016-10-06 00:33:02 +00:00
void setupShapePaint(SkCanvas* canvas, GrColor color, SkBlendMode mode, SkPaint* paint) {
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
paint->setColor(color);
Revert[8] "replace SkXfermode obj with SkBlendMode enum in paints" This reverts commit c245574ba3d0e2ade6c94b2812de3baa383bf4c4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2396953002 TBR= Review-Url: https://codereview.chromium.org/2396953002
2016-10-06 00:33:02 +00:00
if (mode == SkBlendMode::kPlus) {
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
// Check for overflow, otherwise we might get confusing AA artifacts.
int maxSum = SkTMax(SkTMax(SkColorGetA(kBGColor) + SkColorGetA(color),
SkColorGetR(kBGColor) + SkColorGetR(color)),
SkTMax(SkColorGetG(kBGColor) + SkColorGetG(color),
SkColorGetB(kBGColor) + SkColorGetB(color)));
if (maxSum > 255) {
SkPaint dimPaint;
dimPaint.setAntiAlias(false);
Revert[8] "replace SkXfermode obj with SkBlendMode enum in paints" This reverts commit c245574ba3d0e2ade6c94b2812de3baa383bf4c4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2396953002 TBR= Review-Url: https://codereview.chromium.org/2396953002
2016-10-06 00:33:02 +00:00
dimPaint.setBlendMode(SkBlendMode::kDstIn);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
if (255 != paint->getAlpha()) {
// Dim the src and dst colors.
dimPaint.setARGB(255 * 255 / maxSum, 0, 0, 0);
paint->setAlpha(255 * paint->getAlpha() / maxSum);
} else {
// Just clear the dst, we need to preserve the paint's opacity.
dimPaint.setARGB(0, 0, 0, 0);
}
hide trivial helpers on canvas BUG=skia: Change-Id: I42d4ca92897bde5bd8e0575a0104b12b83134cef Reviewed-on: https://skia-review.googlesource.com/8852 Reviewed-by: Cary Clark <caryclark@google.com> Commit-Queue: Mike Reed <reed@google.com>
2017-02-22 18:21:42 +00:00
canvas->drawRect({ -kShapeSpacing/2, -kShapeSpacing/2,
kShapeSpacing/2 + 3 * kShapeTypeSpacing, kShapeSpacing/2 },
dimPaint);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
}
}
}
Revert[8] "replace SkXfermode obj with SkBlendMode enum in paints" This reverts commit c245574ba3d0e2ade6c94b2812de3baa383bf4c4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2396953002 TBR= Review-Url: https://codereview.chromium.org/2396953002
2016-10-06 00:33:02 +00:00
void drawShape(SkCanvas* canvas, Shape shape, const SkPaint& paint, SkBlendMode mode) {
SkASSERT(mode <= SkBlendMode::kLastMode);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
SkPaint shapePaint(paint);
shapePaint.setAntiAlias(kSquare_Shape != shape);
Revert[8] "replace SkXfermode obj with SkBlendMode enum in paints" This reverts commit c245574ba3d0e2ade6c94b2812de3baa383bf4c4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2396953002 TBR= Review-Url: https://codereview.chromium.org/2396953002
2016-10-06 00:33:02 +00:00
shapePaint.setBlendMode(mode);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
switch (shape) {
case kSquare_Shape:
hide trivial helpers on canvas BUG=skia: Change-Id: I42d4ca92897bde5bd8e0575a0104b12b83134cef Reviewed-on: https://skia-review.googlesource.com/8852 Reviewed-by: Cary Clark <caryclark@google.com> Commit-Queue: Mike Reed <reed@google.com>
2017-02-22 18:21:42 +00:00
canvas->drawRect({ -kShapeSize/2, -kShapeSize/2, kShapeSize/2, kShapeSize/2 },
shapePaint);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
break;
case kDiamond_Shape:
canvas->save();
canvas->rotate(45);
hide trivial helpers on canvas BUG=skia: Change-Id: I42d4ca92897bde5bd8e0575a0104b12b83134cef Reviewed-on: https://skia-review.googlesource.com/8852 Reviewed-by: Cary Clark <caryclark@google.com> Commit-Queue: Mike Reed <reed@google.com>
2017-02-22 18:21:42 +00:00
canvas->drawRect({ -kShapeSize/2, -kShapeSize/2, kShapeSize/2, kShapeSize/2 },
shapePaint);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
canvas->restore();
break;
case kOval_Shape:
canvas->save();
Revert[8] "replace SkXfermode obj with SkBlendMode enum in paints" This reverts commit c245574ba3d0e2ade6c94b2812de3baa383bf4c4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2396953002 TBR= Review-Url: https://codereview.chromium.org/2396953002
2016-10-06 00:33:02 +00:00
canvas->rotate(static_cast<SkScalar>((511 * (int)mode + 257) % 360));
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
canvas->drawPath(fOval, shapePaint);
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
canvas->restore();
break;
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
case kConcave_Shape:
canvas->drawPath(fConcave, shapePaint);
break;
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
default:
SkFAIL("Invalid shape.");
}
}
private:
SkPaint fLabelPaint;
Added concave paths to aaxfermodes GM GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1534823004 Review URL: https://codereview.chromium.org/1534823004
2015-12-17 18:58:28 +00:00
SkPath fOval;
SkPath fConcave;
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
Revert "Revert "replace Arithmetic xfermode with imagefilter"" This reverts commit 10ff5bfa789b6b602464e8511fdf676c6f5b1bd4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2680 TBR= Change-Id: Iac2f4d48d227ff426ccf02b7eae280f382ad3580 Reviewed-on: https://skia-review.googlesource.com/2680 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com>
2016-09-26 22:51:24 +00:00
typedef skiagm::GM INHERITED;
Implement Porter Duff XP with a blend table Removes the runtime logic used by PorterDuffXferProcessor to decide blend coeffs and shader outputs, and instead uses a compile-time constant table of pre-selected blend formulas. Separates out the dst read fallback into its own XP. Introduces a new blend strategy for srcCoeff=0 that can apply coverage with a reverse subtract blend equation instead of dual source blending. Adds new macros in GrBlend.h to analyze blend formulas both runtime. Removes kSetCoverageDrawing_OptFlag and GrSimplifyBlend as they are no longer used. Adds a GM that verifies all xfermodes, including arithmetic, with the color/coverage invariants used by Porter Duff. Adds a unit test that verifies each Porter Duff formula with every color/coverage invariant. Major changes: * Uses a reverse subtract blend equation for coverage when srcCoeff=0 (clear, dst-out [Sa=1], dst-in, modulate). Platforms that don't support dual source blending no longer require a dst copy for dst-in and modulate. * Sets BlendInfo::fWriteColor to false when the blend does not modify the dst. GrGLGpu will now use glColorMask instead of blending for these modes (dst, dst-in [Sa=1], modulate ignored for [Sc=1]). * Converts all SA blend coeffs to One for opaque inputs, and ISA to Zero if there is also no coverage. (We keep ISA around when there is coverage because we use it to tweak alpha for coverage.) * Abandons solid white optimizations for the sake of simplicity (screen was the only mode that previous had solid white opts). Minor differences: * Inconsequential differences in opt flags (e.g. we now return kCanTweakAlphaForCoverage_OptFlag even when there is no coverage). * Src coeffs when the shader outputs 0. * IS2C vs IS2A when the secondary output is scalar. BUG=skia: Committed: https://skia.googlesource.com/skia/+/9a70920db22b6309c671f8e5d519bb95570e4414 Review URL: https://codereview.chromium.org/1124373002
2015-05-27 22:08:33 +00:00
};
Revert "Revert "replace Arithmetic xfermode with imagefilter"" This reverts commit 10ff5bfa789b6b602464e8511fdf676c6f5b1bd4. BUG=skia: GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2680 TBR= Change-Id: Iac2f4d48d227ff426ccf02b7eae280f382ad3580 Reviewed-on: https://skia-review.googlesource.com/2680 Reviewed-by: Mike Reed <reed@google.com> Commit-Queue: Mike Reed <reed@google.com>
2016-09-26 22:51:24 +00:00
DEF_GM( return new AAXfermodesGM; )
Reference in New Issue Copy Permalink