We were only respecting this feature in runtime shaders. Note that use
of any tagged matrices will cause color filter creation to fail, but
color transformation is a totally sensible thing to want in a color
filter.
Change-Id: I482226b287ab794cb341367fce453381cb581966
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308507
Reviewed-by: Mike Klein <mtklein@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Some of these checks are currently redundant (we don't allow color
filters to have children right now). But the next CL will re-add that
capability, and the unit tests here will ensure we don't re-break things
by allowing child-sampling to violate the color filter invariant.
Change-Id: I54c10d8b1d1e376c13347296765185d42b9f644a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308285
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
For direct masks, the rectangle passed through the SubRun and the rectangle in the atlas have
the same dimensions. This information is redundant. Instead of passing the glyph origin in pos,
pass the top left corner of the glyph. In onPrepare, use the dimension information to produce the
entire glyph bounding box.
Change-Id: I135d4a144366b39560742797572641e96fe482df
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/307842
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Eventually the small path renderer will have an atlas manager it
can only access through the flushState.
This is heading towards:
https://skia-review.googlesource.com/c/skia/+/307776 (Split the small path renderer into record-time and flush-time pieces)
Change-Id: Ie8ad7d421575548294c78dcb4623cb465952a07e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308281
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This is a reland of c5f8e7df81
Original change's description:
> Switch GN defs on Mac and iOS to use frameworks, take 2.
>
> GN will be removing support for frameworks in the libs variable, so
> might as well get ahead of it.
>
> Change-Id: I990ffdb945c85a9868883d0cab1c2bdcdbdf545c
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308189
> Reviewed-by: Mike Klein <mtklein@google.com>
> Commit-Queue: Jim Van Verth <jvanverth@google.com>
Change-Id: Ie93c4b6a5e82bd61738367c68c2a2bac7bdf23f1
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308199
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Android migration landed in Android CL 12234077
Chrome migration is landing in Chrome CL 2335812
Note: makeFromCompressedTexture is not used by Chrome.
Bug: skia:104662
Change-Id: Ibbe6d412cf22e87188926383d10b21f780208e48
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305102
Commit-Queue: Adlai Holler <adlai@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Auto-Submit: Adlai Holler <adlai@google.com>
This function isn't used by Chrome so we migrate directly.
Flutter migration is at https://github.com/flutter/engine/pull/20235
Bug: skia:104662
Change-Id: I9d875acdbd162f50a6d86b3a4cae3f400e4dd38f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305180
Commit-Queue: Adlai Holler <adlai@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
This is just a mechanical cut & paste.
When the atlasing is separate from the small path renderer both the
atlas manager and the path renderer need to share these objects.
This is pulled out of the omnibus CL:
https://skia-review.googlesource.com/c/skia/+/307776 (Split the small path renderer into record-time and flush-time pieces)
Change-Id: Ibae9d785fa8a3934fa81f9e98d7e1027d48dc5a3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308276
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
6c873cfde9..bedac4f06b
2020-08-06 syoussefi@chromium.org Move EXT_external_objects validation to validationESEXT.cpp
2020-08-05 timvp@google.com Vulkan: Don't always end the render pass when updating the scissor
2020-08-05 ynovikov@chromium.org Make roll_chromium_deps executable
2020-08-05 ynovikov@chromium.org Rename roll_deps to roll_chromium_deps
2020-08-05 syoussefi@chromium.org Vulkan: Support ANGLE_external_objects_flags
2020-08-05 angle-autoroll@skia-public.iam.gserviceaccount.com Roll SwiftShader from 74cc8ea2b8df to be7c55a2a8ce (2 revisions)
2020-08-05 syoussefi@chromium.org Add back glslang include check
2020-08-05 syoussefi@chromium.org Entry points for ANGLE_external_objects_flags
2020-08-05 syoussefi@chromium.org Spec language for ANGLE_external_objects_flags
2020-08-05 angle-autoroll@skia-public.iam.gserviceaccount.com Roll SPIRV-Tools from 92a71657fcba to 8bc27a1cfbc1 (3 revisions)
2020-08-05 timvp@google.com Roll chromium_revision fb6cbaf6e5..ed7912a248 (790363:794870)
2020-08-05 jmadill@chromium.org Vulkan: Squash Texture+ImageView Serial and improve caching.
2020-08-05 nguyenmh@google.com Batch capture run + change how results are logged
If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/angle-skia-autoroll
Please CC herb@google.com on the revert to ensure that a human
is aware of the problem.
To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug
Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md
Cq-Include-Trybots: skia/skia.primary:Build-Debian10-Clang-x86_64-Release-ANGLE;skia/skia.primary:Test-Win10-Clang-AlphaR2-GPU-RadeonR9M470X-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-Golo-GPU-QuadroP400-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC5i7RYH-GPU-IntelIris6100-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC6i5SYK-GPU-IntelIris540-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC8i5BEK-GPU-IntelIris655-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUCD34010WYKH-GPU-IntelHD4400-x86_64-Debug-All-ANGLE
Bug: chromium:1104516
Tbr: herb@google.com
Test: Test: CQ
Change-Id: I0fffe515042cf25559ac2222d2554eb8419c305a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308417
Reviewed-by: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
This reverts commit c5f8e7df81.
Reason for revert: This breaks CLion.
Original change's description:
> Switch GN defs on Mac and iOS to use frameworks, take 2.
>
> GN will be removing support for frameworks in the libs variable, so
> might as well get ahead of it.
>
> Change-Id: I990ffdb945c85a9868883d0cab1c2bdcdbdf545c
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308189
> Reviewed-by: Mike Klein <mtklein@google.com>
> Commit-Queue: Jim Van Verth <jvanverth@google.com>
TBR=mtklein@google.com,jvanverth@google.com,senorblanco@google.com
Change-Id: I5479693596f9bb07686325489b58489e89f60c05
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308338
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Bug: skia:10570
Change-Id: I257380fda62d99c7c27efabf2105478d20f23c57
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308336
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
The SkVM backend can draw all the existing GMs.
Image diffs look good. Rough local perf changes:
spiral_rt: 14.3ms -> 5.9ms w/o JIT, or 2.5ms w/JIT
overdrawcolorfilter: 4.2 4.9* 2.6
overdraw_canvas: 10.1 8.9 2.7
runtimecolorfilter: 8.5 8.5 7.5
with one regression marked with a *.
Change-Id: Ib4f11422f3c2e8756960794a965385fa616ad8bf
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308193
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
This lets us notice that an SkRasterPipelineSpriteBlitter failed to set
up its stages (e.g. an SkColorFilter it can't handle) and instead of
using a NullBlitter, give SkVMSpriteBlitter a chance, just like we do in
the non-sprite blitters.
This is arguably working around the SkSpriteBlitter helper abstract
class' limitations, and removing that intermediate to SkBlitter might
make things clearer in the end. Notice how SkVMSpriteBlitter is an
SkBlitter directly, not an SkSpriteBlitter, and it gets all its
information at once rather than split between constructor and setup().
We could make all the sprite blitters like that.
Change-Id: I905c744fdcf7b76e8a62634e8d7eb96d9d0ed391
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308277
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
GN will be removing support for frameworks in the libs variable, so
might as well get ahead of it.
Change-Id: I990ffdb945c85a9868883d0cab1c2bdcdbdf545c
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308189
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Change-Id: I9ed98859814e462c63ab29b94f0365ccc57d2e9b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/307706
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Previously, `takeOwnership` was overloaded to take Symbols or IRNodes.
However, Symbols themselves are a subclass of IRNode, so attempting to
call `takeOwnership` with any subclass of Symbol was considered
ambiguous. This was fixed by splitting the API into
`takeOwnershipOfSymbol` and `takeOwnershipOfIRNode`, letting the caller
decide explicitly rather than forcing the compiler to choose.
These APIs also returned their input back to the caller as
`const Symbol*` and `const IRNode*`, although generally the caller
actually wanted the object in its original type. This meant most uses
of the return type were immediately cast back to their actual type.
(Some callers accidentally stripped const as well, although it would be
wrong to mutate an object in the symbol table.) Added templates to
allow the input to be returned as its original type, and removed
the now-unnecessary casts.
The String `takeOwnership` call was updated to `takeOwnershipOfString`
for consistency with its peers.
Change-Id: Ib44a08097d89caec2c30deb4a25463330b7d2bf2
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/307841
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
I think we decided at some point that we had implemented all existing
color filters with SkRasterPipeline, and then stopped checking for
appendStages() failure. It's all still plumbed through, and I'd like to
start making some color filters fail to appendStage() and use
SkVMBlitter, at least some of the time.
Change-Id: I520151e7bfc01c4cf333a6ed515f27bee6cd55ee
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308195
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
We can still serialize things, but we have to rely on the type-name
code-path (because the factory methods don't exist).
Change-Id: I7a054b07168859154c691b9672e760e5b658ae50
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308260
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
This reverts commit 07438b0cda.
Bug: skia:10369
Bug: skia:10371
This will allow Skia clients developing for Android 11+ to rely on
Android's NDK APIs for decoding, which will allow them to decode
without including their own decoding libraries (e.g. libjpeg-turbo).
Using these APIs also provides support for static HEIF images.
Run ImageGenSrc in kPlatform_Mode on Android to verify decoding
visually.
Add tests and a grayscale png.
Update some test bots running Android R to specify ndk_api so they will
run the new code.
Change-Id: I4ca07d832dbd6a9d8cff0faea975fd70da00718f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308185
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
As noted at line 341, the SmallPathOp doesn't reliably know the
proxies that will be used by the atlas when visitProxies is called.
Deleting this code gives us a bit more flexibility for upcoming
refactorings.
Change-Id: I0292e750ed46c8555a75d1b5fa2710f1f616b70b
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308220
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
SkOverdrawColorFilter is already implemented in an
identical way to overdrawcolorfilter_runtime, so it's
kind of just testing two copies of the same thing.
Change-Id: I8d3b28649ae1c686633ba8e8e1017e8dc5100b3d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308192
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
This also follows up on the comments Rob left in his zombie review.
Bug: skia:104662
Change-Id: I7031f7b2832eb3f47025e5d269248b854cc912e0
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/307780
Reviewed-by: Robert Phillips <robertphillips@google.com>
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
Auto-Submit: Adlai Holler <adlai@google.com>
The most interesting part of the CL is that we recheck gSkVMAllowJIT in
Program::eval() even though we've already checked it in the constructor.
This allows Viewer to toggle the JIT on and off without having to worry
about program caching. This is not something that you'd expect to come
up in practice if a program just sets gSkVMAllowJIT at the start of
main(); for real clients I think we can avoid all this with a simple
SkGraphics::allowJIT() that only lets clients opt-in, never back out.
I toyed with making '!' rotate through a tristate in Viewer, until I
realized that these really are independent bits: GMs like threshold_rt
that use both ordinary effects and SkVM-only effects demonstrate
different behavior and performance in all four modes. So '!' continues
to toggle SkVMBlitter, and now '@' toggles the JIT.
I've left the test program default settings unchanged, with the JIT
enabled unless --nojit is passed. Where we previously simplified the
command line by conflating --dylib with --skvm, we now conflate --dylib
with --jit.
Change-Id: If86bf524c657298c0846bcd33c706e3c3f91e788
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308184
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Incidentally this also adds the guard for the case where the request is
coming from GrContext::setBackendTextureState. I still thing the client
should be tracking what state requests go into the GrContext call so that
they are not requesting illegal transitions. However, it is easier to
make the fix here for surface that guards both then try to separate out
the two calls so that we only guard the surface version.
Bug: chromium:1108040
Change-Id: I5af0afe033eea55df19efb68f97a6a5aaa9f2da3
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308178
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Auto-Submit: Greg Daniel <egdaniel@google.com>
The root cause of this error is that Metal does not support upcasting
a small matrix into a larger matrix (I was unfamiliar with this GLSL
ability). Proper support in Metal for that type of cast will be added in
a followup CL. For now, this CL adds defensive bounds-checking to
placate ASAN.
Change-Id: Ieb0d6b14f9bc7ecc9ef6223364c606612dc79e43
Bug: oss-fuzz:22776
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308182
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
This reverts commit cfef980939.
Reason for revert: Breaking Google3 roll
Original change's description:
> Add an SkImageGenerator that uses NDK APIs
>
> Bug: skia:10369
> Bug: skia:10371
>
> This will allow Skia clients developing for Android 11+ to rely on
> Android's NDK APIs for decoding, which will allow them to decode
> without including their own decoding libraries (e.g. libjpeg-turbo).
> Using these APIs also provides support for static HEIF images.
>
> Run ImageGenSrc in kPlatform_Mode on Android to verify decoding
> visually.
>
> Add tests and a grayscale png.
>
> Update some test bots running Android R to specify ndk_api so they will
> run the new code.
>
> Change-Id: Ica782339b2414d472ede0b61729a127ce41892a5
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305689
> Reviewed-by: Mike Klein <mtklein@google.com>
> Commit-Queue: Leon Scroggins <scroggo@google.com>
TBR=djsollen@google.com,mtklein@google.com,scroggo@google.com,brianosman@google.com,reed@google.com
Change-Id: Ifed506a76a0ff5903d101c1bf7330d319b8376a6
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:10369
Bug: skia:10371
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308180
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
This is a 'Create' and returns a reference. Be sure to release the
reference.
Bug: skia:10542
Change-Id: If4ed7a37a43efe5f8f17a4a15a0f82cdae7d9276
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/307795
Auto-Submit: Ben Wagner <bungeman@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Bug: skia:10369
Bug: skia:10371
This will allow Skia clients developing for Android 11+ to rely on
Android's NDK APIs for decoding, which will allow them to decode
without including their own decoding libraries (e.g. libjpeg-turbo).
Using these APIs also provides support for static HEIF images.
Run ImageGenSrc in kPlatform_Mode on Android to verify decoding
visually.
Add tests and a grayscale png.
Update some test bots running Android R to specify ndk_api so they will
run the new code.
Change-Id: Ica782339b2414d472ede0b61729a127ce41892a5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/305689
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
709472c9dd..6c873cfde9
2020-08-05 xinyi.he@arm.com Disable RobustBufferAccessWebGL2ValidationStateChangeTest on Mali
2020-08-05 jmadill@chromium.org Vulkan: Redo Sampler Serials.
2020-08-04 jmadill@chromium.org Vulkan: Refactor ImageViewHelper serials.
2020-08-04 timvp@google.com Create empty //tools/android/md5sum/BUILD.gn
2020-08-04 ccameron@chromium.org macOS/ARM: Allow populating GPU based on AGXAccelerator
2020-08-04 syoussefi@chromium.org Vulkan: EXT_multisampled_render_to_texture2 support
2020-08-04 syoussefi@chromium.org Vulkan: Fix blit/resolve of mixed-samples framebuffers
2020-08-04 jonahr@google.com Suppress SurfacelessContext test on Android/NVIDIA
2020-08-04 cnorthrop@google.com Tests: Remove more optimizations from traces
2020-08-04 angle-autoroll@skia-public.iam.gserviceaccount.com Roll glslang from 928b7b26bd73 to f881f08358b0 (2 revisions)
2020-08-04 angle-autoroll@skia-public.iam.gserviceaccount.com Roll SPIRV-Tools from b78f4b1518e0 to 92a71657fcba (1 revision)
2020-08-04 angle-autoroll@skia-public.iam.gserviceaccount.com Roll Vulkan-Headers from a0f1f5149d91 to a6d08c75d219 (1 revision)
2020-08-04 angle-autoroll@skia-public.iam.gserviceaccount.com Roll SwiftShader from 4568d8587998 to 74cc8ea2b8df (2 revisions)
If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/angle-skia-autoroll
Please CC herb@google.com on the revert to ensure that a human
is aware of the problem.
To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug
Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md
Cq-Include-Trybots: skia/skia.primary:Build-Debian10-Clang-x86_64-Release-ANGLE;skia/skia.primary:Test-Win10-Clang-AlphaR2-GPU-RadeonR9M470X-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-Golo-GPU-QuadroP400-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC5i7RYH-GPU-IntelIris6100-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC6i5SYK-GPU-IntelIris540-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUC8i5BEK-GPU-IntelIris655-x86_64-Debug-All-ANGLE;skia/skia.primary:Test-Win10-Clang-NUCD34010WYKH-GPU-IntelHD4400-x86_64-Debug-All-ANGLE
Bug: chromium:1110408,chromium:1110421
Tbr: herb@google.com
Test: Test: angle_perftests --gtest_filter="TracePerfTest.Run/*nba*"Test: Test: gn gen out/Android
Change-Id: Id3c460176cb85f7fdf1959d4ffbba21b70ab1204
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/308057
Reviewed-by: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: skia-autoroll <skia-autoroll@skia-public.iam.gserviceaccount.com>
Since the fibonacci sequence is always multiplied by
fFirstHeapAllocationSize, change the names ofr fFib0 and
fFib1 to fNextHeapAlloc, and fYetNextHeapAlloc, and assign
both to be fFirstHeapAllocationSize. This will create the
same sequence of heap allocation sizes. Now
fFirstHeapAllocationSize is only used in reset(). This will
simplify CL/307348 which move reset to a different class.
Change-Id: I54da1980239645acd36992476915137dd68ab9da
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/307349
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Using a direct mask, the pixels in the cache are in 1:1 correspondence to the device pixels.
There is no need to scale.
Change-Id: I7d3f48cfb44a51882370527401a3bd0d162d9e76
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/307875
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Herb Derby <herb@google.com>
This was broken by the addition of the sk_sp visualizer.
Change-Id: Ibf7519a8a2e0388275e543a1e5a69d942d50ed96
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/307867
Commit-Queue: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
SK_IGNORE_MAC_OPSZ_FORCE should be removed since it is no longer defined
by any users.
Change-Id: Ib99082513292e55465131a42a7f676c5d1179113
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/307863
Reviewed-by: Mike Klein <mtklein@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Auto-Submit: Ben Wagner <bungeman@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
Commit-Queue: Herb Derby <herb@google.com>
AFAICT, there were two unrelated things happening here:
1) .fp files supported referencing class variables as if they existed in
sk_Args. We never do this (they can be directly referenced).
2) When baking the values of 'in' variables into runtime SkSL, we
generated intermediate 'Setting' objects that wrapped the literal
values supplied by the user. Those Setting objects were named
'sk_Args.<field_name>', even though this logic had nothing to do with
the fake SkArgs type. In this case, the Setting objects don't appear
to add any value, so just inject the literals directly (which is what
constant propagation would do on the next iteration).
Change-Id: Ia54193c5282737653cf10b7aae77e38e5ab78c79
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/307793
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
