This arrangement allows the backend texture to outlive the YCbCr SkImage.
Change-Id: I34939d05bf1091c8efcacb687dc1900729d4cbe5
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/296478
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Previously, when REPORTER_ASSERT reported a failure, there was no
additional information describing the FP that could not be cloned in the
error log. Now, we print out a very simple tree of the FP and its
children.
Change-Id: I141cfb17ca431864a6f555d56f0335293f259c4e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/296452
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
In the "ignore input" mode, the input FP contributes nothing and is
never sampled. In the "modulate input" cases, the input FP is sampled
as one would expect.
Change-Id: I96717d63d8e3d7ef6aa4eaaf88154c6e5ce47e55
Bug: skia:10217
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/296299
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
When collapsing static switches, we were not handling the symbol table
in SkSL::Block correctly, which led to assertion failures in some cases.
This CL is based on the fixes in http://review.skia.org/296178
but removes the need to track pointers-to-unique-pointers.
We do this by splitting the work into two parts:
1 - determine range of statements to move
2 - actually move statements
Because the statements are all consecutive, keeping track of this range
is not actually that difficult and we don't need to do any checks twice.
Change-Id: I71ad745ef1e4b4f5f6b753762e65fa49b2399adc
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/296440
Commit-Queue: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
In PDF files, "names" and "strings" are not the same thing, but I was
conflating them. Separate out the interfaces for adding attributes to
PDF struct tree elements so there's a way to add either a name or a string,
and similarly for arrays of names or arrays of strings.
Fix the table test to correctly use a name for the "Scope" attribute
and an array of strings for the "Headers" attribute.
Bug: chromium:607777
Change-Id: Ib30bded2bbcf96e31ba6925fb062615558dea0db
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/296338
Reviewed-by: Ben Wagner <bungeman@google.com>
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Auto-Submit: Dominic Mazzoni <dmazzoni@chromium.org>
Commit-Queue: Derek Sollenberger <djsollen@google.com>
This will remove common boilerplate from our gen-code, and gives us a
place to put common child-cloning boilerplate.
Change-Id: I6101655af89d4c5844ec908b81ce4f6e5d59f834
Bug: skia:10217
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/296177
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Change-Id: I44f64a8c98c019a8f4878b0b6f6d82489aa8252c
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/296179
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Mike Reed <reed@google.com>
We also fix getReadSwizzle with this change.
Change-Id: I1989d8347dc97d7a4c75aa9094a0146419c6d8fa
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295819
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
This new api will eventually replace the version that takes an
SkSurface::BackendAccess.
Change-Id: I48cd013725e14027f386b0b111223459944ac44a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295567
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Options used to rely on an external call to SanitizeOptions to
ensure they were correct. Now that defaults are set directly,
make sure that the values coming in make sense.
Change-Id: If6cfc027722b6a7717a920b482ec5be8f7526367
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/296040
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Herb Derby <herb@google.com>
If GrClips know their dimensions then getConservativeBounds() does not
need any arguments and isRRect() can remove its rtBounds argument.
I also updated GrFixedClip to report the render target bounds as a
degenerate rrect in its isRRect implementation if it was wide open. Its
apply() function was also simplified to take advantage of the prior
GrScissorState work where the rectangle was always valid to access and
contained within the render target bounds.
Change-Id: I627b97976cb176b1c80627462027034b06ad2cb2
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/290957
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
This CL is not fully comprehensive; for instance, it does not contain
fixes for backends that don't compile on Mac. But it does resolve the
vast majority of cases that trigger -Wimplicit-fallthrough.
A few minor bugs were found and fixed, but none that were likely to
affect normal operation.
Change-Id: I43487602b0d56200ce8b42702e04f66390d82f60
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295916
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
The old factories (makeShader) will behave as before: they will inherit
the filter-quality from the paint.
The new factory takes an explicit filter setting, and will use that
regardless of the paint.
Big follow-ups:
- update callers to not rely on setting in SkPaint
- revise/enhance settings in imageshader
- settings for scaling up and down
- control over trilerp, etc.
- other: 4x4 kernels? trilerp bias?
- move mipmaps to always be explicit requests a SkImage factory time
Bug: skia:10344
Change-Id: If87b06d4fd6eafd8b9cdecda7c00d69897066ef8
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295086
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
This should make the YCbCr sampler helper usable in a GM (with a normal Vk context).
Change-Id: I75451f6ca934f7b59c48349c77234856d0946a12
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295766
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
The GrTextContext::options would flow from the GrRecordingContext to
the GrDrawingManager to the GrTextContext, and finally to the
GrRenderTargetContext.
Just find them on the GrRecording context off the GrRenderTargetContext.
Change-Id: I902481d20072c2470905261ab81c2b6456f25343
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295559
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
The current implementation ignores return value from
`registerChildProcessor` and, surprisingly, assumes that a cloned FP
index will match the original FP index. This version honors the return
value.
(In practice, I have not seen any cases where the current implementation
has caused actual breakage.)
Updating common code-gen had large ripple effects in the SkSL unit
tests. While repairing the tests, I also took the opportunity to use
raw-strings to pass the source SkSL text, and annotated the `expectedH`
and `expectedCPP` blocks to make the tests easier to understand at a
glance.
Change-Id: I71be69d9e4620963b3ef49ad8e0dba3b40af7f4f
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295452
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
When collapsing static switches down to a single statement, we detect
break statements and don't copy them. But the logic was broken; we
weren't copying the entire statement in which the break occurred, which
could be a block, causing some of the code to simply be omitted.
Change-Id: Ic5b59c11d12326c93d49080193a0a5297732bfb0
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295776
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Reviewed-by: John Stiles <johnstiles@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Bug: skia:10362
Change-Id: I56d1f618dcdf96133f1932b7ac8d4602ddb93a59
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295575
Commit-Queue: Stephen White <senorblanco@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
The aim here is to unify VkTestHelper with the VkYcbcrSamplerHelper's context creation code. AFAICT they have a ~90% overlap.
Change-Id: Iba8d1482b8c1d7164682f90e19d2183d8cfe45d4
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295583
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
Now that we have the GrContext setBackendTextureState calls we no longer
need these to be on flush.
Bug: skia:10254
Change-Id: I7c44667a327de11dd853e3e71b114959a7bcee86
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295447
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Bug: b/145995037
Bug: skia:5457
Android is switching to the new behavior with ag/10041691, so the
legacy flag is no longer needed.
TBR=djsollen@google.com
Change-Id: I786807e4cb8b4ed4cd41fd2776ae2915f8bc62cc
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/263176
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
After removing all unused API from SkReader32, it only had a handful
of functions, and it was (rightly) only used by SkReadBuffer. Remove
the temptation to use it by just folding it into SkReadBuffer.
SkWriter32 had some unnecessary functions as well (only used in unit
tests), so those are gone. There is still a strange relationship:
SkWriteBuffer is just an interface - SkReadBuffer is actually the
complement of SkBinaryWriteBuffer/SkWriter32. Those two classes produce
data in the exact same format, but with slightly different interfaces.
(The choice about which one is used is mostly about high-level
serialization vs. low-level helpers).
Change-Id: I1e823755febecd2e053ea732b21295d8f4d9d832
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295557
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
The intent is to reuse this helper to create a YCbCr GM.
Change-Id: I4d6af42745dbf845e28753bec670ad4a75c393b2
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295443
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
These are much safer than SkReader32/SkWriter32 (they do validation and
ensure we never read past the end of the buffer).
Where we used to just assert that the contents of the cache were valid,
we now validate everything, and fail gracefully by discarding the cache
contents if it's corrupted or invalid.
Reland includes a new skipByteArray API. The previous technique for
reading into an std::string relied on data(), which doesn't return a
writeable pointer until the C++17 standard library.
Bug: skia:9402
Change-Id: I3b88efbf8ca590c8ad4f8164f7c07eee12696ec6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/295441
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This is currently only supported for the Vulkan backend
Bug: skia:10254
Change-Id: I9274799098dc00dec5abcbcec95ce7cc23fec537
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/293844
Commit-Queue: Greg Daniel <egdaniel@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
From what I can tell no one is currently depending on this feature and
continuing to support it is starting to become a pain. Clients will
be able to manually request VkImages (via flush api) go back to their
original queue, so there is a path if someone really needs this.
Bug: skia:10254
Change-Id: I91283bba451fa9bf52b466f587d3f7c5c9365242
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/294657
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
This check isn't correct, since most backends will use a fallback path if
writePixels directly to a surface fails.
That fallback path doesn't work for Dawn currently, so skip the test
on the Dawn backend.
Change-Id: I07f9a25640bc579d6b0952e208237368d007b06d
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/294702
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Change-Id: I0af800900a7fbd9d16af0058ee0754358ebc3875
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/293562
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Dawn does not support writePixels to a TextureAsRenderTarget, so update its caps to reflect that.
Also update the WritePixelsNonTexture_Gpu test to check the caps.
Bug: skia:10333
Change-Id: I4dcf2e0fecc34cba9586e5cab71739e427301efb
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/294597
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Change-Id: Ifbe0f3ae6e01d65f18351903da8aef63170ce6c7
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/294457
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Auto-Submit: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Adds a method to determine the worst-case number of tessellated line
segments that a path might require, and disables hardware tessellation
if it is more segments than are supported (falling back on indirect
draw shaders).
If the path requires even more segments than are supported by the
indirect draw shaders (1024), we crop the path to the viewport. The
required number of segments is proportional to the square root of the
bounding box's diagonal, so we won't start cropping paths until their
device-space bounding box diagonal is nearly 175,000 pixels long.
Change-Id: I8a9435e70bb93dda3464cc11a3e44fbe511744ae
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/293691
Reviewed-by: Greg Daniel <egdaniel@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
This reverts commit 4926b07217.
Reason for revert: fix wip
Original change's description:
> Revert "Improve scissor state tracking in GrRTC"
>
> This reverts commit 3b923a880b.
>
> Reason for revert: GrAppliedHardClip isn't tracking scissor state properly
>
> Original change's description:
> > Improve scissor state tracking in GrRTC
> >
> > At a low level, this changes GrScissorState from a rect+bool to a rect+size.
> > The scissor test is considered enablebd if the rect does not fill the
> > device bounds rect specified by the size. This has a number of benefits:
> >
> > 1. We can always access the scissor rect and know that it will be
> > restricted to the render target dimensions.
> > 2. It helps consolidate code that previously had to test the scissor rect
> > and render target bounds separately.
> > 3. The clear operations can now match the proper backing store dimensions
> > of the render target.
> > 4. It makes it easier to reason about scissors applying to the logical
> > dimensions of the render target vs. its backing store dimensions.
> >
> > Originally, I was going to have the extra scissor guards for the logical
> > dimensions be added in a separate CL (with the cleanup for
> > attemptQuadOptimization). However, it became difficult to ensure correct
> > behavior respecting the vulkan render pass bounds without applying this
> > new logic at the same time.
> >
> > So now, with this CL, GrAppliedClips are sized to the backing store
> > dimensions of the render target. GrOpsTasks also clip bounds to the
> > backing store dimensions instead of the logical dimensions (which seems
> > more correct since that's where the auto-clipping happens). Then when
> > we convert a GrClip to a GrAppliedClip, the GrRTC automatically enforces
> > the logical dimensions scissor if we have stencil settings (to ensure
> > the padded pixels don't get corrupted). It also may remove the scissor
> > if the draw was just a color buffer update.
> >
> > Change-Id: I75671c9cc921f4696b1dd5231e02486090aa4282
> > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/290654
> > Commit-Queue: Michael Ludwig <michaelludwig@google.com>
> > Reviewed-by: Brian Salomon <bsalomon@google.com>
>
> TBR=bsalomon@google.com,csmartdalton@google.com,michaelludwig@google.com
>
> Change-Id: Ie98d084158e3a537604ab0fecee69bde3e744d1b
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/294340
> Reviewed-by: Michael Ludwig <michaelludwig@google.com>
> Commit-Queue: Michael Ludwig <michaelludwig@google.com>
TBR=bsalomon@google.com,csmartdalton@google.com,michaelludwig@google.com
# Not skipping CQ checks because this is a reland.
Change-Id: I2116e52146890ee4b7ea007f3c3d5c3e532e4bdd
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/294257
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
This reverts commit 3b923a880b.
Reason for revert: GrAppliedHardClip isn't tracking scissor state properly
Original change's description:
> Improve scissor state tracking in GrRTC
>
> At a low level, this changes GrScissorState from a rect+bool to a rect+size.
> The scissor test is considered enablebd if the rect does not fill the
> device bounds rect specified by the size. This has a number of benefits:
>
> 1. We can always access the scissor rect and know that it will be
> restricted to the render target dimensions.
> 2. It helps consolidate code that previously had to test the scissor rect
> and render target bounds separately.
> 3. The clear operations can now match the proper backing store dimensions
> of the render target.
> 4. It makes it easier to reason about scissors applying to the logical
> dimensions of the render target vs. its backing store dimensions.
>
> Originally, I was going to have the extra scissor guards for the logical
> dimensions be added in a separate CL (with the cleanup for
> attemptQuadOptimization). However, it became difficult to ensure correct
> behavior respecting the vulkan render pass bounds without applying this
> new logic at the same time.
>
> So now, with this CL, GrAppliedClips are sized to the backing store
> dimensions of the render target. GrOpsTasks also clip bounds to the
> backing store dimensions instead of the logical dimensions (which seems
> more correct since that's where the auto-clipping happens). Then when
> we convert a GrClip to a GrAppliedClip, the GrRTC automatically enforces
> the logical dimensions scissor if we have stencil settings (to ensure
> the padded pixels don't get corrupted). It also may remove the scissor
> if the draw was just a color buffer update.
>
> Change-Id: I75671c9cc921f4696b1dd5231e02486090aa4282
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/290654
> Commit-Queue: Michael Ludwig <michaelludwig@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
TBR=bsalomon@google.com,csmartdalton@google.com,michaelludwig@google.com
Change-Id: Ie98d084158e3a537604ab0fecee69bde3e744d1b
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/294340
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
At a low level, this changes GrScissorState from a rect+bool to a rect+size.
The scissor test is considered enablebd if the rect does not fill the
device bounds rect specified by the size. This has a number of benefits:
1. We can always access the scissor rect and know that it will be
restricted to the render target dimensions.
2. It helps consolidate code that previously had to test the scissor rect
and render target bounds separately.
3. The clear operations can now match the proper backing store dimensions
of the render target.
4. It makes it easier to reason about scissors applying to the logical
dimensions of the render target vs. its backing store dimensions.
Originally, I was going to have the extra scissor guards for the logical
dimensions be added in a separate CL (with the cleanup for
attemptQuadOptimization). However, it became difficult to ensure correct
behavior respecting the vulkan render pass bounds without applying this
new logic at the same time.
So now, with this CL, GrAppliedClips are sized to the backing store
dimensions of the render target. GrOpsTasks also clip bounds to the
backing store dimensions instead of the logical dimensions (which seems
more correct since that's where the auto-clipping happens). Then when
we convert a GrClip to a GrAppliedClip, the GrRTC automatically enforces
the logical dimensions scissor if we have stencil settings (to ensure
the padded pixels don't get corrupted). It also may remove the scissor
if the draw was just a color buffer update.
Change-Id: I75671c9cc921f4696b1dd5231e02486090aa4282
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/290654
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
These tests create a wgpu::Texture which outlives the wgpu::Device
which created it. This is problematic on Dawn, so just skip
the tests for now.
Bug: skia:10326
Change-Id: I29d5328e313baca19a2bbc2eedd80f2768cae77e
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/293936
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Bug: skia:10178
These functions can be performed generically using SkRasterPipeline
or skcms. Further, the reason we used a function pointer anyway was
so that we could call the same function on each row separately. But
libwebp's API doesn't let us do a single row at a time anyway.
Simplify this method by using readPixels when necessary and
skipping conversion entirely when possible.
Add support for encoding from unpremul 4444. It is simpler to support
it, and it's not obvious why we didn't support it before.
Keep the behavior of not supporting A8, and apply the same to the
other alpha-only formats. Note that we could support encoding such an
image to alpha, r=0, g=0, b=0, but I'd rather leave adding that
feature to a separate change, which enables it for all encoders (and
accounts for the internal use of PNGs as a round-trip for
kAlpha_8_SkColorType).
Add GMs to test the newly supported SkColorTypes.
Change-Id: I4d86c5621792fb6dc3cb68b736a1eb35d577e3a6
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/292962
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
These classes are much safer (there's no way to safely deserialize a
string with SkReader32 without knowledge of how it works internally).
Prior to this CL, SkVertices was the only complex type that had manual
serialization using the lower level types - now it works like everything
else. Additionally: the versioning can now be tied to picture versions
going forward (like everything else).
Bug: oss-fuzz:22909
Bug: oss-fuzz:22918
Bug: skia:9984
Bug: skia:10304
Change-Id: I3cf537eb765b5c8ce98b554c0f200e5d67c33d14
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/293349
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Bug: 932080
Bug: b/142252770
An ICO file has a directory of images that are stored later in the file.
The directory contains the offset and size of the images. SkIcoCodec
uses these to create embedded SkPng/SkBmpCodecs. The old implementation
allocated a block of memory for each image and copied the stream into
those blocks so that the embedded SkCodecs could independently read
their encoded data.
Although SkIcoCodec checks for null, this still allows large (albeit
temporary - since we'll discard them if the stream does not contain
enough data to fill them) allocations and the potential for over-
commit.
Instead, read the entire stream into a contiguous buffer. If the stream
is already actually a buffer, just use that directly. In this case, the
new code will do less work. Otherwise, the memory we allocate is
limited by the size of the stream.
Note that this is a behavior change for a stream that contains two
consecutive ICOs, where the client expects to be able to read the second
one later. This was an issue for PNGs on Android (b/34073812), but I
suspect no one is relying on this behavior for ICO. Update Codec_end
test to remove the ICO test.
Alternatives to consider:
- only buffer the individual encoded images. This will allow us to
continue passing the Codec_end test.
- lazily read the embedded streams. Currently we read their start to
verify they are valid images (at least in the header) and read their
actual sizes and bit-depths, which could differ from that listed in
the directory. We use those to make a guess at the "best" image to use.
An image with mismatched sizes may now decode differently.
Change-Id: I30e5f6c8c2e5a0fa135348f61efe151a7f5d4756
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/277058
Auto-Submit: Leon Scroggins <scroggo@google.com>
Commit-Queue: Derek Sollenberger <djsollen@google.com>
Reviewed-by: Derek Sollenberger <djsollen@google.com>
This reverts commit 59aa4b7187.
Reason for revert: ASAN failures were actually unrelated to this CL.
Original change's description:
> Revert "Remove double support from SkSL."
>
> This reverts commit 71a35d49b7.
>
> Reason for revert: ASAN failures on swarming bots
>
> Original change's description:
> > Remove double support from SkSL.
> >
> > Doubles are not supported by Metal or GLSL pre-4.0, are not supported in
> > most backends, and aren't used in any GMs. There isn't any good way to
> > use them in new code as it would just degrade to float on many of our
> > supported platforms. (This is assuming that our backends actually know
> > how to degrade doubles to floats, which is not universally the case.)
> >
> > Change-Id: Ieacc69db4bdacca104a15a6eef33e05f977d1ffa
> > Bug: skia:10299
> > Reviewed-on: https://skia-review.googlesource.com/c/skia/+/292846
> > Commit-Queue: John Stiles <johnstiles@google.com>
> > Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
> > Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
> > Auto-Submit: John Stiles <johnstiles@google.com>
>
> TBR=bsalomon@google.com,ethannicholas@google.com,johnstiles@google.com
>
> Change-Id: I175e42420bcae8dfacd0bfeb269dd84e0b3c9d25
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Bug: skia:10299
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/293268
> Reviewed-by: John Stiles <johnstiles@google.com>
> Commit-Queue: John Stiles <johnstiles@google.com>
TBR=bsalomon@google.com,ethannicholas@google.com,johnstiles@google.com
# Not skipping CQ checks because this is a reland.
Bug: skia:10299
Change-Id: I5cdd71a1512228175514a0d29e19ae91afc78b6a
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/293273
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
Auto-Submit: John Stiles <johnstiles@google.com>
This reverts commit 71a35d49b7.
Reason for revert: ASAN failures on swarming bots
Original change's description:
> Remove double support from SkSL.
>
> Doubles are not supported by Metal or GLSL pre-4.0, are not supported in
> most backends, and aren't used in any GMs. There isn't any good way to
> use them in new code as it would just degrade to float on many of our
> supported platforms. (This is assuming that our backends actually know
> how to degrade doubles to floats, which is not universally the case.)
>
> Change-Id: Ieacc69db4bdacca104a15a6eef33e05f977d1ffa
> Bug: skia:10299
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/292846
> Commit-Queue: John Stiles <johnstiles@google.com>
> Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
> Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
> Auto-Submit: John Stiles <johnstiles@google.com>
TBR=bsalomon@google.com,ethannicholas@google.com,johnstiles@google.com
Change-Id: I175e42420bcae8dfacd0bfeb269dd84e0b3c9d25
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:10299
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/293268
Reviewed-by: John Stiles <johnstiles@google.com>
Commit-Queue: John Stiles <johnstiles@google.com>
This reverts commit 97fe0cbed2.
Reason for revert: ASAN failures
Original change's description:
> Omit dead SkSL functions
>
> Now that SkSL inlines functions, dead functions are very common. This
> change causes them to be omitted from the final output.
>
> Change-Id: Ie466a3f748812eff1a368498365c89d73ab0b7be
> Reviewed-on: https://skia-review.googlesource.com/c/skia/+/292684
> Reviewed-by: Brian Osman <brianosman@google.com>
> Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
TBR=bsalomon@google.com,brianosman@google.com,ethannicholas@google.com
Change-Id: Id20c5be67dd574d30d6f978ba610e43aa5018416
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/293241
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Now that SkSL inlines functions, dead functions are very common. This
change causes them to be omitted from the final output.
Change-Id: Ie466a3f748812eff1a368498365c89d73ab0b7be
Reviewed-on: https://skia-review.googlesource.com/c/skia/+/292684
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
