- remove ALWAYS_INLINE until we find we need it
- make bit_puns explicit
- implement everything recursively so, e.g.
sqrt(float8) picks up sqrt(float4) when
not otherwise specialized.
- implement SSE specializations:
of the operations I tested, only sqrt, rcp, and rsqrt
needed any help. The others look good as-is.
Change-Id: I1b679c7bd9a99f952272b118d7ade2469b55d604
Reviewed-on: https://skia-review.googlesource.com/c/190222
Auto-Submit: Mike Klein <mtklein@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
GrBuffer is a base class for GrGpuBuffer and GrCpuBuffer. GrGpuBuffer is a
GrGpuResource and the others are not. This allows GrCpuBuffers to exist
outside of the GrGpuResourceCache.
Also removes flags from GrResourceProvider buffer factory function. The
only flag still in use was kRequireGpuMemory. Now CPU buffers are made
without using GrResourceProvider.
Change-Id: I82670d1316e28fd6331ca36b26c8c4ead33846f9
Reviewed-on: https://skia-review.googlesource.com/c/188823
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Change-Id: Ie576bc8442e5a646c33191100be1be5f89d72290
Reviewed-on: https://skia-review.googlesource.com/c/189483
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
There are two fixes here. The first pads out the uniform buffer to
match the alignment of the largest element in the buffer, as required
by Metal. The second ensures that the RTHeight uniform is only added
to the shader if required by the program inputs, as otherwise it wasn't
being included in the uniform data.
Bug: skia:8737
Change-Id: Icbe57456f3ec79faada231278c58137ca9865dd2
Reviewed-on: https://skia-review.googlesource.com/c/189875
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Greg Daniel <egdaniel@google.com>
I happened to be disassembling a test object file and noticed
that this was emitted in every test object file.
Change-Id: I1d09f5bb80d9f8a5530dc9f81a2dd95872e34a10
Reviewed-on: https://skia-review.googlesource.com/c/190202
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Bug: skia:
Change-Id: I0a3b3d607e1b74787480b03520a07839c87486dc
Reviewed-on: https://skia-review.googlesource.com/c/189980
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Centralize these for my sanity. Most will also be parceled out to other contexts.
Change-Id: If0e7e98bcf66c4d8a3391f9b04e643ccc91af4ad
Reviewed-on: https://skia-review.googlesource.com/c/189488
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Change-Id: Ie3e5b353f84e74d398a5350dc0baff5541789119
Reviewed-on: https://skia-review.googlesource.com/c/189982
Commit-Queue: Mike Klein <mtklein@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Auto-Submit: Mike Klein <mtklein@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Obviously lots of these new operations like sqrt() will want platform
specialization. That'll come later.
Change-Id: Ia0758425d4ec5911968a3d0ad63fa387b9b4cb39
Reviewed-on: https://skia-review.googlesource.com/c/189848
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Auto-Submit: Mike Klein <mtklein@google.com>
Change-Id: I1cb8113af243ed6327179d295835295834a752aa
Reviewed-on: https://skia-review.googlesource.com/c/189581
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Auto-Submit: Mike Klein <mtklein@google.com>
This is a reland of 91c1d08bc3
Original change's description:
> SkSL is now pickier about type conversions
>
> Bug: skia:
> Change-Id: I4e8b8f229f4e4344f160b0dbb41832764d0b75bd
> Reviewed-on: https://skia-review.googlesource.com/c/188311
> Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
Bug: skia:
Change-Id: I727cad061afc0a5ee6f4d2df789330d809dd110a
Reviewed-on: https://skia-review.googlesource.com/c/189643
Reviewed-by: Ethan Nicholas <ethannicholas@google.com>
Commit-Queue: Ethan Nicholas <ethannicholas@google.com>
Bug: skia:
Change-Id: I43a0136d67471030056da2af19894c28bc96d22b
Reviewed-on: https://skia-review.googlesource.com/c/189495
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Renamed to GrGpuBufferType in anticipation of splitting GrBuffer
into GrGpuBuffer and GrCpuBuffer types.
There were two unused values in the enum that are removed, DrawIndirect
and Texel.
Change-Id: Icb6b3da689adbd8e10495c10fd0470a6ee0120b5
Reviewed-on: https://skia-review.googlesource.com/c/189280
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Mechanical. This makes the priv() accessor the same for all the context types.
Change-Id: I40850eb05a33b8d7cc3eabdd42226d24b2ba58aa
Reviewed-on: https://skia-review.googlesource.com/c/189164
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
This enables four different options in the compiler, described
below. I also added enough masks to satisfy RTCc when running
all GMs in both 8888 and gl configs.
---
/RTCc - Detects when a value is assigned to a smaller data
type and results in data loss. This happens even when casting.
Masking is required to suppress this.
/RTCs - Various stack-related checks, including uninitialized
data (by initializing locals to a non-zero value), array bounds
checking, and stack pointer corruption that can occur with a
calling convention mismatch.
/RTCu - Reports when a variable is used without having been
initialized. Mostly redundant with compile-time checks.
/guard:cf - This is more of a security option, that computes
all possible targets for indirect calls at compile time, and
verifies that those are the only targets reached at compile
time. Also generates similar logic around switch statements
that turn into jump tables.
Bug: skia:
Change-Id: I7b527af8fd67dec0b6556f38bcd0efc3fd505856
Reviewed-on: https://skia-review.googlesource.com/c/188625
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Bug: skia:
Change-Id: I280556d4691a762402d2a94959bd6fda614bc39c
Reviewed-on: https://skia-review.googlesource.com/c/189168
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Added as part of https://skia-review.googlesource.com/c/skia/+/132261/
As I understand it, this is no longer needed to prevent linear blending.
Change-Id: Ieddf1a86f2d5c6c874217549099a4168d2c44b47
Reviewed-on: https://skia-review.googlesource.com/c/189166
Commit-Queue: Leon Scroggins <scroggo@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Auto-Submit: Leon Scroggins <scroggo@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Now that we're keeping refs on GrBuffers during flush, we shouldn't
need to specify this flag.
Change-Id: Idbe08e84d23690d6c9e94f23f2c4507a2e965aff
Reviewed-on: https://skia-review.googlesource.com/c/188635
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
The GrContext's ID isn't really unique any more (since it can be shared among a family of contexts). Change its name to reflect the new reality.
Additionally, no client seems to be using it so make it private.
Change-Id: Ibb9004d699fe6ca7876b3be94142e612b5b9efbd
Reviewed-on: https://skia-review.googlesource.com/c/188308
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
The issue is around needing to go from a float rect to a clipped (intersected
with the clip bounds) irect.
The old code promoted the clip bounds to rect,
intersected with the rect, and then rounded to ints. This is problematic
since convert float --> int can produce float values that are larger than
their int counter-parts. (e.g. int:67108863 becomes float:67108864).
This error goes on to confuse subsequent logic.
The fix is to round the rect to int first (knowing that very large floats
will saturate to maxint), and then intersect.
Bug: 927075
Change-Id: If6c7d72ff15184700b87532e1922e576d1348ee8
Reviewed-on: https://skia-review.googlesource.com/c/188307
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Florin Malita <fmalita@chromium.org>
This copies the hardware buffer unit test from, which was cherry picked
straight into skqp last year, into Skia proper.
There are not functional changes, but a few of the APIs have been updated
to work with newer Skia.
Bug: skia:
Change-Id: I2d7b2ed8b0b9314ca3e03e703a6a5ac53805275e
Reviewed-on: https://skia-review.googlesource.com/c/188034
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
SkiaRenderer does not delete promise image textures when they are released
but not done. Refulfilling promise image textures takes a significant amount
of CPU time. This allows us to fulfill each promise image once.
Bug: skia:8736
Change-Id: I7ad7fa9678ed0ec4bb714b71fbf920ab4a845409
Reviewed-on: https://skia-review.googlesource.com/c/188039
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Bug: b/122338652
Bug: b/77339676
getSampledDimensions is supposed to report dimensions that
SkAndroidCodec can decode to. But if SkAndroidCodec is respecting an
orientation that requires swapping the width and the height,
getSampledDimensions needs to do so, too.
Change-Id: I0cacefc2b02eb9832b8c029c88bcaaedcb2ba82b
Reviewed-on: https://skia-review.googlesource.com/c/188035
Auto-Submit: Leon Scroggins <scroggo@google.com>
Commit-Queue: Derek Sollenberger <djsollen@google.com>
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Change-Id: I03d15a563eeb7ea187c7d48deffe00357dc55177
Reviewed-on: https://skia-review.googlesource.com/c/188032
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
This begins the process of splitting GrContext into:
GrContext_Base, GrImageContext, GrRecordingContext and GrDirectContext.
Change-Id: I3c43045f2a5549b049e95791d65f74d4e16de36f
Reviewed-on: https://skia-review.googlesource.com/c/186878
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
No longer used by Chrome.
Change-Id: I842482490f132bb3aa3e85f847d8d4167f5595df
Reviewed-on: https://skia-review.googlesource.com/c/188029
Commit-Queue: Robert Phillips <robertphillips@google.com>
Auto-Submit: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
This just shuffles stuff around.
Change-Id: I90e24a73bf44dd3e88fc121a0528bf94e4fdead4
Reviewed-on: https://skia-review.googlesource.com/c/187786
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This just shuffles stuff around.
Change-Id: Ieab35f50945efe87512d7077cb994132f0e0b6ef
Reviewed-on: https://skia-review.googlesource.com/c/186874
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This reverts commit 3c03c851ad.
Modification was to copy/assign fBitfieldsUInt rather than fBitfields,
since fBitfields might be smaller, and therefore ignore some bits that
we do use in hashing, etc.
Bug: skia:
Change-Id: I5965d5d64bd22bd60f0da797fdfeeffb730588ea
Reviewed-on: https://skia-review.googlesource.com/c/187307
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Mike Reed <reed@google.com>
kUnbudgetedCacheable now means that the resource is never purged
until its unique key is removed.
This fixes an issue where a cached texture for a promise image
might get purged by cache pressure. This in turn could cause
Skia to call the promise image's Fulfill proc multiple times with
no intervening Release calls. The balancing Release calls would
occur, but the policy is that each Fulfill should be balanced by
Release *before* another Fulfill.
Update/add unit tests.
Bug: chromium:922851
Change-Id: I6411e413b3104721ca4bb6e7f07b3b73d14cbcf9
Reviewed-on: https://skia-review.googlesource.com/c/186361
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Make all wrapped resources be kUnbudgetedUncacheable except those
created by AHardwareBuffer image generators and as backings for promise
images.
Make all non-wrapped unbudgeted resources be kUnbudgetedUncacheable.
Update unit tests to mostly use GrWrapCacheable::kNo except where they
are testing the distinction.
Update unit tests for new expectations.
Bug: chromium:922851
Change-Id: I4d3bdaa161ffc76390f26334bcb7e2b47dd9319d
Reviewed-on: https://skia-review.googlesource.com/c/185004
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Store budgeted status as a tri-state enum on GrGpuResource:
kBudgeted
kUnbudgetedCacheable
kUnbudgetedUncacheable
Uncacheable vs. Cacheable captures the current distinction between
wrapped resources created with purgeImmediately or !purgeImmediately.
Non-wrapped and unbudgeted resources are all kUnbudgetedCacheable to
match current behavior.
This change just introduces the new types. No behavior is changed.
Bug: chromium:922851
Change-Id: Ic2387bf321cf9b56b4c9ffd9dbef8ade60f9cb98
Reviewed-on: https://skia-review.googlesource.com/c/185003
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Most interfaces had migrated to skcms_TransferFunction. Having both was
awkward in several places, so this (almost) finishes the migration. Some
clients are calling SkICC::WriteToICC, so that's left intact. After this
CL, those clients can switch to using SkWriteICCProfile directly, and
WriteToICC can be deleted.
Bug: skia:
Change-Id: I46ebaeb1f5b20bf0c620e8a620e73ee323a1de31
Reviewed-on: https://skia-review.googlesource.com/c/186541
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
* Remove GetTypefaceOrDefault from SkPaint and SkFont
* Remove RefTypefaceOrDefault from SkPaint and SkFont
Change-Id: I04ae777142c2bdec849508b611b844418bbaedff
Reviewed-on: https://skia-review.googlesource.com/c/185781
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Enforces this constraint from one central location, rather than
relying on each Op to remember to check if overlap is allowed from its
onCombineIfPossible method.
Fixes an issue where we need to check the total bounds of both chains
for overlap (not the bounds of individual Ops).
Bug: skia:8671
Change-Id: I163651c868847884459acfc00d13ffdfca3a27c3
Reviewed-on: https://skia-review.googlesource.com/c/185815
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Note: This removes from the public API. The two most commonly
checked values from gammaNamed can still be tested with
gammaCloseToSRGB and gammaIsLinear.
Others can be tested for (exact or approximate) equality by
calling isNumericalTransferFn, and testing the resulting
skcms_TransferFunction in any method the client wants.
Bug: skia:
Change-Id: I9b63ffea81bba461795cad946050c7083b02eec8
Reviewed-on: https://skia-review.googlesource.com/c/185684
Reviewed-by: Brian Osman <brianosman@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This reverts commit 5d052dac3a.
Reason for revert: BREAKS CHROMIUM ROLL.
https://cs.chromium.org/chromium/src/third_party/blink/renderer/platform/fonts/win/font_platform_data_win.cc?l=49
| FAILED: obj/third_party/blink/renderer/platform/platform/font_platform_data_win.obj
| ../../third_party/blink/renderer/platform/fonts/win/font_platform_data_win.cc(49,29):
| error: no member named 'kLCDRenderText_Flag' in 'SkPaint'
| if (text_flags & SkPaint::kLCDRenderText_Flag) {
| ~~~~~~~~~^
| ../../third_party/blink/renderer/platform/fonts/win/font_platform_data_win.cc(51,36):
| error: no member named 'kAntiAlias_Flag' in 'SkPaint'
| } else if (text_flags & SkPaint::kAntiAlias_Flag) {
| ~~~~~~~~~^
| ../../third_party/blink/renderer/platform/fonts/win/font_platform_data_win.cc(56,52):
| error: no member named 'kSubpixelText_Flag' in 'SkPaint'
| font->setSubpixel(SkToBool(text_flags & SkPaint::kSubpixelText_Flag));
| ~~~~~~~~~^
| ../../third_party/blink/renderer/platform/fonts/win/font_platform_data_win.cc(64,29):
| error: no member named 'kAntiAlias_Flag' in 'SkPaint'
| if (text_flags & SkPaint::kAntiAlias_Flag)
| ~~~~~~~~~^
| ../../third_party/blink/renderer/platform/fonts/win/font_platform_data_win.cc(85,27):
| error: no member named 'kAntiAlias_Flag' in 'SkPaint'
| ? SkPaint::kAntiAlias_Flag
| ~~~~~~~~~^
| ../../third_party/blink/renderer/platform/fonts/win/font_platform_data_win.cc(91,35):
| error: no member named 'kLCDRenderText_Flag' in 'SkPaint'
| ? SkPaint::kLCDRenderText_Flag
| ~~~~~~~~~^
| ../../third_party/blink/renderer/platform/fonts/win/font_platform_data_win.cc(93,27):
| error: no member named 'kAntiAlias_Flag' in 'SkPaint'
| text_flags = SkPaint::kAntiAlias_Flag | lcd_flag;
| ~~~~~~~~~^
| ../../third_party/blink/renderer/platform/fonts/win/font_platform_data_win.cc(101,28):
| error: no member named 'kAntiAlias_Flag' in 'SkPaint'
| text_flags |= SkPaint::kAntiAlias_Flag;
| ~~~~~~~~~^
| 8 errors generated.
Original change's description:
> hide hinting and flag enums with PAINT_FLAGS
>
> A preflight for https://skia-review.googlesource.com/c/skia/+/185460
>
> Bug: skia:2664
> Change-Id: I83f7f9cd500e8507b559462835477ba8ae7c11ab
> Reviewed-on: https://skia-review.googlesource.com/c/185540
> Reviewed-by: Mike Reed <reed@google.com>
> Commit-Queue: Mike Reed <reed@google.com>
> Auto-Submit: Mike Reed <reed@google.com>
TBR=reed@google.com
Change-Id: I4d1581a7e29267ad459ed31d4885b823fa00066f
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:2664
Reviewed-on: https://skia-review.googlesource.com/c/185680
Reviewed-by: Hal Canary <halcanary@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>
A later CL will make this return a unique_ptr<SkDocument>.
Bug: skia:5972
Change-Id: Ie10d6c07d5f2524ecb71d906db0d37427827225d
Reviewed-on: https://skia-review.googlesource.com/c/181660
Reviewed-by: Hal Canary <halcanary@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>
fulfills a different SkImage.
Bug: skia:8613
Change-Id: I7ee14112c69f8aaef223a37dda455259b501a2bf
Reviewed-on: https://skia-review.googlesource.com/c/184440
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Brings back GL ETC1 support and adds Vulkan support as well.
Bug: skia:8684
Change-Id: Ie65da0a3172793081f0e4072f161bfb9b14678bc
Reviewed-on: https://skia-review.googlesource.com/c/179724
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
This looks like a relic from nvpr text.
Bug: skia:8671
Change-Id: I417c7b4565208e3bf91a3891fa59ba50308a8139
Reviewed-on: https://skia-review.googlesource.com/c/183884
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
All of this extra code/parameter-passing for fonts to the paint unflattener
can go away when we can drop support for older skps
Bug: skia:
Change-Id: I4d626f0925a54564b7c386ed80953a0291c8bc1a
Reviewed-on: https://skia-review.googlesource.com/c/183863
Reviewed-by: Florin Malita <fmalita@chromium.org>
Commit-Queue: Mike Reed <reed@google.com>
motivation: The Document has become intrusive enough that we almost
always have a pointer to the Document on hand anyways.
Also: update document.
Also: forward declare more things in headers.
Also, don't try to clean up resources when abort() or close() is called.
It is easy enough for the client to delete the Document when done.
Change-Id: I21aeed37f26ba16d68af041d9317adf00e9a61f2
Reviewed-on: https://skia-review.googlesource.com/c/180646
Commit-Queue: Hal Canary <halcanary@google.com>
Reviewed-by: Herb Derby <herb@google.com>
This makes the API easier to use in Chrome.
It is no longer required to pass the SkPromiseImageTexture to the
release proc.
Bug: skia:
Change-Id: I6636401f6a7915d3ad15e890718638bc91a58cc4
Reviewed-on: https://skia-review.googlesource.com/c/183383
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Bug: skia:
Change-Id: Ib029f337f5e61366e2550e77dc99310b44d03f84
Reviewed-on: https://skia-review.googlesource.com/c/182970
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
The SkImages aren't guaranteed to have a full fledged GrContext (e.g., if they are promise images created outside of a DDL recorder). This CL starts the process if lessening Skia's reliance on what ever the SkImage has cached.
Change-Id: I0181d09417e1ea2ced7706a722e2815f3ceee870
Reviewed-on: https://skia-review.googlesource.com/c/182974
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
Change-Id: I21a5ed81d3e161ddf8a838ea9ac0414265b00ae6
Reviewed-on: https://skia-review.googlesource.com/c/182701
Auto-Submit: Herb Derby <herb@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Herb Derby <herb@google.com>
This reverts commit 559c617137.
Reason for revert: breaking things
Original change's description:
> Reuse GrTexture instances when the same GrBackendTexture is used to
> repeatedly fulfill a promise SkImage.
>
> Bug: skia:8613
>
> Change-Id: I35c76435d630d2daa034e0c3efb59666bfd6882a
> Reviewed-on: https://skia-review.googlesource.com/c/175820
> Commit-Queue: Brian Salomon <bsalomon@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
TBR=bsalomon@google.com,robertphillips@google.com
Change-Id: I7548809945d0a875fdb9387398bbc45e733c0846
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:8613
Reviewed-on: https://skia-review.googlesource.com/c/182960
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Unlike other ops, always use half-float colors. The logic around the
Instance struct makes dynamically switching color size tricky.
CCPR stores color in a per-instance attribute though, so the cost of
always using FP16 is much lower.
Bug: skia:
Change-Id: I9c0c64940f74f915a18417a5830030558e065d28
Reviewed-on: https://skia-review.googlesource.com/c/182760
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
This file was adjusted 3 times, beginning here
commit 212e9060ed
Author: Mike Reed <reed@google.com>
Date: Tue Dec 25 17:35:49 2018 -0500
to try to update it to use SkFont. Along the way we broke the part where
it compares drawing with the typeface pre and post streaming. This CL
restores the original test, and then updated it to use SkFont.
After this CL, we should be able to move SK_SUPPORT_LEGACY_PAINT_FONT_FIELDS
to clients, having them opt-in.
Bug: skia:
Change-Id: Id786160322f8530f5d0c65b30a0c228139b39d61
Reviewed-on: https://skia-review.googlesource.com/c/182805
Reviewed-by: Mike Reed <reed@google.com>
Auto-Submit: Mike Reed <reed@google.com>
This should help avoid confusion from T**.
Change-Id: I1851baa2a55714721fa935d234b6a4a1c6d6504f
Reviewed-on: https://skia-review.googlesource.com/c/182562
Commit-Queue: Mike Klein <mtklein@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Auto-Submit: Mike Klein <mtklein@google.com>
Reviewed-by: Herb Derby <herb@google.com>
This is a reland of 5820b0c3f3
It is updated in patchset 2 to clean up pointers passed into memcpy, and to
optimize the bounds calculation in GrPerspQuad. This should fix a performance
regression caused by the move away from caching 1/w. The Sk4f::invert() does not
always preserve 1/1 == 1, which led to bounds slightly outside of clips and
thus forced Skia to keep the scissor test enabled. The fix also restores the
optimization of skipping the 1/w division when the quad is known to be 2D.
Original change's description:
> Use specialized quad lists in rectangle ops
>
> Hopefully reduces memory footprint of GrFillRectOp and GrTextureOp
>
> The original rect code (GrAAFillRectOp) stored 2 SkMatrices (18 floats), 2
> SkRects (8 floats) an SkPMColor4f (4 floats) and a flag (1 int) for a total
> of 124 bytes per quad that was stored in the op.
>
> The first pass at the rectangle consolidation switched to storing device and
> local quads as GrPerspQuads (32 floats), an SkPMColor4f (4 floats) and a flag
> (1 int) for a total of 148 bytes per quad. After landing, several memory
> regressions appeared in Chrome and our perf monitor.
>
> Several intertwined approaches are taken here. First, GrPerspQuad no longer
> caches 1/w, which makes a quad 12 floats instead of 16. Second, a specialized
> list type is defined that allows storing the x, y, and extra metadata together
> for quads, but keeps the w components separate. When the quad type isn't
> perspective, w is not stored at all since it is implicitly 1 and can be
> reconstituted at tessellation time. This brings the total per quad to either
> 84 or 116 bytes, depending on if the op list needs perspective information.
>
> Bug: chromium:915025
> Bug: chromium:917242
> Change-Id: If37ee122847b0c32604bb45dc2a1326b544f9cf6
> Reviewed-on: https://skia-review.googlesource.com/c/180644
> Commit-Queue: Michael Ludwig <michaelludwig@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
Bug: chromium:915025, chromium:917242
Change-Id: I98a1bf83fd7d393604823d567c57d7e06fad5e55
Reviewed-on: https://skia-review.googlesource.com/c/182203
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Chris Dalton <csmartdalton@google.com>
for f in $(git grep -l drawSimpleText); do
sed -i 's/wSimpleText(\([[:alnum:]]\), strlen(\1), kUTF8_SkTextEncoding,/wString(\1,/g' "$f"
sed -i 's/wSimpleText(\([[:alnum:]]*\)\.c_str(), \1\.size(), kUTF8_SkTextEncoding,/wString(\1,/g' "$f"
done
git checkout @~ include/core/SkCanvas.h samplecode/SampleCusp.cpp
git cl format
Change-Id: Ibc0c3afd30f86465a645fadb26177d61919f57fd
Reviewed-on: https://skia-review.googlesource.com/c/181561
Reviewed-by: Hal Canary <halcanary@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>
Auto-Submit: Hal Canary <halcanary@google.com>
This is a reland of 6a3dc8be46
Original change's description:
> ccpr: Support caching of paths that span multiple tiles
>
> Adds an accumulative "hit rect" for each cache entry that tracks the
> region of the path that has been drawn during its lifetime. Now, a
> path mask can be cached once the "hit rect" covers 50% of the path.
> This allows us to cache a path that spans multiple tiles.
>
> To guard against unnecessarily caching gigantic path masks, we also
> require that 10% of the path be visible during the draw when it is
> cached.
>
> Bug: skia:8462
> Change-Id: Iab2c277102b7a774eaa909c9663211694554c5a5
> Reviewed-on: https://skia-review.googlesource.com/c/180700
> Commit-Queue: Chris Dalton <csmartdalton@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
Bug: skia:8462
Change-Id: Ia2b10430acd2dffac78b5abd432763ead79bc902
Reviewed-on: https://skia-review.googlesource.com/c/181983
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
This reverts commit 6f44647e52.
Reason for revert: checking to see if this is blocking the Chrome roll
Original change's description:
> Make SkBitmap/SkPixmap::erase* do so in sRGB
>
> We generally consider untagged colors to be sRGB, so this makes us more
> consistent with other parts of the API.
>
> Add a test.
>
> Cq-Include-Trybots: luci.chromium.try:linux-blink-rel
> Change-Id: I5468c86ad92164797a65ffd9fbe471e01a97a2ca
> Reviewed-on: https://skia-review.googlesource.com/c/179245
> Commit-Queue: Leon Scroggins <scroggo@google.com>
> Reviewed-by: Mike Reed <reed@google.com>
> Reviewed-by: Mike Klein <mtklein@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
TBR=djsollen@google.com,mtklein@google.com,scroggo@google.com,brianosman@google.com,reed@google.com
Change-Id: Ia592adf2c790d294da1e32c1e83f9f34e81d79cc
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Cq-Include-Trybots: luci.chromium.try:linux-blink-rel
Reviewed-on: https://skia-review.googlesource.com/c/182083
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This reverts commit 6a3dc8be46.
Reason for revert: <INSERT REASONING HERE>
Original change's description:
> ccpr: Support caching of paths that span multiple tiles
>
> Adds an accumulative "hit rect" for each cache entry that tracks the
> region of the path that has been drawn during its lifetime. Now, a
> path mask can be cached once the "hit rect" covers 50% of the path.
> This allows us to cache a path that spans multiple tiles.
>
> To guard against unnecessarily caching gigantic path masks, we also
> require that 10% of the path be visible during the draw when it is
> cached.
>
> Bug: skia:8462
> Change-Id: Iab2c277102b7a774eaa909c9663211694554c5a5
> Reviewed-on: https://skia-review.googlesource.com/c/180700
> Commit-Queue: Chris Dalton <csmartdalton@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
TBR=bsalomon@google.com,robertphillips@google.com,brianosman@google.com,csmartdalton@google.com
Change-Id: Ibae9d46333e3178856fd623f26317366102dd344
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:8462
Reviewed-on: https://skia-review.googlesource.com/c/181982
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
We generally consider untagged colors to be sRGB, so this makes us more
consistent with other parts of the API.
Add a test.
Cq-Include-Trybots: luci.chromium.try:linux-blink-rel
Change-Id: I5468c86ad92164797a65ffd9fbe471e01a97a2ca
Reviewed-on: https://skia-review.googlesource.com/c/179245
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Adds an accumulative "hit rect" for each cache entry that tracks the
region of the path that has been drawn during its lifetime. Now, a
path mask can be cached once the "hit rect" covers 50% of the path.
This allows us to cache a path that spans multiple tiles.
To guard against unnecessarily caching gigantic path masks, we also
require that 10% of the path be visible during the draw when it is
cached.
Bug: skia:8462
Change-Id: Iab2c277102b7a774eaa909c9663211694554c5a5
Reviewed-on: https://skia-review.googlesource.com/c/180700
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
This is a reland of d6fa45472c
Original change's description:
> ccpr: Rework the path cache to support sporadic flushing
>
> Removes the notion of a stashed atlas that we store from the previous
> flush. Now we just cache every atlas we ever render. Cached atlases
> can either be 16-bit or 8-bit.
>
> The "reuse" and "animation" cases should both behave exactly the same
> as before: Where before we would copy from the stashed atlas to 8-bit
> atlases, we now copy from a cached 16-bit atlas and then invalidate
> it. Where before we would recycle the stashed atlas's backing texture
> object, we now recycle this same texture object from an invalidated
> 16-bit cached atlas.
>
> The main difference is that cases like tiled rendering now work. If
> you draw your whole scene in one flush, you still get one big 16-bit
> cached atlas, just like the "stashed atlas" implementation. But if you
> draw your scene in tiles, you now get lots of little cached 16-bit
> atlases, which can be reused and eventually copied to 8-bit atlases.
>
> Bug: skia:8462
> Change-Id: Ibae65febb948230aaaf1f1361eef9c8f06ebef18
> Reviewed-on: https://skia-review.googlesource.com/c/179991
> Commit-Queue: Chris Dalton <csmartdalton@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
Bug: skia:8462
Change-Id: I2f64b0c37e2cd644a202dfc786366dda5d238391
Reviewed-on: https://skia-review.googlesource.com/c/181450
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Moved named common transfer functions and gamuts to constexpr values in
SkColorSpace.h, in SkNamedTransferFn and SkNamedGamut namespaces.
Converted nearly all SkColorSpace::MakeRGB calls within Skia to use the
new factory with the named values. Multiple clients want a way to
extract named transfer function and gamut - this still doesn't provide
that, but this may be a better path forward for honestly advertising how
SkColorSpace works internally.
Originally landed as:
https://skia.googlesource.com/skia/+/a9549ab31630fc244094e6f1692371cbaf87f666
Re-landing with a new serialization format, but maintaining ability to
load old serialized color spaces, for SKP compatibility.
Bug: skia:
Change-Id: Ib84a6e1cd5d7d9816175773fdbaff2ca32658667
Reviewed-on: https://skia-review.googlesource.com/c/181176
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This reverts commit d6fa45472c.
Reason for revert: Assertion failures
Original change's description:
> ccpr: Rework the path cache to support sporadic flushing
>
> Removes the notion of a stashed atlas that we store from the previous
> flush. Now we just cache every atlas we ever render. Cached atlases
> can either be 16-bit or 8-bit.
>
> The "reuse" and "animation" cases should both behave exactly the same
> as before: Where before we would copy from the stashed atlas to 8-bit
> atlases, we now copy from a cached 16-bit atlas and then invalidate
> it. Where before we would recycle the stashed atlas's backing texture
> object, we now recycle this same texture object from an invalidated
> 16-bit cached atlas.
>
> The main difference is that cases like tiled rendering now work. If
> you draw your whole scene in one flush, you still get one big 16-bit
> cached atlas, just like the "stashed atlas" implementation. But if you
> draw your scene in tiles, you now get lots of little cached 16-bit
> atlases, which can be reused and eventually copied to 8-bit atlases.
>
> Bug: skia:8462
> Change-Id: Ibae65febb948230aaaf1f1361eef9c8f06ebef18
> Reviewed-on: https://skia-review.googlesource.com/c/179991
> Commit-Queue: Chris Dalton <csmartdalton@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
TBR=bsalomon@google.com,robertphillips@google.com,csmartdalton@google.com
Change-Id: Iad74a14fcb09da12f32b9b78f803b8472a5d60ae
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:8462
Reviewed-on: https://skia-review.googlesource.com/c/181444
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Removes the notion of a stashed atlas that we store from the previous
flush. Now we just cache every atlas we ever render. Cached atlases
can either be 16-bit or 8-bit.
The "reuse" and "animation" cases should both behave exactly the same
as before: Where before we would copy from the stashed atlas to 8-bit
atlases, we now copy from a cached 16-bit atlas and then invalidate
it. Where before we would recycle the stashed atlas's backing texture
object, we now recycle this same texture object from an invalidated
16-bit cached atlas.
The main difference is that cases like tiled rendering now work. If
you draw your whole scene in one flush, you still get one big 16-bit
cached atlas, just like the "stashed atlas" implementation. But if you
draw your scene in tiles, you now get lots of little cached 16-bit
atlases, which can be reused and eventually copied to 8-bit atlases.
Bug: skia:8462
Change-Id: Ibae65febb948230aaaf1f1361eef9c8f06ebef18
Reviewed-on: https://skia-review.googlesource.com/c/179991
Commit-Queue: Chris Dalton <csmartdalton@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
When parsing numbers with long decimals, the fast-path decimal factor
may collapse to zero. Catch this condition (and defer to the slow parse
path).
Bug: skia:8499
Change-Id: I4e364402aecdcca9558d027a55ff297170e2a195
Reviewed-on: https://skia-review.googlesource.com/c/181179
Reviewed-by: Kevin Lubick <kjlubick@google.com>
Commit-Queue: Florin Malita <fmalita@chromium.org>
This is slightly preemptive, since paint still has getters/setters for
these fields, but skia no longer looks at them, so...
Reduces serial size by 20-bytes + size of a typeface
Bug: skia:2664
Change-Id: If1997c823db5b3d83b04cb7e1bc1d6787795b70a
Reviewed-on: https://skia-review.googlesource.com/c/181174
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Florin Malita <fmalita@chromium.org>
This reverts commit a9549ab316.
Reason for revert: SKPs changed?
Original change's description:
> Add SkColorSpace factory from 3x3 row-major gamut and transfer function
>
> Moved named common transfer functions and gamuts to constexpr values in
> SkColorSpace.h, in SkNamedTransferFn and SkNamedGamut namespaces.
>
> Converted nearly all SkColorSpace::MakeRGB calls within Skia to use the
> new factory with the named values. Multiple clients want a way to
> extract named transfer function and gamut - this still doesn't provide
> that, but this may be a better path forward for honestly advertising how
> SkColorSpace works internally.
>
> Bug: skia:
> Change-Id: I9296d67e8f0dab5ceb49869cb3ba24e98a05f3c4
> Reviewed-on: https://skia-review.googlesource.com/c/180360
> Reviewed-by: Mike Klein <mtklein@google.com>
> Reviewed-by: Brian Osman <brianosman@google.com>
> Commit-Queue: Brian Osman <brianosman@google.com>
TBR=mtklein@google.com,brianosman@google.com,reed@google.com
Change-Id: Ie888f877b3c1dba33e1a8c0f5fa92594628de7fb
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:
Reviewed-on: https://skia-review.googlesource.com/c/181300
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This reverts commit 5820b0c3f3.
Reason for revert: Unanticipated gold image differences and performance regressions
Original change's description:
> Use specialized quad lists in rectangle ops
>
> Hopefully reduces memory footprint of GrFillRectOp and GrTextureOp
>
> The original rect code (GrAAFillRectOp) stored 2 SkMatrices (18 floats), 2
> SkRects (8 floats) an SkPMColor4f (4 floats) and a flag (1 int) for a total
> of 124 bytes per quad that was stored in the op.
>
> The first pass at the rectangle consolidation switched to storing device and
> local quads as GrPerspQuads (32 floats), an SkPMColor4f (4 floats) and a flag
> (1 int) for a total of 148 bytes per quad. After landing, several memory
> regressions appeared in Chrome and our perf monitor.
>
> Several intertwined approaches are taken here. First, GrPerspQuad no longer
> caches 1/w, which makes a quad 12 floats instead of 16. Second, a specialized
> list type is defined that allows storing the x, y, and extra metadata together
> for quads, but keeps the w components separate. When the quad type isn't
> perspective, w is not stored at all since it is implicitly 1 and can be
> reconstituted at tessellation time. This brings the total per quad to either
> 84 or 116 bytes, depending on if the op list needs perspective information.
>
> Bug: chromium:915025
> Bug: chromium:917242
> Change-Id: If37ee122847b0c32604bb45dc2a1326b544f9cf6
> Reviewed-on: https://skia-review.googlesource.com/c/180644
> Commit-Queue: Michael Ludwig <michaelludwig@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
TBR=robertphillips@google.com,csmartdalton@google.com,michaelludwig@google.com
Change-Id: I6067b6c0e103d08787626a0a8eff753a0f0c97b6
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:915025, chromium:917242
Reviewed-on: https://skia-review.googlesource.com/c/181167
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Moved named common transfer functions and gamuts to constexpr values in
SkColorSpace.h, in SkNamedTransferFn and SkNamedGamut namespaces.
Converted nearly all SkColorSpace::MakeRGB calls within Skia to use the
new factory with the named values. Multiple clients want a way to
extract named transfer function and gamut - this still doesn't provide
that, but this may be a better path forward for honestly advertising how
SkColorSpace works internally.
Bug: skia:
Change-Id: I9296d67e8f0dab5ceb49869cb3ba24e98a05f3c4
Reviewed-on: https://skia-review.googlesource.com/c/180360
Reviewed-by: Mike Klein <mtklein@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Hopefully reduces memory footprint of GrFillRectOp and GrTextureOp
The original rect code (GrAAFillRectOp) stored 2 SkMatrices (18 floats), 2
SkRects (8 floats) an SkPMColor4f (4 floats) and a flag (1 int) for a total
of 124 bytes per quad that was stored in the op.
The first pass at the rectangle consolidation switched to storing device and
local quads as GrPerspQuads (32 floats), an SkPMColor4f (4 floats) and a flag
(1 int) for a total of 148 bytes per quad. After landing, several memory
regressions appeared in Chrome and our perf monitor.
Several intertwined approaches are taken here. First, GrPerspQuad no longer
caches 1/w, which makes a quad 12 floats instead of 16. Second, a specialized
list type is defined that allows storing the x, y, and extra metadata together
for quads, but keeps the w components separate. When the quad type isn't
perspective, w is not stored at all since it is implicitly 1 and can be
reconstituted at tessellation time. This brings the total per quad to either
84 or 116 bytes, depending on if the op list needs perspective information.
Bug: chromium:915025
Bug: chromium:917242
Change-Id: If37ee122847b0c32604bb45dc2a1326b544f9cf6
Reviewed-on: https://skia-review.googlesource.com/c/180644
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
- avoid using legacy paint->font converter
- call out that drawText exists just for testing
- simplify drawText to assume UTF8
Bug: skia:
Change-Id: Ide14d8581d4744827d2282e7983cc5e19070b21e
Reviewed-on: https://skia-review.googlesource.com/c/180641
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Before, we could only wait for all reserved objects to serialize.
The new unit test `SkPDF_abort_jobs` crashed before this change.
Change-Id: Ia2fcb33251c6c32c125f631ed3eeb619f616bef3
Reviewed-on: https://skia-review.googlesource.com/c/179856
Auto-Submit: Hal Canary <halcanary@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Skia can now build if we mark drawPosText as private,
Will hide/remove next (after Chrome CL)
Bug: skia:
Change-Id: I156560b025c119af302545bb5bd60678f7b8e8f7
Reviewed-on: https://skia-review.googlesource.com/c/179985
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
Auto-Submit: Mike Reed <reed@google.com>
Paths to intersect have two nearly coincident cubics. Where they
cross, the intersection error makes the curves start at slightly
different points. To sort the intersection, one curve is translated
to the start of the opposite point, moving it from one side to the
other, introducing a winding error.
The fix looks for that error in a very tiny range (enlarging that
range causes other tests that now pass to fail). This fix is very
fragile and points to the need for a better approach than sorting
angles to find winding values, as documented in the bug.
Also renamed some angle functions to show that they operate only
on lines and not general curves.
All tests pass with this fix:
./out/release/pathops_unittest -V -x
./out/debug/pathops_unittest -V -x
TBR=reed@google.com
Bug: skia:8380
Change-Id: I04e53d4c6a96035f661a4c9f31a17055ce13e3eb
Reviewed-on: https://skia-review.googlesource.com/c/179241
Commit-Queue: Cary Clark <caryclark@skia.org>
Reviewed-by: Cary Clark <caryclark@skia.org>
This reverts commit 576633cf57.
Fix: add new version to both variants of SkReadBuffer
Bug: skia:
Change-Id: I8ba60d374860718402328398a4a09b8c97d65e33
Reviewed-on: https://skia-review.googlesource.com/c/179845
Commit-Queue: Mike Reed <reed@google.com>
Auto-Submit: Mike Reed <reed@google.com>
Reviewed-by: Mike Reed <reed@google.com>
This reverts commit f1ae66bedd.
Reason for revert: need to guard for no-serialization builds (I think)
Original change's description:
> serialize fonts instead of paints
>
> Some changes from before:
> - we don't force serializing a typeface if the field is null
> - we don't serialize the blob bounds (don't trust them)
>
> Bug: skia:
> Change-Id: I41281b2aa63a1026de87330023346f1da5378c1f
> Reviewed-on: https://skia-review.googlesource.com/c/179735
> Commit-Queue: Mike Reed <reed@google.com>
> Reviewed-by: Florin Malita <fmalita@chromium.org>
> Reviewed-by: Herb Derby <herb@google.com>
TBR=bungeman@google.com,herb@google.com,fmalita@chromium.org,reed@google.com
Change-Id: Id891198bc683b9b5e7417a30c2f7adb0bd978e30
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:
Reviewed-on: https://skia-review.googlesource.com/c/179843
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
Some changes from before:
- we don't force serializing a typeface if the field is null
- we don't serialize the blob bounds (don't trust them)
Bug: skia:
Change-Id: I41281b2aa63a1026de87330023346f1da5378c1f
Reviewed-on: https://skia-review.googlesource.com/c/179735
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Florin Malita <fmalita@chromium.org>
Reviewed-by: Herb Derby <herb@google.com>
This reverts commit fdcbe206f5.
Reason for revert: Produced a further perf regression.
Original change's description:
> Tweak Plot size to try to get better perf.
>
> Large plots allow tighter packing of large glyphs, but smaller plots
> take less time to update. Trying a size in between for the largest
> A8 size.
>
> Bug: skia:8642
> Change-Id: Ib6d7651b70fe3a2101a3bfc5b367c433123cb914
> Reviewed-on: https://skia-review.googlesource.com/c/179242
> Commit-Queue: Jim Van Verth <jvanverth@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
TBR=jvanverth@google.com,robertphillips@google.com
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: skia:8642
Change-Id: I9bc32abfae1fb4a6380b2fd5c0a1ea269447c534
Reviewed-on: https://skia-review.googlesource.com/c/179726
Reviewed-by: Jim Van Verth <jvanverth@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
* sed 's/sk_sp<SkPDF\(Object\|Array\|Dict\)>/std::unique_ptr<SkPDF\1>/g'
* Added SkPDFMakeDict()
* Let SkPDFMakeArray() take zero arguments.
Bug: skia:8630
Change-Id: I08dc11e974f1d71e09a33af6bdbcb5f01b07d186
Reviewed-on: https://skia-review.googlesource.com/c/179063
Commit-Queue: Hal Canary <halcanary@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Auto-Submit: Hal Canary <halcanary@google.com>
A proc can be registered with a GrTexture. The proc will be called when
it is safe to delete the texture is "idle." Idle means it referred to
outside of GrResourceCache and that the I/O operations on the GPU are
completed (this latter part applieas to Vulkan only).
The intended use case for this is to call promise image texture release
procs once we start caching GrTextures for deinstantiated promise
images.
Bug= skia:8613
Change-Id: Idce9a4292fef7b15370a053060d8878a9d6828fa
Reviewed-on: https://skia-review.googlesource.com/c/178937
Reviewed-by: Greg Daniel <egdaniel@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Large plots allow tighter packing of large glyphs, but smaller plots
take less time to update. Trying a size in between for the largest
A8 size.
Bug: skia:8642
Change-Id: Ib6d7651b70fe3a2101a3bfc5b367c433123cb914
Reviewed-on: https://skia-review.googlesource.com/c/179242
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
It's probably nicer to draw these images than not draw them.
Cq-Include-Trybots: luci.chromium.try:linux-blink-rel
Change-Id: I9f8df081c8256d345ad2d8bbbb30dbad982ad94e
Reviewed-on: https://skia-review.googlesource.com/c/178981
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
All PDF Streams are immediatly serialized, and are never long-lived
in memory.
if EXPERIMENTAL fExecutor is set on the Document, streams are
compressed in parallel.
Results for PDFBigDocBench:
without patch 1807885.01 μs
with patch without executor 1802808.35 μs
with patch with executor 246313.72 μs
SkPDFStreamOut() function replaces SkPDFStream classes.
Page resources are all serialized early.
Several Document-level objects are serialzied early.
SkUUID introduced as top-level object.
Many {insert|append}ObjRef() converted to memory efficient
{insert|append}Ref().
Bug: skia:8630
Change-Id: Ic336917d0c8b9ac1c2423b43bfe9b49a3533fbff
Reviewed-on: https://skia-review.googlesource.com/c/176588
Auto-Submit: Hal Canary <halcanary@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>
I've removed a bunch of unused code,
inlined routines that are only used once,
removed platform code where a portable code works fine,
and renamed methods and updated comments.
I'm thinking about getting rid of Sk4px entirely, and boiling it down
instead to a couple helper methods on Sk16b/Sk16h. This is a start.
Change-Id: I35b6c44710aa2cefe8c1e07fc2cb877a4042e8a5
Reviewed-on: https://skia-review.googlesource.com/c/178985
Auto-Submit: Mike Klein <mtklein@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Reviewed-by: Herb Derby <herb@google.com>
This flag is annoying to set correctly when writing test code.
Fortunately it has become redundant with the "is GL FBO 0" flag.
Change-Id: Ifd88292d2d6ea05bfe0d269e853baff857e70bfe
Reviewed-on: https://skia-review.googlesource.com/c/178929
Reviewed-by: Greg Daniel <egdaniel@google.com>
Reviewed-by: Chris Dalton <csmartdalton@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
This test used to assert/fail under ASAN because copyPath's
fLastMoveToIndex was garbage after the internal copy (in SkPath::transform).
Now the code sets it to the src's value.
Bug:883596
Change-Id: I3cba82fb63398bf4aa0abc9d3f2a0067d8ad9006
Reviewed-on: https://skia-review.googlesource.com/c/178931
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Mike Reed <reed@google.com>
The current system works if all the DDLs (generated from a single DDL Recorder) are all replayed into the same surface but would probably fail otherwise.
Change-Id: I07fb0ca9ecfb7a6aa27766bdb62e5b41311021d4
Reviewed-on: https://skia-review.googlesource.com/c/178900
Commit-Queue: Brian Salomon <bsalomon@google.com>
Auto-Submit: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
This reverts commit 99e1709ef5.
Reason for revert: DDL failures
Original change's description:
> Fix recording multiple DDLs from a single DDL Recorder
>
> The current system works if all the DDLs (generated from a single DDL Recorder) are all replayed into the same surface but would probably fail otherwise.
>
> Change-Id: I0e76ecbf36aa3975049e4f48aaadd35b113a22eb
> Reviewed-on: https://skia-review.googlesource.com/c/178260
> Commit-Queue: Robert Phillips <robertphillips@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
TBR=bsalomon@google.com,robertphillips@google.com
Change-Id: I958c0419647eee4be621bd09c2421ff3801ffe71
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/178860
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
The current system works if all the DDLs (generated from a single DDL Recorder) are all replayed into the same surface but would probably fail otherwise.
Change-Id: I0e76ecbf36aa3975049e4f48aaadd35b113a22eb
Reviewed-on: https://skia-review.googlesource.com/c/178260
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Change-Id: I784e85cb154be077a1ca1949d7806ac95269d8c8
Reviewed-on: https://skia-review.googlesource.com/c/178281
Reviewed-by: Mike Klein <mtklein@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Moves method to get GrBackendFormat from GrBackendTexture from GrCaps
to GrBackendTexture so that a GrContext is not required.
Uses kUnknown_GrPixelConfig as failure return from GrCaps functions
rather than an GrPixelConfig* out param and bool result.
Having the texture type be part of GrBackendFormat made removing the
GrCaps function that goes from GrBackendRenderTarget to GrPixelConfig
awkward so that was left alone for now.
Change-Id: If9be0f898c538be4a7b24022b6011f63441a0317
Reviewed-on: https://skia-review.googlesource.com/c/175991
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Uses a fixed quantized table to compute the ARGB texture dimensions,
and compute the A8 dimensions from that. Plot sizes are computed
based on the respective atlas's dimensions.
Bug: skia:
Change-Id: I28c56557094efa49d1a011a1d6145ef3a79b7cac
Reviewed-on: https://skia-review.googlesource.com/c/177068
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Herb Derby <herb@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Remove SkPathPriv::IsConvex() if SK_LEGACY_PATH_CONVEXITY
is defined to fix the Google3 roll.
TBR=ethannicholas@google.com
Bug: skia:
Change-Id: I2d2177213ec43dd048f15685d0afe49fb07656fe
Reviewed-on: https://skia-review.googlesource.com/c/177680
Commit-Queue: Cary Clark <caryclark@google.com>
Auto-Submit: Cary Clark <caryclark@google.com>
Reviewed-by: Cary Clark <caryclark@google.com>
With skia_use_wuffs=true, calling getFrameCount after the partial
decode results in the full decode failing.
Bug: skia:8235
Change-Id: I9ac5496e3fecf309f1e303a8d3e8725a6d00731b
Reviewed-on: https://skia-review.googlesource.com/c/176590
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Nigel Tao <nigeltao@google.com>
Chrome added a public method to validate SkPathRef,
but always called it when validating SkPath. We did too.
Remove the SkPathRef entry point, validate SkPathRef
when validating SkPath, and remove Skia's callers.
(Chrome has already been fixed.)
TBR=reed@google.comR=fmalita@chromium.org
Bug:913930
Change-Id: I0828b00b42cc1f031b4216ddeace50f80aa21e62
Reviewed-on: https://skia-review.googlesource.com/c/177065
Commit-Queue: Cary Clark <caryclark@skia.org>
Auto-Submit: Cary Clark <caryclark@skia.org>
Reviewed-by: Florin Malita <fmalita@chromium.org>
This separates the existing convexity logic into
two passes. The first pass detects concavity by
counting the changes in direction.
The second pass computes the cross product to
see that all angles bend in the same direction, and
computes the dot product to see if the angle
doubles back on itself.
The second pass treats axis-aligned vectors
separately, and computes the dot and cross products
by comparing point values; it does not use arithmetic
to determine convexity, so it works with all finite
values.
A compile time switch enables returning concave
for co-linear diagonal points:
If successive points are not axis-aligned, and
those points are co-linear along a diagonal;
the path is treated as concave. This is conservative
but avoids paths that change convexity when the
are translated or scaled, since transforming the
path may cause the midpoint to shift to either
side of a line formed by the endpoints.
The compile time switch is set so that co-linear
diagonal points do not affect convexity. Note that
this permits shapes formerly considered concave, such
as stroked lines with round caps, to become convex;
this accounts for many of the GM differences.
A path may double back on itself and be convex;
for instance, a path containing a single line.
Path may have multiple initial moveTo verbs, or
trailing moveTo verbs, and still evaluate as convex.
A separate entry point, SkPathPriv::IsConvex()
allows passing an array of points instead of a path.
A legacy define has been checked into Chrome to
use the old code until layout tests have been
rebaselined.
R=reed@google.com,bsalomon@google.com
Bug:899689
Change-Id: I392bbe04836ffb19666ad92ab2a2404c56543019
Reviewed-on: https://skia-review.googlesource.com/c/173427
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Cary Clark <caryclark@google.com>
Commit-Queue: Cary Clark <caryclark@skia.org>
If getPixels returns an unexpected result, do not create a digest
and compare it to digests for premul and unpremul.
Print a better error message, including the name of the image and
string versions of the SkCodec::Result.
Change-Id: Ie98e350a1ef91c0745fe9de0855192e5793e033c
Reviewed-on: https://skia-review.googlesource.com/c/176582
Auto-Submit: Leon Scroggins <scroggo@google.com>
Reviewed-by: Nigel Tao <nigeltao@google.com>
Commit-Queue: Leon Scroggins <scroggo@google.com>
- add drop-in clang-tidy cxx wrapper
- get build clean for bugprone-use-after-move
The wrapper can be used by setting
cxx = "/path/to/skia/tools/clang-tidy.sh"
in GN.
Change-Id: Idbba911e23bd6ef7530b08fd31906b92c1c1b28c
Reviewed-on: https://skia-review.googlesource.com/c/176523
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
This reverts commit e7f165be2e.
Reason for revert: doesn't compile in google3
Original change's description:
> Treat kWEBP encode with quality=100 as lossless
>
> In SkEncodeImage and friends, treat quality of 100 as a lossless encode
> when using kWEBP. This seems a good fit for the intent - which is
> presumably to save the highest quality image. This also matches
> Chromium's blink::ImageEncoder::ComputeWebpOptions, which treats a
> quality of 1 (on a float scale from 0 to 1) as a lossless encode.
>
> FWIW, Chromium has had this behavior since
> https://codereview.chromium.org/1937433002, in response to
> crbug.com/523098. The goal is to "maintain sharpness to
> match the JPEG encoder behavior (use WEBP lossless encoding)".
>
> Add a test to verify the new behavior. This requires making tests
> depend on libwebp to use WebPGetFeatures, since the Skia API does not
> provide a way to determine whether an encoded webp file was encoded
> lossless-ly or lossily.
>
> Bug: skia:8586
> Change-Id: Ie9e09c2f7414ab701d696c4ad9edf405868a716f
> Reviewed-on: https://skia-review.googlesource.com/c/175823
> Commit-Queue: Leon Scroggins <scroggo@google.com>
> Reviewed-by: Derek Sollenberger <djsollen@google.com>
> Reviewed-by: Mike Reed <reed@google.com>
TBR=djsollen@google.com,scroggo@google.com,reed@google.com
Change-Id: I91680f65a2a5e6f0a13b84e97c9541ebe0606b33
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:8586
Reviewed-on: https://skia-review.googlesource.com/c/176584
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
Bug: skia:
Change-Id: Iee57bc970a026de2ad5a0758153e9cbb20753fa1
Reviewed-on: https://skia-review.googlesource.com/c/173105
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
In SkEncodeImage and friends, treat quality of 100 as a lossless encode
when using kWEBP. This seems a good fit for the intent - which is
presumably to save the highest quality image. This also matches
Chromium's blink::ImageEncoder::ComputeWebpOptions, which treats a
quality of 1 (on a float scale from 0 to 1) as a lossless encode.
FWIW, Chromium has had this behavior since
https://codereview.chromium.org/1937433002, in response to
crbug.com/523098. The goal is to "maintain sharpness to
match the JPEG encoder behavior (use WEBP lossless encoding)".
Add a test to verify the new behavior. This requires making tests
depend on libwebp to use WebPGetFeatures, since the Skia API does not
provide a way to determine whether an encoded webp file was encoded
lossless-ly or lossily.
Bug: skia:8586
Change-Id: Ie9e09c2f7414ab701d696c4ad9edf405868a716f
Reviewed-on: https://skia-review.googlesource.com/c/175823
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Reviewed-by: Mike Reed <reed@google.com>
This reverts commit a34e6eac3c.
Reason for revert: Broken gold images on Adreno 530 and 540
Original change's description:
> Fix performance regresion on Mali, while preserving sk_FragCoord.w
>
> Bug: skia:
> Change-Id: I9269288bbb861cd02efa7b5dfe2c9434ec39064e
> Reviewed-on: https://skia-review.googlesource.com/c/174309
> Commit-Queue: Michael Ludwig <michaelludwig@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
TBR=bsalomon@google.com,michaelludwig@google.com
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: skia:
Change-Id: Id344219c101564a5fd06477895386fad2d8dfa78
Reviewed-on: https://skia-review.googlesource.com/c/175986
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Michael Ludwig <michaelludwig@google.com>
This fixes a crash we saw when we switch vulkan copies as draws to creating
their own secondary command buffer. The crash came from the perf blendmode
tests when using an advanced blend mode. They would do 1000 draws which forced
us into creating 2000 command buffers (since the dst copies and the normal draws
each used them). I tested without the copies as draws change and just increasing
the total number of draws we do and was able to repro the crash.
Besides fixing the above OOM crash, I am also seeing a 5-10% perf gain on the
blendmode micro benches which is nice
Bug: skia:
Change-Id: I9266ea0ba02a755f54dabd4ee804963ab0c9b684
Reviewed-on: https://skia-review.googlesource.com/c/175436
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Greg Daniel <egdaniel@google.com>
Uninstantiated: a state of not being instantiated.
Deinstantiate: transition from instantiated to uninstantiated state.
Change-Id: Id7b6b295267674a9915f95105d73fabfcc3555de
Reviewed-on: https://skia-review.googlesource.com/c/175586
Reviewed-by: Greg Daniel <egdaniel@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Bug: skia:
Change-Id: I365c82a058c97e8741ee91e999cb6faab5a5ecf5
Reviewed-on: https://skia-review.googlesource.com/c/175422
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Follow-ups assuming this sticks:
- remove client overrides, so we can remove the (empty) base class virtuals
- mark as obsolete the .skp enums for these entrypoints, and note date
- remove glyphrunlist helpers that now may not have any callers
- remove DrawCommands related to these entrypoints
Bug: skia:
Change-Id: I44b4e6ef3bee61911e77b9b2ab73f82c0817536a
Reviewed-on: https://skia-review.googlesource.com/c/175240
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Florin Malita <fmalita@chromium.org>
Reviewed-by: Herb Derby <herb@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>
Bug: skia:
Change-Id: I9269288bbb861cd02efa7b5dfe2c9434ec39064e
Reviewed-on: https://skia-review.googlesource.com/c/174309
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
This reverts commit 10273c1d5f.
Reason for revert: bad gms
Original change's description:
> Use unnormalized coords all the way through with GL_TEXTURE_RECTANGLE textures.
>
> We used to unnormalize them in the shader via SkSL.
>
> This allows us to support GL_TEXTURE_RECTANGLE without having textureSize()
> available in GLSL.
>
> Change-Id: Ibe63a302228811933ef000251db4cad9aaf4f2ea
> Reviewed-on: https://skia-review.googlesource.com/c/174068
> Reviewed-by: Brian Osman <brianosman@google.com>
> Commit-Queue: Brian Salomon <bsalomon@google.com>
TBR=bsalomon@google.com,brianosman@google.com
Change-Id: I9bf38e1040578becba28ac8cccd81e2af2844278
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/c/175252
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Bug: skia:7903
Change-Id: I69b65fc1cfcc2cc5045bb3b75395f9a256ade278
Reviewed-on: https://skia-review.googlesource.com/c/172979
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Read only textures and proxies fail writePixels, as copy dsts, and mip
regeneration.
Bug: skia:8509
Change-Id: Iaa0b473cc9a9930fde3ef0e91373d5040650de35
Reviewed-on: https://skia-review.googlesource.com/c/174316
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
We used to unnormalize them in the shader via SkSL.
This allows us to support GL_TEXTURE_RECTANGLE without having textureSize()
available in GLSL.
Change-Id: Ibe63a302228811933ef000251db4cad9aaf4f2ea
Reviewed-on: https://skia-review.googlesource.com/c/174068
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
Bug: skia:8579
Zero-initialize the pixel buffer that wuffs uses inside the frame rect.
If the encoded image is incomplete, we will swizzle this memory into the
destination, so this prevents swizzling uninitialized memory.
In addition, zero-initialize the client's memory if the above buffer
will not fill it, and if the frame is independent.
Move decodeFrameConfig into onStartIncrementalDecode. This makes it
clear that we haven't begun decoding the pixels so rowsDecoded is
irrelevant. Update the documentation for onStartIncrementalDecode to
make it clear that it can be called more than once (and when to do so).
If a frame is incomplete and depends on prior frames, do not blend it.
This will overwrite pixels from the prior frames without a way to undo.
Current clients only want to show incomplete images for the first frame
anyway.
Add some debugging information to Codec_partialAnim test, including
writing out pngs when failing the test.
TBR=djsollen@google.com for documentation change in SkCodec.h
Change-Id: I85c8ca4075301306f4738ddfc2f5992a5745108b
Reviewed-on: https://skia-review.googlesource.com/c/174310
Commit-Queue: Leon Scroggins <scroggo@google.com>
Reviewed-by: Nigel Tao <nigeltao@google.com>
Change-Id: I4960b1cd055daf44637e95825f82cb7fe2ce134a
Reviewed-on: https://skia-review.googlesource.com/c/174285
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Bug: skia:
Change-Id: I11ff57dca2c48519bc3a23e36da01bf40d1b329c
Reviewed-on: https://skia-review.googlesource.com/c/174221
Commit-Queue: Mike Reed <reed@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Auto-Submit: Mike Reed <reed@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Bug: skia:8410 chromium:876349
Change-Id: Ib97c82cd165f7e2f2e94c65fc307220b99053df3
Reviewed-on: https://skia-review.googlesource.com/c/174065
Commit-Queue: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
Reviewed-by: Mike Klein <mtklein@google.com>
Bug: skia:
Change-Id: I04b6cbfe083c32a25205e04a4b15672d4b678a9c
Reviewed-on: https://skia-review.googlesource.com/c/174063
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Michael Ludwig <michaelludwig@google.com>
In theory, a convex shape is still convex if transformed by an affine
matrix. However, SkPath segments are specified using floats, and attributes
like collinearity can break under some transforms due to finite precision.
Computing convexity is non-trivial, so there is value in SkPath caching this
calculation. Convexity is useful, as both the CPU and GPU backends can draw
convex shapes faster than non-convex.
To balance these two (fragile float math and value of caching convexity),
this CL invalidates this cached state if the transform could change convexity.
In the general case, it is assumed that convexity could change. Special cases
where it is safe to keep the cached state after transform are:
- identity transform
- scale/translate transform if the path is known to be axis-aligned
All other combinations invalidate the cached state, forcing it to be
recomputed.
"axis-aligned" means the segments in the path are all axis-aligned, horizontal
or vertical (e.g. a rect or rrect)
Bug: 899689
Change-Id: I1381273eaff61d6b7134ae94b4f251c69991081a
Reviewed-on: https://skia-review.googlesource.com/c/173226
Commit-Queue: Ravi Mistry <rmistry@google.com>
Reviewed-by: Cary Clark <caryclark@google.com>
This reverts commit 26d8d77aae.
Reason for revert: speculative, in case this is blocking the chrome roll
Original change's description:
> don't trust convexity with affine transforms
>
> In theory, a convex shape transformed by an affine matrix should still
> be convex. However, due to numerical nastiness of floats, when we try
> to determine if something is convex, we can get different answers pre
> and post a transformation (think of two line segments nearly colinear).
>
> Convex paths take a faster scan converter, but it is only well behaved
> if the path is, in fact, convex. Thus we have to be conservative about
> which paths we mark as convex.
>
> This bug found a case where a "convex" path, after going through a transform,
> became (according to our measure) non-convex. The bug was that we *thought*
> that once convex always convex, but in reality it was not. The fix (hack) is
> to notice when we transform by an affine matrix (we're still assuming/hoping
> that scaling and translate keep things convex (1)...) and mark the convexity
> as "unknown", forcing us to re-compute it.
>
> This will slow down these paths, since it costs something to compute convexity.
> Hopefully non-scale-translate transforms are rare, so we won't notice the
> speed loss too much.
>
> (1) This is not proven. If we find scaling/translation to break our notion of
> convexity, we'll need to get more aggressive/clever to find a fix.
>
>
> Bug: 899689
> Change-Id: I5921eca247428bf89380bc2395fe373fa70deb1d
> Reviewed-on: https://skia-review.googlesource.com/c/173080
> Commit-Queue: Mike Reed <reed@google.com>
> Reviewed-by: Cary Clark <caryclark@google.com>
> Reviewed-by: Jim Van Verth <jvanverth@google.com>
TBR=mtklein@google.com,jvanverth@google.com,caryclark@google.com,reed@google.com,caryclark@skia.org
Change-Id: I5d846798f2c34c6576591a3c3125cfdc3c72dbdc
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 899689
Reviewed-on: https://skia-review.googlesource.com/c/173162
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
In theory, a convex shape transformed by an affine matrix should still
be convex. However, due to numerical nastiness of floats, when we try
to determine if something is convex, we can get different answers pre
and post a transformation (think of two line segments nearly colinear).
Convex paths take a faster scan converter, but it is only well behaved
if the path is, in fact, convex. Thus we have to be conservative about
which paths we mark as convex.
This bug found a case where a "convex" path, after going through a transform,
became (according to our measure) non-convex. The bug was that we *thought*
that once convex always convex, but in reality it was not. The fix (hack) is
to notice when we transform by an affine matrix (we're still assuming/hoping
that scaling and translate keep things convex (1)...) and mark the convexity
as "unknown", forcing us to re-compute it.
This will slow down these paths, since it costs something to compute convexity.
Hopefully non-scale-translate transforms are rare, so we won't notice the
speed loss too much.
(1) This is not proven. If we find scaling/translation to break our notion of
convexity, we'll need to get more aggressive/clever to find a fix.
Bug: 899689
Change-Id: I5921eca247428bf89380bc2395fe373fa70deb1d
Reviewed-on: https://skia-review.googlesource.com/c/173080
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Cary Clark <caryclark@google.com>
Reviewed-by: Jim Van Verth <jvanverth@google.com>
