This is a reland of 78cb579f33
This time, lowp stages are controlled by !defined(JUMPER_IS_SCALAR), not
by defined(__clang__). The two are usually the same, except when we opt
Clang builds into JUMPER_IS_SCALAR artificially.
Some Google3 builds use compilers old enough that they barf when
compiling our NEON code. It's conceivably also possible to define
JUMPER_IS_SCALAR yourself, but I don't think anyone does that.
Original change's description:
> Reland "make SkJumper stages normal Skia code"
>
> This is a reland of 22e536e3a1
>
> Now with fixed #include paths in SkRasterPipeline_opts.h,
> and -ffp-contract=fast for the :hsw target to minimize
> diffs on non-Windows Clang AVX2/AVX-512 bots.
>
> Original change's description:
> > make SkJumper stages normal Skia code
> >
> > Enough clients are using Clang now that we can say, use Clang to build
> > if you want these software pipeline stages to go fast.
> >
> > This lets us drop the offline build aspect of SkJumper stages, instead
> > building as part of Skia using the SkOpts framework.
> >
> > I think everything should work, except I've (temporarily) removed
> > AVX-512 support. I will put this back in a follow up.
> >
> > I have had to drop Windows down to __vectorcall and our narrower
> > stage calling convention that keeps the d-registers on the stack.
> > I tried forcing sysv_abi, but that crashed Clang. :/
> >
> > Added a TODO to up the same narrower stage calling convention
> > for lowp stages... we just *don't* today, for no good reason.
> >
> > Change-Id: Iaaa792ffe4deab3508d2dc5d0008c163c24b3383
> > Reviewed-on: https://skia-review.googlesource.com/110641
> > Commit-Queue: Mike Klein <mtklein@chromium.org>
> > Reviewed-by: Herb Derby <herb@google.com>
> > Reviewed-by: Florin Malita <fmalita@chromium.org>
>
> Change-Id: I44f2c03d33958e3807747e40904b6351957dd448
> Reviewed-on: https://skia-review.googlesource.com/112742
> Reviewed-by: Mike Klein <mtklein@chromium.org>
Change-Id: I3d71197d4bbb19ca4a94961a97fa2e54d5cbfb0d
Reviewed-on: https://skia-review.googlesource.com/112744
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
This reverts commit 78cb579f33.
Reason for revert: lowp should be controlled by defined(JUMPER_IS_SCALAR), not defined(__clang__). So close.
Original change's description:
> Reland "make SkJumper stages normal Skia code"
>
> This is a reland of 22e536e3a1
>
> Now with fixed #include paths in SkRasterPipeline_opts.h,
> and -ffp-contract=fast for the :hsw target to minimize
> diffs on non-Windows Clang AVX2/AVX-512 bots.
>
> Original change's description:
> > make SkJumper stages normal Skia code
> >
> > Enough clients are using Clang now that we can say, use Clang to build
> > if you want these software pipeline stages to go fast.
> >
> > This lets us drop the offline build aspect of SkJumper stages, instead
> > building as part of Skia using the SkOpts framework.
> >
> > I think everything should work, except I've (temporarily) removed
> > AVX-512 support. I will put this back in a follow up.
> >
> > I have had to drop Windows down to __vectorcall and our narrower
> > stage calling convention that keeps the d-registers on the stack.
> > I tried forcing sysv_abi, but that crashed Clang. :/
> >
> > Added a TODO to up the same narrower stage calling convention
> > for lowp stages... we just *don't* today, for no good reason.
> >
> > Change-Id: Iaaa792ffe4deab3508d2dc5d0008c163c24b3383
> > Reviewed-on: https://skia-review.googlesource.com/110641
> > Commit-Queue: Mike Klein <mtklein@chromium.org>
> > Reviewed-by: Herb Derby <herb@google.com>
> > Reviewed-by: Florin Malita <fmalita@chromium.org>
>
> Change-Id: I44f2c03d33958e3807747e40904b6351957dd448
> Reviewed-on: https://skia-review.googlesource.com/112742
> Reviewed-by: Mike Klein <mtklein@chromium.org>
TBR=mtklein@chromium.org,herb@google.com,fmalita@chromium.org
Change-Id: Ie64da98f5187d44e03c0ce05d7cb189d4a6e6663
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/112743
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
This is a reland of 22e536e3a1
Now with fixed #include paths in SkRasterPipeline_opts.h,
and -ffp-contract=fast for the :hsw target to minimize
diffs on non-Windows Clang AVX2/AVX-512 bots.
Original change's description:
> make SkJumper stages normal Skia code
>
> Enough clients are using Clang now that we can say, use Clang to build
> if you want these software pipeline stages to go fast.
>
> This lets us drop the offline build aspect of SkJumper stages, instead
> building as part of Skia using the SkOpts framework.
>
> I think everything should work, except I've (temporarily) removed
> AVX-512 support. I will put this back in a follow up.
>
> I have had to drop Windows down to __vectorcall and our narrower
> stage calling convention that keeps the d-registers on the stack.
> I tried forcing sysv_abi, but that crashed Clang. :/
>
> Added a TODO to up the same narrower stage calling convention
> for lowp stages... we just *don't* today, for no good reason.
>
> Change-Id: Iaaa792ffe4deab3508d2dc5d0008c163c24b3383
> Reviewed-on: https://skia-review.googlesource.com/110641
> Commit-Queue: Mike Klein <mtklein@chromium.org>
> Reviewed-by: Herb Derby <herb@google.com>
> Reviewed-by: Florin Malita <fmalita@chromium.org>
Change-Id: I44f2c03d33958e3807747e40904b6351957dd448
Reviewed-on: https://skia-review.googlesource.com/112742
Reviewed-by: Mike Klein <mtklein@chromium.org>
This reverts commit 22e536e3a1.
Reason for revert: wrong include path :/
Original change's description:
> make SkJumper stages normal Skia code
>
> Enough clients are using Clang now that we can say, use Clang to build
> if you want these software pipeline stages to go fast.
>
> This lets us drop the offline build aspect of SkJumper stages, instead
> building as part of Skia using the SkOpts framework.
>
> I think everything should work, except I've (temporarily) removed
> AVX-512 support. I will put this back in a follow up.
>
> I have had to drop Windows down to __vectorcall and our narrower
> stage calling convention that keeps the d-registers on the stack.
> I tried forcing sysv_abi, but that crashed Clang. :/
>
> Added a TODO to up the same narrower stage calling convention
> for lowp stages... we just *don't* today, for no good reason.
>
> Change-Id: Iaaa792ffe4deab3508d2dc5d0008c163c24b3383
> Reviewed-on: https://skia-review.googlesource.com/110641
> Commit-Queue: Mike Klein <mtklein@chromium.org>
> Reviewed-by: Herb Derby <herb@google.com>
> Reviewed-by: Florin Malita <fmalita@chromium.org>
TBR=mtklein@chromium.org,herb@google.com,fmalita@chromium.org
Change-Id: I2bdc709c80cdfa6b13ff24e024b3721bef887f46
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://skia-review.googlesource.com/112741
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>
Enough clients are using Clang now that we can say, use Clang to build
if you want these software pipeline stages to go fast.
This lets us drop the offline build aspect of SkJumper stages, instead
building as part of Skia using the SkOpts framework.
I think everything should work, except I've (temporarily) removed
AVX-512 support. I will put this back in a follow up.
I have had to drop Windows down to __vectorcall and our narrower
stage calling convention that keeps the d-registers on the stack.
I tried forcing sysv_abi, but that crashed Clang. :/
Added a TODO to up the same narrower stage calling convention
for lowp stages... we just *don't* today, for no good reason.
Change-Id: Iaaa792ffe4deab3508d2dc5d0008c163c24b3383
Reviewed-on: https://skia-review.googlesource.com/110641
Commit-Queue: Mike Klein <mtklein@chromium.org>
Reviewed-by: Herb Derby <herb@google.com>
Reviewed-by: Florin Malita <fmalita@chromium.org>
This adds the ability to define long phrases
in one place and refer to those phrases in
many places.
Bookmaker has new syntax to support phrase substitution.
When it encounters
#some_phrase_reference#
It substitutes the body of
#PhraseDef some_phrase_reference
text to substitute when encountering the phrase
##
The phrase label must start with a lowercase letter,
and be bracketed by single hash marks, without spaces
between the label and the hash marks.
Docs-Preview: https://skia.org/?cl=111224TBR=caryclark@google.com
Bug: skia:6898
Change-Id: I12c57d916ccedbd86b421377d117399150ada72a
Reviewed-on: https://skia-review.googlesource.com/111224
Reviewed-by: Cary Clark <caryclark@skia.org>
Commit-Queue: Cary Clark <caryclark@skia.org>
PLEASE NOTE: Instructions on running `am instrument` for the
SkQP APK have changed.
To run a single test, see the section "Running a single test"
in `tools/skqp/README.md`.
No-Try: true
Change-Id: I0a2cbc47755929d6c6a927a3591ff98046779c77
Reviewed-on: https://skia-review.googlesource.com/108780
Reviewed-by: Derek Sollenberger <djsollen@google.com>
Commit-Queue: Hal Canary <halcanary@google.com>
move the #Line directive to make bookmaker happy
Docs-Preview: https://skia.org/?cl=108740
Bug: skia:6898
Change-Id: I8b993bab6b4f69cc266941e281a00a945f32f25e
Reviewed-on: https://skia-review.googlesource.com/108740
Commit-Queue: Cary Clark <caryclark@skia.org>
Reviewed-by: Cary Clark <caryclark@skia.org>
Florin pointed out docs that used SkPoint where
SkIPoint was intended. This was caused by the
original documentation guessing that 'points'
referred to one and not the other.
For now, use IPoint explicitly in the documentation
to get around this.
Docs-Preview: https://skia.org/?cl=108569
Bug: skia:6898
Change-Id: I12bac1d4cf82a614bee98895f12e4f99add2dbd7
Reviewed-on: https://skia-review.googlesource.com/108569
Reviewed-by: Florin Malita <fmalita@chromium.org>
Commit-Queue: Cary Clark <caryclark@skia.org>
generating replacement includes exposed
errors mostly dealing with globals like
SkAlphaType and members. Rewrite finding
and resolving links to hopefully make this
area more robust.
TBR=caryclark@google.com
Docs-Preview: https://skia.org/?cl=107160
Bug: skia:6898
Change-Id: I9b8025160203d204286f3f6ca0cebd70da6253b4
Reviewed-on: https://skia-review.googlesource.com/107160
Commit-Queue: Cary Clark <caryclark@skia.org>
Reviewed-by: Cary Clark <caryclark@skia.org>
Must use double quotes; no comma after last element of array.
Change-Id: I2556508babacc9bc00a5cddceed9ee454dc4fc96
Reviewed-on: https://skia-review.googlesource.com/106282
Reviewed-by: Ben Wagner <bungeman@google.com>
Reviewed-by: Cary Clark <caryclark@google.com>
Commit-Queue: Ben Wagner <bungeman@google.com>
This reverts the revert 9ff8c8c073.
Original:
This is a performance-only hint that no one but fuzzers
is using. It's even explicitly filtered out in Android.
The fuzzers have noticed they can trick us into allocating
uninitialized memory and treating it as opaque, blending
uninitialized pixels, etc.
Since no one's using this, we can just kill the bit.
Bug: skia:7566, chromium:808830
Docs-Preview: https://skia.org/?cl=105282
Change-Id: I4326c663f777aa373ff7ec9f319519da9729350d
Reviewed-on: https://skia-review.googlesource.com/105282
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Mike Klein <mtklein@chromium.org>
add self-check looking for #In markup on every method, pointing
to an existing #Subtopic to reference the method.
Docs-Preview: https://skia.org/?cl=104325
Bug: skia:6898
Change-Id: I749a25b9a43033ae68d193249b2c0b810dcf8fc8
Reviewed-on: https://skia-review.googlesource.com/104325
Commit-Queue: Cary Clark <caryclark@skia.org>
Reviewed-by: Cary Clark <caryclark@skia.org>
This reverts commit 9a7a2ee5ad.
Reason for revert: still need to update blink_headless in Google3
Original change's description:
> remove SkCanvas::kIsOpaque_SaveLayerFlag
>
> This is a performance-only hint that no one but fuzzers
> is using. It's even explicitly filtered out in Android.
>
> The fuzzers have noticed they can trick us into allocating
> uninitialized memory and treating it as opaque, blending
> uninitialized pixels, etc.
>
> Since no one's using this, we can just kill the bit.
>
> Bug: skia:7566, chromium:808830
>
> Change-Id: Id74a85e51bc4d0907b4127eb9ac5b02576f8f0a7
> Reviewed-on: https://skia-review.googlesource.com/104441
> Reviewed-by: Mike Reed <reed@google.com>
> Commit-Queue: Mike Klein <mtklein@chromium.org>
TBR=mtklein@chromium.org,fmalita@chromium.org,reed@google.com
Change-Id: I10d3c7e5184b9322715a5bfb6a7106292c8876a3
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: skia:7566, chromium:808830
Reviewed-on: https://skia-review.googlesource.com/104781
Reviewed-by: Mike Klein <mtklein@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
This is a performance-only hint that no one but fuzzers
is using. It's even explicitly filtered out in Android.
The fuzzers have noticed they can trick us into allocating
uninitialized memory and treating it as opaque, blending
uninitialized pixels, etc.
Since no one's using this, we can just kill the bit.
Bug: skia:7566, chromium:808830
Change-Id: Id74a85e51bc4d0907b4127eb9ac5b02576f8f0a7
Reviewed-on: https://skia-review.googlesource.com/104441
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
This CL implements |SkSVGDevice::drawAnnotation|, overridden from
|SKBaseDevice|. |drawAnnotation| supports annotating rectangular areas
of a Skia device. Previous to this change, annotations are being used
in |SkPDFDevice| to include hyperlinked rectangular areas in .pdf
documents. This CL implements the SVG equivalent of this PDF feature.
BUG=skia:7581
Docs-Preview: https://skia.org/?cl=104680
Change-Id: I92ae01ceb7ae10cd2010bebab2a58dcfe48ef253
Reviewed-on: https://skia-review.googlesource.com/104680
Commit-Queue: Florin Malita <fmalita@chromium.org>
Reviewed-by: Florin Malita <fmalita@chromium.org>
- mark the interfaces that use SkMask as deprecated
- add more autogenerated subtopics
- make subtopic names singular, avoiding collision with Skia names
- simplify #Deprecated and #Bug tags
- add "#Deprecated soon" to note things to be deprecated
- fix some spelling errors
- refresh web docs
- add self-check functionality to find methods outside subtopics
TBR=caryclark@google.com
Docs-Preview: https://skia.org/?cl=102150
Bug: skia:6898
Change-Id: I0e742a56d49dccd4409bb68eed9167c8ad7611ce
Reviewed-on: https://skia-review.googlesource.com/102150
Commit-Queue: Cary Clark <caryclark@skia.org>
Reviewed-by: Cary Clark <caryclark@skia.org>
Replace manually entered summary tables with ones that are populated
and sorted by bookmaker. This introduces a slight regression for
anonymous enums but fixes a lot of bugs and omissions.
The format is
#Topic somethingTopical
#Populate
##
where somethingTopical is one of Subtopics, Constructors, Constants,
Classes_and_Structs, Members, Member_Functions, and Related_Functions.
Fix the bad formatting in SkCanvas reference. The #Error tag was
was corrupting the markdown table. Remove the tag and replace it
with #NoExample.
Next up: revise self-check to know about populated topics.
TBR=caryclark@google.com
Docs-Preview: https://skia.org/?cl=102080
Bug: skia:6898
Change-Id: Idef5d1c14c740c18a81d6a5106182788dd2a09e1
Reviewed-on: https://skia-review.googlesource.com/102080
Commit-Queue: Cary Clark <caryclark@skia.org>
Reviewed-by: Cary Clark <caryclark@skia.org>
- make descriptions of table entries phrases instead of sentences;
lower case start, no ending period (not enforced, yet)
- add #Line markup to move one line descriptions to the #Method
body. Later, will generate tables like Member_Functions from this
- add #In markup to associate a #Method with a #Subtopic. Later, will
generate tables of related methods from this
- remove return type from operator overloads in tables
- add new colorTypes to examples that index into arrays of strings
to name them
Docs-Preview: https://skia.org/?cl=100422TBR=caryclark@google.com
Bug: skia:6898
Change-Id: I8558048866369f419f1944832b99c05da3fd52bb
Reviewed-on: https://skia-review.googlesource.com/100422
Reviewed-by: Cary Clark <caryclark@skia.org>
Commit-Queue: Cary Clark <caryclark@skia.org>
