This has the side effect of requiring SkNullGLContext to use the null GL interface.
It exposes SkNullGLContext and also removes null context support from SampleApp.
Review URL: https://codereview.chromium.org/916733002
The language was being set to garbage, now set to part of the file name.
Add a test to ensure we continue to parse fallback directories correctly.
BUG=chromium:422180
Review URL: https://codereview.chromium.org/912053003
Qt changes the context on many conditions. Abandon GrContext
in these cases.
Also call GrContext::resetContext during Qt GL paint callback,
the GL state may be touched by Qt.
Fixes the bug where changing between gpu and msaa would start
erroring in framebuffer binds, if the .skp had many layers.
Review URL: https://codereview.chromium.org/915573002
Reason for revert:
Going to punt on 16-bit float support for now. Can't figure out ARM 64.
Original issue's description:
> GYP groudwork for half-float opts support.
>
> This sets us up two new opts targets with the immediate goal of adding half-float (SkHalf.h) opts:
> - opts_neon_fp16: uses hardware support on most ARM chips with NEON to do 4 conversions at a time;
> - opts_avx: uses hardware support on Intel chips with AVX to do 8 conversions at a time.
>
> opts_avx will be a handy thing to have around later too, especially if we want to work with floats.
>
> This doesn't actually add any new source files to these libraries yet, so they're no-ops for now.
> I'll need to write a parallel change to Chrome's GN and GYPs before we can start adding sources.
>
> This also rolls GYP up to head, to get suppport for EnableEnhancedInstructionSet: '3' on Windows,
> which is how we turn on AVX there. There's no Mac-specific flag, so we use OTHER_CPLUSPLUSFLAGS.
>
> BUG=skia:
>
> TBR=reed@google.com
>
> Committed: https://skia.googlesource.com/skia/+/46b80833394d7919cadf2abf2b93802141dd21c5TBR=reed@google.com,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/912223002
Prior to this change, SkPDFObject subclasses were required
to track their resources separately from the document
structure. (An object has a resource if it depends, via an
indirect reference, on another object). This led to a lot
of extra code to duplicate effort. I replace the
getResources() function with the much simpler addResources()
function. I only define a non-trivial addResources() method
on arrays, dictionaries, and indirect object references.
All other specialized classes simply rely on their parent
class's implementation.
SkPDFObject::addResources() works by recursively walking the
directed graph of object (direct and indirect) references
and adding resources to a set. It doesn't matter that there
are closed loops in the graph, since we check the set before
walking down a branch.
- Add SkPDFObject::addResources() virtual function, with
four implementations
- Remove SkPDFObject::getResources() virtual function and
all implementations.
- Remove SkPDFObject::GetResourcesHelper()
- Remove SkPDFObject::AddResourceHelper()
- In SkPDFCatalog::findObjectIndex(), add an object to the
catalog if it doesn't exist yet.
- SkPDFCatalog::setSubstitute() no longer sets up resources
- SkPDFDocument.cpp no longer needs the Streamer object
- SkPDFDocument.cpp calls fDocCatalog->addResources to build
the resource list.
- SkPDFFont::addResource() removed
- All SkPDF-::fResource sets removed (they are redundant).
- removed SkPDFImage::addSMask() function
- SkPDFResourceDict::getReferencedResources() removed.
Motivation: this removes quite a bit of code and makes the
objects slightly slimmer in memory. Most importantly, this
will lead the way towards removing SkPDFObject's inheritance
from SkRefCnt, which will greatly simplify everything.
Testing: I usually test changes to the PDF backend by
comparing checksums of PDF files rendered from GMs and SKPs
before and after the change. This change both re-orders and
re-numbers the indirect PDF objects. I used the qpdf
program to normalize the PDFs and then compared the
normalized outputs from before and after the change; they
matched.
Review URL: https://codereview.chromium.org/870333002
Darken was 5-20% slower with this CL, though it might improve things a bit to only look at one component of the vector.
BUG=skia:
Review URL: https://codereview.chromium.org/915633002
Adds an SSE4.1 version of the Color32A_D565 function.
Performance improvement in the following benchmarks:
Xfermode_SrcOver - ~100%
luma_colorfilter_large - ~150%
luma_colorfilter_small - ~60%
tablebench - ~10%
chart_bw - ~10%
(Measured on a Atom Silvermont core)
Signed-off-by: Henrik Smiding <henrik.smiding@intel.com>
Review URL: https://codereview.chromium.org/892623002
This sets us up two new opts targets with the immediate goal of adding half-float (SkHalf.h) opts:
- opts_neon_fp16: uses hardware support on most ARM chips with NEON to do 4 conversions at a time;
- opts_avx: uses hardware support on Intel chips with AVX to do 8 conversions at a time.
opts_avx will be a handy thing to have around later too, especially if we want to work with floats.
This doesn't actually add any new source files to these libraries yet, so they're no-ops for now.
I'll need to write a parallel change to Chrome's GN and GYPs before we can start adding sources.
This also rolls GYP up to head, to get suppport for EnableEnhancedInstructionSet: '3' on Windows,
which is how we turn on AVX there. There's no Mac-specific flag, so we use OTHER_CPLUSPLUSFLAGS.
BUG=skia:
TBR=reed@google.com
Review URL: https://codereview.chromium.org/915693002
The macro is only used in CrashHandler.*
Removes SK_CRASH_HANDLER from Android's SkUserConfig, where it is not
needed.
Review URL: https://codereview.chromium.org/915663002
Properly labels several methods as static.
Use XML_GetBuffer to avoid an extra copy.
Set the memory allocators to Skia's.
Set define in 'defines' instead of cflags.
Update debug dumper.
Review URL: https://codereview.chromium.org/915443002
No algorithmic changes. The new APIs let us avoid a few ugly trips through void*,
and I've made the consume/acquire/release decision explicitly conditioned on TSAN.
This should fix the attached bug, which is TSAN seeing us implementing the
sk_consume_load() with a relaxed load, where we used to pass __ATOMIC_CONSUME
to TSAN. This restores us to the status quo of a couple weeks ago, where we
use relaxed loads (to avoid an extra dmb on ARM) for all setups except TSAN,
who gets the logically correct memory order, consume.
No public API changes.
TBR=reed@google.com
BUG=chromium:455606
Review URL: https://codereview.chromium.org/908943002
In 32 bits, if the "element count" was under 32 bits, but "element count" * "element size" was over the 32 bit limit, the overflow was causing the read operation to appear as if it had succeded, even though it should have failed.
BUG=456828
Review URL: https://codereview.chromium.org/904833003
Removes the opaque-only versions of this function from the factory since
they will never be used. Opaque source colors are handled in
SkRGB16_Opaque_Blitter instead, which doesn't use the factory function.
Signed-off-by: Henrik Smiding <henrik.smiding@intel.com>
Review URL: https://codereview.chromium.org/901593002
