Use DEF_GPUTEST_FOR_*_CONTEXT macros to obtain the
test GPU context.
Makes changing the context -related classes easier,
since not all tests need to be changed.
BUG=skia:2992
Review URL: https://codereview.chromium.org/1448873002
When the parallel stroke to the curve can't be computed from
the intersection of the tangent lines, as straight line connects
the two points instead. Allow the intersection to succeed unless
the ratio isn't finite or the contribution of (1 - ratio) isn't
significant.
R=reed@google.com,fmalita@chromium.org
BUG=skia:4603
Review URL: https://codereview.chromium.org/1484873003
Motivation: for use by fiddle.
Also, add new files to .gitignore
CQ_EXTRA_TRYBOTS=client.skia.compile:Build-Ubuntu-GCC-x86_64-Release-CMake-Trybot,Build-Mac10.9-Clang-x86_64-Release-CMake-Trybot
Review URL: https://codereview.chromium.org/1484053002
The wbmp version of SkImageDecoder will support decoding an image where
the second byte can be masked away with 0x9F. Prior to this CL, SkCodec
checked that the entire byte was zero. The SkCodec implementation
appears to be more correct (at least according to Wikipedia [1]), but
it also means we could regress if someone was using an image that did
not quite fit the specification.
[1] https://en.wikipedia.org/wiki/Wireless_Application_Protocol_Bitmap_Format
BUG=skia:3257
Review URL: https://codereview.chromium.org/1473673005
I'm not sure if this is an improvement but this is an attempt to start moving the handling of pathEffects up the call stack (ideally to SkGpuDevice).
Note that the duplicate path sniffing in GrDrawContext::drawPath is still useful since it can be invoked if a path has to be explicitly drawn for a mask filter.
Review URL: https://codereview.chromium.org/1472333003
Much of the Skia code dealing with clip bounds does not handle
overflow, so if we start doing any math on these rects, the code will
break. Use explicit reasonable values instead.
BUG=skia:
Review URL: https://codereview.chromium.org/1479563002
The crop rect should be applied to the union of the input bounds, not
to the src input's bounds. These are often the same, since the saveLayer
offscreen size is computed as the union of all the required bounds, but
is not correct if the input primitive is not used (e.g., if all inputs
are connected to SkImageSources). But this will change as we
move to more accurate intermediate bounds computations (getting rid of the join() hacks as described in skbug.com/3194).
Since we can't know this without actually processing the inputs,
split SkMergeImageFilter processing into:
- filter all inputs
- applyCropRect to the union'ed bounds
- allocate the destination
- do the merge
BUG=3194
Review URL: https://codereview.chromium.org/1475793002
If the client wants no caching, and we haven't already cached it, pass the
caller's dst-buffer directly down to the generator, avoiding the (previous)
extra memcpy.
BUG=skia:4594
Review URL: https://codereview.chromium.org/1473373002
Replaces "nvpr mode" with "offscreen mode" and a flag to enable/
disable nvpr. Assigns new config names to the various combinations of
flags and begins printing them with the bench results in order to match
nanobench.
BUG=skia:
Review URL: https://codereview.chromium.org/1473253002
Modifies NvprWrappedBenchmark to present its offscreen canvas via
GrContext::copySurface rather than going through the SkCanvas API.
Only copies a subrectangle the size of the benchmark, not the entire
canvas.
BUG=skia:
Review URL: https://codereview.chromium.org/1443263002
Testing:
The testing I did is non-standard. I just compared images using XPS viewer.
They all looked the same, but the XPS device is suffering from bit rot.
BUG=skia:
Review URL: https://codereview.chromium.org/1471043002
Use DEF_GPUTEST_FOR_*_CONTEXT macros to obtain the
test GPU context.
Makes changing the context -related classes easier,
since not all tests need to be changed.
BUG=skia:2992
Review URL: https://codereview.chromium.org/1454663002
Skia'a nanobench can hold onto an image, that causes the lifetime
of GrGLGpu to extend past that of the GL interface.
Kill reference to surface and image before killing the interface
Review URL: https://codereview.chromium.org/1472433002
The GrDrawContext drawOval and drawRRect methods are only called from SkGpuDevice so there is no value in delaying the conversion to a path when dashing is involved.
Review URL: https://codereview.chromium.org/1470103002
