We don't need to syncronize the mapped memory writes to the buffer since
all CPU writes are already syncronized when we submit a command buffer. And we are using coherent memory for buffers so we don't need to call vkFlushMappedMemory
BUG=skia:
Review-Url: https://codereview.chromium.org/2289973002
This ought to support compiles for now.
Am I picking up my CIPD ndk packages right?
The main thing to note is that I'm passing the target_arch directly through
as target_cpu. This means these bots will have a slightly different naming
convention than we've been using, but it'll agree with what you must type
yourself when using GN to build for Android:
- Arm7 -> arm
- Arm64 -> arm64
- Mips -> mipsel
- Mips64 -> mips64el
- x86 -> x86 (unchanged)
- x86_64 -> x64
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2292663002
Review-Url: https://codereview.chromium.org/2292663002
Changes the Android shadow sample to use stroked roundrects when
we can (mainly if stroked geometry area < fill geometry area).
Also changes the setup for the overstroke geometry so that it computes
the correct distance to the outer edge.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2283003003
Review-Url: https://codereview.chromium.org/2283003003
ANGLE is using the wrong external type (GL_HALF_FLOAT) when querying for
ReadPixels format, but it validates parameters to TexImage2D against the
correct type (GL_HALF_FLOAT_OES). Rather than introduce a lot of
complexity to work around the inconsistency, I'm just going to remove
support for now.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2288233002
Review-Url: https://codereview.chromium.org/2288233002
This makes it considerably easier to use ccache with the Android NDK.
You can now just set
compiler_prefix = "ccache"
ndk = "/path/to/ndk"
and we'll use the NDK clang, wrapped with ccache.
The name compiler_prefix is stolen from / compatible with Chrome.
If you have ccache, you can just always leave compiler_prefix="ccache" enabled.
This should make it an unusual thing for humans to have to change cc or cxx.
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2281163002
Review-Url: https://codereview.chromium.org/2281163002
Instead of mapping invaid glyphIDs to zero or maxGlyphID,
don't draw them at all.
Validate glyphs when glyph is written, not ahead of time.
Don't allocate array to copy user-provided glyphs.
Easy early exit from SkPDFDevice::internalDrawText()
GlyphPositioner::flush() called ~GlyphPositioner()
SkScopeExit class now exists.
Assume SkTypeface* pointers are now never null in more
places.
precalculate alignmentFactor to clean up code.
SkPDFDevice::updateFont must be called with validated
glyphID. Skip bad glyphs to make this true.
SkPDFDevice::updateFont always succeeds.
SkPDFFont::GetFontResource always succeeds (preconditions are
asserted). If GetMetrics fails, don't call GetFontResource.
SkPDFFont::glyphsToPDFFontEncodingCount() becomes
SkPDFFont::countStretch() and is inlined.
SkPDFFont::glyphsToPDFFontEncoding now works one Glyph at a
time and is inlined.
SkPDFFont::noteGlyphUsage() operates one glyph at a time.
Add SkScopeExit.h; also a unit test for it.
SkPostConfig: Fix SK_UNUSED for Win32.
No public API changes.
TBR=reed@google.com
BUG=625995
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2278703002
Review-Url: https://codereview.chromium.org/2278703002
Currently when building for Android framework the freetype target
exports the expat include directory, but the freetype target neither
uses nor logically provides this information. Instead, the expat target
itself should do this.
Review-Url: https://codereview.chromium.org/2283903003
With this version, it's now exposing GL_EXT_color_buffer_half_float, but
it's (incorrectly) using GL_HALF_FLOAT rather than GL_HALF_FLOAT_OES.
In addition, release builds were crashing in EGLImageTest due to an
apparent disagreement about calling convention on function-pointer to
exported symbol. There is a proper typedef for that in one of their
headers, but I can't seem to include it without creating more problems,
so I just fixed the decoration on our local typedef.
With those changes, all tests pass, and (on my Windows machine), all
GMs and SKPs produce identical results vs. master in angle and angle-gl.
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2282733003
Review-Url: https://codereview.chromium.org/2282733003
Chrome's complex text path uses HarfBuzz OpenType glyph lookup
functions. These do not support glyph lookup in Type 1 / Postscript
fonts and we do not wish to support those in Chrome any longer.
In order to avoid matching fonts against Type 1 fonts possible present
on Linux installations and exposed through fontconfig, this CL filters
those out in Skia's font matching code.
This change might not be desirable for all context in which Skia is
used, hence making it conditional on a Chrome context define.
BUG=skia:5685
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2280053002
Review-Url: https://codereview.chromium.org/2280053002
This seems to work well for miter and bevel joins with the resulting stroke and fill path remaining convex. There seems to be an issue with round joins where the outer generated shell is usually not convex. Without this CL the resulting stroke & filled paths are always concave.
Perf-wise (on Windows):
convex-lineonly-paths-stroke-and-fill bench
(in ms) w/o w/CL %decrease
8888 2.88 2.01 30.2
gpu 4.4 1.38 68.6
GOLD_TRYBOT_URL= https://gold.skia.org/search?issue=2275243003
Review-Url: https://codereview.chromium.org/2275243003
