The path renderer bitfield will soon be private (for Skia testing
only), so switch to using the public flag.
Test: Simple refactoring change, still builds
Change-Id: I6715ba8addd0e104a46842f2398320d1904bb187
Release VectorDrawable cache surface on render thread.
This is fixing an assert in skia GrSingleOwner.h:33.
Test: Ran gmail before and after the change.
Bug: 64842607
Change-Id: I46e0c2557ac5b2fc3be2cc2d35abf96f6d6c9399
1. automatically parse privacy options and generate lookup table
2. create FdBuffer iterator API in order to remove dependency on Reporter.h
Bug: 64687253
Test: Unit test for iterator API, and manually tested lookup table
Change-Id: I1ea376a4481fc4afc7bdf447936f767b63690fd3
bug:34809371
In some applications, the first draw is not opaque - either because the
application is misbehaved, or because hwui is not able to reliably tell
whether the layer is opaque or translucent. This is undefined behaviour
in OpenGL ES and has a significant performance and bandwidth impact on
some tiler GPUs as it requires loading the previous frame's color data.
This change disables blending in that case and also for effectively
opaque blend modes (SRC=GL_ONE, DST=GL_ZERO). It increases performance
by ~10% for Leanback CTS on some low-end GPUs (gradient layer that hwui
incorrectly believes to be translucent).
Test: manual - visual inspection on fugu (nexus player)
Change-Id: I2cbf1c76678acae1a36923e72fd18ed55cd89dc2
* Increase the ascent and descent of individual lines in StaticLayout
as needed, if any fallback fonts that end up getting used call for
it. For backward compatibility, this is hidden behind a builder
flag.
* Document in Paint.java that the returned parameters are only for
the default font, and a layout may need more space based on
fallbacks used.
Also update for changes in minikin API:
* MinikinFont now requires a method for getting vertical extents
(ascent, descent, and line gap).
* minikin API now allows asking for vertical extents of laid out
text.
* minikin API's LineBreaker now returns ascents and descents for each
line.
Finally, added performances test for creating a StaticLayout.
Follwing are the numbers on a marlin with a stable clock before and
after this CL.
For fixed text almost always hitting the cache:
Before: mean=260684 median=260188 min=258532 standardDeviation=1897
After: mean=262432 median=261509 min=260429 standardDeviation=2185
For random text almost never hitting the cache:
Before: mean=5971827 median=5991126 min=5886871 standardDeviation=83724
After: mean=6337093 median=6317010 min=6311222 standardDeviation=40213
Bug: 28963299
Bug: 29063863
Bug: 32057121
Bug: 37756858
Test: bit FrameworksCoreTests:android.text.
Test: bit CtsTextTestCases:*
Change-Id: I482a98ff8f472e8bab4f0ba9d1d7b368858038ff
Filter empty paths at record time. Quick settings pull test has
about 18% empty paths.
Test: Ran quick settings pull jank test
Bug: 64487466
Change-Id: I5b1650060ae4d157d313efb435ecda26c03d4b72
The main purpose of this CL is reducing font cache size of
low-density device.
The memory usage for the small RGBA texture will be
Nexus 6P: 7,928,856 bytes (1408x1408)
Nexus 5X: 4,734,976 bytes (1088x1088)
These used to be 4,194,304 bytes
Test: manually checked
Bug: 64400885
Change-Id: Ied064a6d59909ad7fbeff74332973206436fbd34
This is the second attempt now that we have addressed many of the
correctness and performance concerns.
Bug: 63910712
Change-Id: I3df4ba52c2bc59930f577ffc8d889cb23a16a439
Test: CtsGraphicsTestCases, CtsUiRenderingTestCases, CtsViewTestCases
Remove all ro.hwui.* tuning props and instead
calculate them from the screen resolution.
Or just hardcode them to what all devices
were hardcoding them to anyway.
Bug: 63741221
Test: Check cache size results on sailfish
Change-Id: I8b0d210572a246f4fefb076935cf5156a70c274c
Merged-In: I8b0d210572a246f4fefb076935cf5156a70c274c
(cherry picked from commit 8dc02f99d0)
Using a shared atlas surface is causing a variance in AA pixels
when drawing the same content. This is causing some CTS tests in
CtsPreference2TestCases to fail because of up to RGB(1,1,1)
difference in AA pixels in screenshots of a recreated activity.
Another effect is reducing memory usage ("GL mtrack" down by
10MB for clock app).
Test: Ran CtsPreference2TestCases tests.
Bug: 64487466
Change-Id: I1e5a547285c885136b6885d129d566d5b0c2ed76
Bug: 64269689
Test: m checkbuild
Test: bit CtsGraphicsTestCases:*
Test: bit CtsViewTestCases:*
Test: bit CtsTextTestCases:*
Change-Id: Idc9f2b4612741be949204cd06efc41e1738d6587
Ensure that RenderNode fitsOnLayer() is true before assigning
it a layer.
Bug: 63814070
Test: repro steps in bug no longer crash
Change-Id: I28bb2cb173a5efde24e2384f2606fea85b394ac8
Implement non-rectangular clips by writing into the stencil
buffer using new Skia API. This CL is fixing
PathClippingTests#testWebViewClipWithCircle test, which was
failing on some devices.
Bug: 34454070
Bug: 31489986
Test: CtsUiRenderingTestCases tests passed for Skia pipeline.
Change-Id: I29d7af02e2af53943540a91393f5d7a8c4e44049
