This reverts commit 9c7b131996.
Reason for revert: Fixed build failures due to merge
Change-Id: I7d7bfed3a3234b966f3fe3fd6e0cbc63d2bedf83
Test: unit test
The problem of creating idmap data consists of three steps:
1. create the mapping overlay resid -> target resid
2. fix the padding between the target resids
3. write the mapping as a binary stream
Because we iterate over the resources in the overlay package, there are
no guarantees on the order in which target resources are processed. This
means we can't fix the padding until after step 1. On the other hand,
after step 2 the mapping should be immutable.
Introduce new data structures that help enforce immutability and
ownership of data, as outlined above.
Bug: 80150169
Test: make libandroidfw_tests
Test: atest OverlayHostTests OverlayDeviceTests
Change-Id: I45689e8003c96d251a63a758f77609972ae963df
A tiny scaleX/scaleY is sneaking into the matrix,
throwing off the nearest/bilerp calculation.
This tiny scaleX/scaleY appears to be coming from
the inverse texture matrix necessary to workaround
skia issue https://bugs.chromium.org/p/skia/issues/detail?id=7075
So add another workaround for SkMatrix::getType()
reporting this as having a scale, even though there really
isn't one.
Bug: 119783323
Test: atest android.view.cts.TextureViewTest#testSamplingWithTransform
Change-Id: I3e675102ef99ce093f698460242c19dfe7e90345
Change idmap to iterate over the resources in the overlay package
instead of the target package when scanning for resources defined in
both packages. This cuts down the runtime cost of creating an idmap
considerably since the algorithm now scales with the number of resources
in the overlay package (a handful) and not the number of resources in
the target package (android: 10k, SystemUI: 8k) at a minor cost to code
complexity.
Improvements on the runtime of ResTable::createIdmap (systrace on an
emulator running aosp_x86_64-eng):
- target=android: 12.5 ms -> 3.0 ms
- target=SystemUI: 8.6 ms -> 1.0 ms
The bulk of the cost of creating an idmap from installd is now the fork
and execl to call "idmap --fd ..." which weigh in at 16 ms.
Bug: 80150169
Test: make libandroidfw_tests
Test: atest OverlayHostTests OverlayDeviceTests
Change-Id: I98e18d5958c0cd835a73055b714f5bf0f4f95a09
This changes native InputManagerService to use the new backend library
for inputflinger: libinputreader.
Bug: 119264687
Test: manual: Build, run, and test input (touch, mouse)
Change-Id: I619492d5492cd8444d7c2cc7af9717c19c938256
When themes have two different AssetManagers, we only copy attributes
that exist in both the themes. If the value of the attribute or
reference is equal to 0 (a null reference but not using the TYPE_NULL
type), do not attempt to fix the package id of that null reference.
Bug: 119522708
Test: manual test of broken app
Change-Id: Id615d5a47b7f121b08bbba885d106b9cd3f54afc
The variable sizeResult is an iterator pointing into mCurrentValues,
and clearing mCurrentValues invalidates the iterator.
Bug: 119773465
Change-Id: I940ea397af87561d33f74a9d52abc3afd0a7538f
C++'s "override" keyword implies "virtual". When "virtual" and
"override" are both specified it means something slightly different from
just "override", which is what was intended here.
I teach the C++11 class here at Google. I asked the C++ arbitors to help
me understand this difference. They warned that it is dangerous and not
worth knowing. But that it is worth avoiding.
Test: Tested on local device -- no noticable change.
BUG=115613038
Change-Id: I372821053755fc4cbb1d001b71af7f02dffde5b2
BatteryStats calculates power usage of the device and various components
(such as apps). This information is used, e.g., in the battery panel of
Settings. We now log it to statsd. It can be used for validating how
good the information displayed in Settings is. In the long-term, it is
likely not ideal for off-device calculations, since that can be
hopefully estimated using statsd's raw data.
Three atoms: one for the total power use, one for the power use of each
uid, and one for each non-uid component. Since they will all likely be
pulled together, StatsCompanionService will provide stale data for
BatteryStats pulls called within a second of a previous BatteryStats
pull.
Also in this cl:
Remove StatsLogEventWrapper.writeDouble. Statsd doesn't support actually
writing doubles into its proto reports, so having this function is
misleading (the data will get to statsd and then be completely ignored).
It's less confusing if we don't pretend it does something.
Change-Id: If80bab8ea938afa4632535bb88ff59879fbe8099
Fixes: 119111972
Test: cts-tradefed run cts-dev -m CtsStatsdHostTestCases -t android.cts.statsd.atom.UidAtomTests#testDeviceCalculatedPowerUse
Test: cts-tradefed run cts-dev -m CtsStatsdHostTestCases -t android.cts.statsd.atom.UidAtomTests#testDeviceCalculatedPowerBlameUid
Test: BatteryStatsHelperTest#testDrainTypesSyncedWithProto
Recreate VulkanSurface, if ANativeWindow has been resized. This
new code is hit when wallpaper is resized from 64x64 to real size.
Bug: 119111018
Test: Wallpaper is shawn correctly with Vulkan rendering
Change-Id: I5e43310e5ee8597a7f326d51b1773e2cf68b603a
Prepare the idmap and asset managers for interfacing with idmap2 instead
of today's installd + idmap pipeline, but don't make the switch just
yet.
Instead, idmap2 runs as its own native daemon with an AIDL interface.
This removes the need for installd to fork and exec on each idmap call,
saving about 50 ms per call.
Bug: 78815803
Test: atest OverlayDeviceTests OverlayHostTests
Change-Id: I60852e15d99329896ff9de6559d1e7cd1c67e33d
Split out the View/ViewRootImpl bits from the
hardware rendering bits.
Create a potential public API surface for
hardware rendering
Bug: 112709971
Test: builds & boots
Change-Id: I9e6f44b07a170574a905f42338282c4bb7e95f56
Also for a canvas wrapping a bitmap the colorspace of the bitmap
will be used to correctly blend content.
Test: CtsUiRenderingTestCases
Bug: 111436479
Change-Id: I63ad7a30605a7f725cc0ef4716d42ea978fb03e3
The entry pool is encoded in UTF-8 not UTF-16. This would cause stringAt
to return a pointer to a cache that holds the UTF-8 string converted to
UTF-16. If the cache becomes invalidated, the UTF-16 pointer is no
longer valid and causes seg faults when looked up.
Test: libandroidfw_tests
Change-Id: I0f13dbeadd0d148de2805efd6cadb8e220368e3b
