An earlier CL with benchmarks has shown that sending strings as UTF-8
is 50% faster for US-ASCII strings, and still 68% faster for complex
strings referencing higher Unicode planes. (So an improvement in
both cases!)
Since code across the OS still makes heavy assumptions about Parcel
strings typically being UTF-16, we need to carefully migrate
Parcelables by hand, which is what this CLs begins doing.
Bug: 154436100
Test: manual
Change-Id: I9a675473e0ce3a4a5c5e305dd851b40bb1560e1c
This is a temporary failure, does not fail the session, but requires
caller to re-commit. E.g. there are connectivity issues which can be
fixed later.
Bug: b/153874006
Test: atest PackageManagerShellCommandTest PackageManagerShellCommandIncrementalTest IncrementalServiceTest
Change-Id: I02791a2963130dbecb510c4a7cafcf04f6245761
When a package that is part of a shareduid is installed / updated its
lineage is compared against the current signature / lineage for the
shareduid. If the new package has more than one signature in the lineage
an IndexOutOfBoundsException is thrown by SigningDetails#hasAncestor
crashing the system server. This commit ensures the ancestor check
is only performed against the current signer for the shareduid.
Fixes: 155134046
Test: atest SigningDetailsTest
Change-Id: Icef9f21c6901e255e5276085259d2f773f41e858
This change removes the initial target SDK check which prevents apps
that target pre-Q from testing that they fail to be install if the
compatibility change is enabled for pre-Q packages.
Bug: 132742131
Test: adb shell am compat enable 132742131 for app with compressed ARSC
that targets pre-R
Change-Id: I7e568a9e99045c09565bb372e454f573b954fea5
This reverts commit 472371c6ad.
Reason for revert: The feature should never be enabled on R and above.
Bug: 111577843
Change-Id: I4e15231909d120d548fa7b7623276e518e32f7a5
Test: ClipboardManagerTest
This simplifies:
- resource management - no extra copies of controls,
- state management - all states in one place, no more hidden (bound but
not created) state.
Bug: b/153874006
Test: atest PackageManagerShellCommandTest PackageManagerShellCommandIncrementalTest IncrementalServiceTest
Change-Id: I3d16a099c7f42fcf14637c5a8e96bd6f99e073d1
This reverts commit 472371c6ad.
Reason for revert: The feature should never be enabled on R and above.
Bug: 111577843
Change-Id: I4e15231909d120d548fa7b7623276e518e32f7a5
Test: ClipboardManagerTest
This change is to prevent misuse of window context from app
and leads to performance drop on system by limit the numer of window
context an app can use. Code snippet below is a sample to cause
this issue:
```
Rect getBounds() {
Context windowContext = context.createWindowContext(...);
return windowContext.getSystemService(WindowManager.class)
.getCuttentWindowMetrics().getBounds()
}
```
This method could be invoked dozens of times and produce dozens of window
tokens. It would slow down the speed of window traversalling. These
token won't be removed until system server has been GC'd.
Test: atest WindowContextTests WindowContextPolicyTests
fixes: 152934797
Bug: 153369119
Change-Id: I927e85a45c05c4d90b51a624ea408ff3a3ffce93
An earlier CL with benchmarks has shown that sending strings as UTF-8
is 50% faster for US-ASCII strings, and still 68% faster for complex
strings referencing higher Unicode planes. (So an improvement in
both cases!)
Since code across the OS still makes heavy assumptions about Parcel
strings typically being UTF-16, we need to carefully migrate
Parcelables by hand, which is what this CLs begins doing.
This is a purely mechanical refactoring with no functional changes.
Bug: 154436100
Test: manual
Exempt-From-Owner-Approval: trivial refactoring
Change-Id: Ia9e581efd7c40269342b7528ca07363deb843c0f
Recently while investigating some Binder limits, I discovered that
we're still sending Strings across Binder as UTF-16, which is very
wasteful for two reasons:
1. The majority of data flowing through APIs like PackageManager is
already limited to US-ASCII, and by sending UTF-16 we're wasting
half of our transactions on null-byte overhead.
2. Internally ART is already "compressing" simple strings by storing
them as US-ASCII instead of UTF-16, meaning every time we want to
write a simple string to Binder, we're forced to first inflate it
to UTF-16.
This change first updates Parcel.cpp to accept char* UTF-8 strings,
similar to how it accepts char16_t* for UTF-16. It then offers
both UTF-8 and UTF-16 variants to Parcel.java via JNI. We also
update the String8 handling to behave identical to String16.
This change adds benchmarking to show that these new methods are
about 50% faster for US-ASCII strings, and about 68% faster for
complex strings that reference higher Unicode planes. (So an
improvement in both cases!)
Bug: 154436100
Test: atest FrameworksCoreTests:ParcelTest
Test: make core-libart conscrypt okhttp bouncycastle vogar caliper && vogar --mode app_process --benchmark frameworks/base/core/tests/benchmarks/src/android/os/ParcelStringBenchmark.java
Change-Id: I22a11d3497486d922ec8e14c85df66ca096b8f2a
