If the callback passed to Once panics it was leaving the waiter
in place that would never be completed. Move writing the value
and signalling the waiter to defer.
Test: TestOncePerPanic
Change-Id: Icc4d3b779a79914fcd881d61d38dffcc2f591c39
There are a couple of makefiles that have etc with $(TARGET_OUT)/usr/share
as the local module path. Added the conversion in androidmk for target
and host.
Bug: b/132123818
Test: Wrote and ran unit test cases, did a test conversion on
external/neven/Android.mk makefile.
Change-Id: Iafed89f4cee499f561a1235f8870b1f1329e99bc
Writing $OUT/dexpreopt.config during build setup and then using
it during the build causes bad incremental builds if a cleanspec
removes $OUT/dexpreopt.config. Create a rule in Soong to write
out a copy so that it is a normal build artifact.
Bug: 132379813
Test: m checkbuild
Change-Id: I7755a7f8c1d6561fc83022a7f4cf7e4a64e989c9
This implements the uncompress-dexs macro's behavior in Soong.
Test: Converted TvSampleLeanbackLauncher + zipinfo
Change-Id: I9477aa21429d055f3f36ca90c7fd2c345c999029
Running any executable built with asan and libc_scudo results in crashes.
Bug: 131810078
Test: Built the media processes with the sanitizer enabled verifying
Test: that the libc_scudo.so shared library is not listed in the
Test: mediaextractor and the mediaswcodec binaries.
Test: Build the media processes without the sanitizer enabled verifying
Test: that the libc_scudo.so shared library is listed in the
Test: mediaextractor and the mediaswcodec binaries.
Change-Id: Ie55a0414088787ec11d85fffcb04592ed4f23c57
The ndk_library, ndk_prebuilt_object and ndk_prebuilt_static_stl
modules are natural targets to have native bridge support enabled,
since they build user facing object which have to be supported for
translated architectures.
Bug: http://b/77159578
Test: make
Change-Id: Ic556f4c1c41e5b3dc92f9c290b4482dee8faed33
The current EXPERIMENTAL_USE_OPENJDK9 environment variable is a legacy
from when the 8 -> 9 toolchain upgrade was happening. That migration
is done and the variable only affects the language level, so it should
have a name that reflects that. (The current situation would be
especially confusing if we started a 9 -> 11 toolchain upgrade,
presumably controlled by a variable like EXPERIMENTAL_USE_OPENJDK11,
since the two settings look incompatible but are actually orthogonal.)
The current variable historically allowed a value or "1.8" which meant
"use the OpenJDK 9 toolchain but target language level 8". That value
no longer has any meaning and the new variable doesn't allow it.
Bug: 131678633
Test: `make` with `EXPERIMENTAL_JAVA_LANGUAGE_LEVEL_9=true`
Test: `javap -v $OUT_DIR/soong/.intermediates/libcore/core-oj/android_common/javac/classes/java/util/List.class | grep 'major version'` shows 53
Test: Audit all mentions of the old string here and downstream
Change-Id: Idad808c7f07913baba1a777627322d5452dabcef
This is required to build the recovery variant of boot control HAL.
Bug: 78793464
Test: make
Bug: 34254109
Change-Id: Ia5d7b1a97926acbe4ba89df26aeb66fac6bbb1ab
Merged-In: Ia5d7b1a97926acbe4ba89df26aeb66fac6bbb1ab
(cherry picked from commit 1a3b62757c)
Colin Cross