There was a bug that when an unbundled app is built without SDK
(LOCAL_SDK_VERSION isn't set), it is linked to the nonexisting stubs
library from java_sdk_library. The built stubs libs are not created for
unbundled builds. For unbundled builds, we need to use the prebuilt
stubs lib.
Bug: 77575606
Test: m -j
Test: make -j dist
ANDROID_BUILDSPEC=vendor/google/build/app_build_spec.mk
Change-Id: I70cced1c872053e151fea157e3a6d218b1fab9dc
Fix some incorrect dependencies after I12d33e52019aebac6ea33271939228e4690a4173.
Bug: 73885582
Test: m java FORCE_AAPT2=true
Change-Id: Ica2758991603e1605f629d843d19df5e2ef4fae1
Renderscript was using LOCAL_RENDERSCRIPT_TARGET_API := 0 to mean
current. Use current instead, and check for it when comparing
if LOCAL_RENDERSCRIPT_TARGET_API >= 21.
Fixes an issue where LOCAL_RENDERSCRIPT_TARGET_API := 0 was being
treated as < 21, causing the renderscript resources to be passed
to aapt2:
out/target/common/obj/APPS/RSTestBackward_intermediates/flat-res/zip_res.flata.contents/raw/bc32: error: resource file cannot be a directory.
Bug: 79481102
Test: m java FORCE_AAPT2=true
Change-Id: I7726d0463ce1693bc9653b4d869ced181c491036
LLVM r328258 turned on a feature called temporary dtor inlining by
default for all of C++ in clang-tidy. This feature appears to be
somewhat over-aggressive when objects are being passed by value. For
example, given:
void foo(std::unique_ptr<int> i);
void bar() {
auto x = std::make_unique<int>();
int *i = x.get();
foo(std::move(x));
*i = 99;
}
...clang-tidy will complain about `*i = 99;` being a definite
use-after-free. This is incorrect, however: `foo` could stash the
`unique_ptr` it's given in a global, or a class member, or ...
Until upstream fixes this bug, it's probably best to keep this disabled.
Bug: None
Test: Ran the analyzer across Android locally. Nothing broke; number of
complaints dropped significantly.
Change-Id: I742eedf598a72a533285d913d191bfbbf0ce1688
For SDK builds, $(AAPT) points to a prebuilt and aapt is not
necessarily in the path.
Test: make -j32 <some SDK build> tests
Bug: 70934104
Bug: 67345922
Change-Id: Icfd62a4e3ec284ed59a61714765e5a909a27947d
Targets can define 'TARGET_USERIMAGES_SPARSE_EXT_DISABLED := true' to
generate non-sparse system images, but validate_target_files.py doesn't
work with such images. This CL adds a workaround to temporarily skip the
file consistency check for such images.
Bug: 79616357
Test: Run validate_target_files.py on a target_files.zip that's not
using sparse image.
Test: Run validate_target_files.py on marlin target_files.zip (which
uses sparse image).
Change-Id: I1f4066c5b3fec595b10cab10283d62c1c5a6c624
When linking in the UBSan minimal runtime, don't export the symbols.
This was resulting in an edge case where symbols were sometimes
undefined at runtime on address sanitized builds if static library
dependencies were integer overflow sanitized.
Bug: 78766744
Test: readelf on libraries show either inclusion of the shared library
or no undefined symbols related to the minimal runtime.
Change-Id: I4382cc72baefd7fa96cd83e8349e82f7b083f5aa
Properly check the value of PRIVATE_ENFORCE_USES_LIBRARIES instead of
only checking that the variable is non-empty.
Change the logic to use product libraries instead of the common ones.
This is required so that dex2oat can find the oat files when stripping
is enabled.
Bug: 70934104
Bug: 67345922
Test: manual
Change-Id: Iee71c2b26f3c0e136a8971ab913725b2b6e100b8
Building with FORCE_AAPT2=true will turn on AAPT2 for all modules
unless they set LOCAL_USE_AAPT2 := false. The build system will
attempt to rewrite common AAPT patterns into AAPT2 patterns,
including removing --extra-packages for support library packages,
removing LOCAL_RESOURCE_DIR point to support library resources,
adding a default empty manifest file if it doesn't exist, and
converting LOCAL_STATIC_JAVA_AAR_LIBRARIES to
LOCAL_STATIC_ANDROID_LIBRARIES.
Bug: 79481102
Test: m checkbuild
Change-Id: I8d9d55fe4d5d5c965c64b0407efe74e0afc35c3a
Patch Iaf5853361767f0ea3e9ad4ec908aadd72d639c79 missed one rename
of my_support_library_sdk_raise to my_proguard_sdk_raise.
Test: m checkbuild
Change-Id: Ie3d4536bd80654a7eb4f209ef6d6a8de502a6cdc
Allow non-negative integers instead of positive integers in math
functions so that LOCAL_RENDERSCRIPT_TARGET_API := 0 can be compared
against 21.
Test: make -f ${ANDROID_BUILD_TOP}/build/make/core/math.mk RUN_MATH_TESTS=true
Test: $(get_build_var CKATI) -f ${ANDROID_BUILD_TOP}/build/make/core/math.mk RUN_MATH_TESTS=true
Change-Id: If2504b7de81f8dd676af749a04c9f3e580f298b6
Turn commented out tests into assertion tests, and add more for
functions that didn't have tests.
Test: make -f ${ANDROID_BUILD_TOP}/build/make/core/math.mk RUN_MATH_TESTS=true
Test: $(get_build_var CKATI) -f ${ANDROID_BUILD_TOP}//build/make/core/math.mk RUN_MATH_TESTS=true
Change-Id: I411844fbf00268f863260b7ee42d1795f05c0d16
Required libraries are specified with LOCAL_USES_LIBRARIES
Optional libraries are specified with LOCAL_OPTIONAL_USES_LIBRARIES
The make rule cross references the libraries against what's stored
in the manifest.
Verification is enabled if LOCAL_ENFORCE_USES_LIBRARIES is true. This
defaults to true if either of LOCAL_USES_LIBRARIES or
LOCAL_OPTIONAL_USES_LIBRARIES are specified.
Bug: 70934104
Bug: 67345922
Test: manual
(cherry picked from commit 09f3b97f4b)
Merged-In: Ifca7d1a993620e9d0e42dc497a4a5d7a6c3f4172
Change-Id: I670431f938c31115a7812c1857c31b9f71675632
It's been converted to Soong, so can no longer use LOCAL_MODULE_TAGS. It
appears that some of the test runners expect to have it pre-installed.
Test: m; see it installed.
Change-Id: I189c9bbc2f2add90ac4f6841f4bf13852e4d579f
Soong needs PLATFORM_VERSION to pass as --version-name when building
framework-res.apk.
Bug: 78324052
Test: aapt dump badging $OUT/system/framework/framework-res.apk | grep -i version
Change-Id: If47839182a1dd7c7a768b1253dd8b575d777ca7e
Merged-In: If47839182a1dd7c7a768b1253dd8b575d777ca7e
(cherry picked from commit 34aba6cf72)
This reverts commit da8febf39c.
Bug: http://b/73768157
Reason for revert: breaks several bionic/debuggerd/libunwind_stack gtests
Change-Id: I161cf85cf352b80644524fd4bb70ff7c75fbc1d0
We used to write a fake post-timestamp into package metadata for
"--override_timestamp", which allows a package to be pushed based on
the fake timestamp. This CL stops doing that by using the designated
"ota-downgrade=yes" instead, but keeping the "--override_timestamp" flag
for backward compatibility.
Now both "--override_timestamp" and "--downgrade" allow pushing packages
with downgrade timestamps. The only differenece is that "--downgrade"
enforces a data wiping, because it's intended a real downgrade (based on
Android version or security patch level). "--override_timestamp" serves
the path of a legit "upgrade", but unfortunately with build timestamps
in reverse order (e.g. the two builds are cut from different branches).
With this CL, we write "post-timestamp" to all packages.
Bug: 77811049
Test: Generate an incremental pakcage with "--override_timestamp".
Test: Generate an incremental pakcage with "--downgrade".
Test: python -m unittest test_ota_from_target_files
Change-Id: I4ae4ce77cb600a06ffd9b15b6d4c730314aa0b79
This reverts commit 44988be862.
Three known build failures with lld are fixed by
(1) adding --no-fatal-warnings to link
(2) disable LLD for some PGO build targets
(3) disable LLD for some art oatdump tests, pending c/681542
Change-Id: Icebf6661509bf4f6051f91dc74eaa07dbc7c1d42
Jiyong Park