it is meaningful only if it is installed in the device
Bug: 277909042
Test: cargo2android.py with/without device, and then check
product/vendor_available
Change-Id: I9ca7328ac953849db224949089412e181b659945
The suffix ".map.txt" triggers NDK-API-Review.
example.map.txt has been replaced with map.txt in the same directory.
Test: ANDROID_BUILD_TOP=`realpath .` \
PATH=$PATH:`realpath out/soong/dist/bin` \
development/vndk/tools/header-checker/tests/test.py
Bug: 298626434
Change-Id: Ie2e21fe87e5bfce7f081fe32eb55547d8aa611ad
Get the size of build artifacts using df.
It collect the amount of disk space used for build for all targets.
Bug: 302075773
Test: combined_build_test.sh -t <target> -v userdebug \
-a <alter_target> -o -c droid
Change-Id: I4c59bc181164f4e8e13277719dc16158d3814f49
Use the newly created srcjar for the bootclasspath as opposed to the
previous method of scanning the tree for *.java files.
The new srcjar har some known limitations:
- jarjar renaming / zapping is not taken into consideration
- optional classpath libraries (<library>) are not included in the
sources at the moment.
- when prebuilts are used for bcp jars, their sources are not included
Overall it should still be a large improvement over the current jar
which contains a lot of irrelevant cruft, and is very incomplete.
Also delete the old script.
Bug: 151360309
Test: m sdk sdk_repo dist; inspect the sources.zip file
Change-Id: I1cf96966334e15c202a0ea38d6133581bb02b80c
rust_library + host_suffix + rlib_suffix is the right order
For example, rust_library_host_dylib, rust_library_host_rlib
Bug: 277909042
Test: cargo2android --run --force-rlib
Change-Id: Ib156b96853eb51bae1f68229ebe76b87b0f8b363
The version blocks ending with _PLATFORM and _PRIVATE are unstable ABIs.
To exclude them, header-abi-linker --exclude-symbol-version should
accept patterns like *_PLATFORM and *_PRIVATE.
Test: ANDROID_BUILD_TOP=`realpath .` \
PATH=$PATH:`realpath out/soong/dist/bin` \
development/vndk/tools/header-checker/tests/test.py
Test: LD_LIBRARY_PATH=out/soong/dist/lib64 \
out/soong/host/linux-x86/nativetest64/header-checker-unittests/header-checker-unittests
Bug: 298626434
Change-Id: Id82bc013b8de62b5ba8e5bd29f399b1b72ab5d39
This is helpful for crates like protobuf and syn for which we have
multiple versions, and will let us eliminate several patch files
Test: Ran it on everything in external/rust/crates.
Change-Id: I29ac1ee69be99b7164e1458e0d8fc1cb028b15a0
build-prebuilts.sh compiles header-checker-unittests by default. The
continuous builds ensure that the tests are compilable. The developers
can use the script to build the tests locally.
Test: OUT_DIR=out BUILD_BROKEN_DISABLE_BAZEL=1 \
development/vndk/tools/header-checker/android/build-prebuilts.sh \
header-checker-unittests
Test: OUT_DIR=out BUILD_BROKEN_DISABLE_BAZEL=1 DIST_DIR=/tmp/distdir \
development/vndk/tools/header-checker/android/build-prebuilts.sh
Bug: 280008249
Change-Id: I32d6ade3c527785d68802c1737c94f963ead5b3c
One flag '-f' corresponds to one directory that stores a parsed report.
Bug: 296988808
Test: ./compare_cts_reports.py -r test_result_1.xml [test_result_2.xml] \
[-r report.zip] [-f parsed_dir_1/] [-f parsed_dir_2/] -m n -d output_dir/
Change-Id: I26310982dc8ef18f5dbfb1a8382ea3a600508fbe
This is cleaner than using .unwrap_or(true) wherever they are used.
Bug: 293289578
Test: atest cargo_embargo.test
Change-Id: I763adf0cc51721408adf55b966ed4304e2766c00
