This can be used to ship source data as test artifacts next to native
tests. It works for both local builds and the test bundles using
package_modules.mk.
You just specify a file list relative to the local directory, and those
files will be copied next to the executable under
/data/nativetest*/<module>/...:
LOCAL_MODULE := mytest
LOCAL_TEST_DATA := data/file1 file2
/data/nativetest/mytest/mytest
/data/nativetest/mytest/data/file1
/data/nativetest/mytest/file2
If the data is in another directory, you may also specify a different
prefix for the source files:
LOCAL_TEST_DATA := external/skia:resources/f.xml
/data/nativetest/skia_test/resources/f.xml
And there's a new convenience macro to find a list of files in this
format:
LOCAL_TEST_DATA := $(call find-test-data-in-subdirs,external/skia,"*.xml",resources)
I'll expand this to native benchmarks and fuzz tests in a later change,
since they don't have their own module classes yet.
Bug: 30564705
Test: m -j minikin_tests; ls $OUT/data/nativetest*/minikin_tests
Test: m -j continuous_native_tests dist; zipinfo -1 out/dist/*continuous_native_tests*.zip
Change-Id: Ic76a7b62e7f567f259c4ab1510ee97d26600ba9a
Moved common variables from other sections back to the common section,
then sort each section.
Test: build-aosp_angler.ninja is identical before/after
Change-Id: Ibb75935205aa80aaa392c358f8d0599ba624f90d
Merged-In: Ibb75935205aa80aaa392c358f8d0599ba624f90d
The PDK uses pattern rules in order to install files from the PDK. When
those files already have build rules, the explicit rules override the
pattern rules, and everything works. But because Make (and Kati) doesn't
attempt to clean the file paths, if one of the rules has a redundant /,
we'll export two ninja rules, and ninja will error out with a dupbuild
error.
The PDK pattern rules are clean, but the explicit notice file creation
was not, it was always adding a double // in between NOTICE_FILES/src
and the module path.
Some modules were also setting a LOCAL_MODULE_PATH with a trailing /,
which is redundant, and also hits the above problem. Instead of fixing
all of the modules, just strip a trailing / from my_module_path.
Bug: 33451638
Test: Build with a PDK
Change-Id: Iff3e98fd191ea90626b9b89f179537e8a75f5ef2
Previously, clang-host would build the clang module for 32-bit and
64-bit for the host and for windows. Make clang-host only compile
for host, and add clang-host-cross for windows. Also add targets
that only build 32-bit or 64-bit: clang-host32, clang-host64,
clang-host-cross32, clang-host-cross64, clang-target32, and
clang-target64.
Test: inspect build.ninja
Change-Id: I7676f9497dfc852f2a0255dda8da06c88eec0db8
The core-junit library has been replaced with the legacy-test
library. The latter included everything that was in core-junit
plus some android.test classes that were in frameworks.
Bug: 30188076
Test: make checkbuild
Change-Id: Id2ba827705dbd2c27a1183e1153b03c11a8a4074
Add support for specifying number of inodes when creating
system, vendor, oem partitions. These are all read-only
and have no use for extra inodes. Removing extra inodes
saves a lot of space.
Bug: 32246383
Change-Id: I13f1d4614b64a4abc752c42a1c65d3d151481c21
(cherry picked from commit b59eca3586)
When creating a sanitized image, skip creating JNI library
symlinks in the second round.
Bug: 33279120
Test: m && m SANITIZE_TARGET=address & ls -l $OUT/system/app/NfcNci/lib/*
Change-Id: Ib5eace9a49eb8b693603ba5cc59e392d575c44e3
So that it can properly install proprietary components in system/vendor/
or vendor/
Test: lunch aosp_arm-eng; m -j vulkan.default
Test: lunch aosp_flounder-eng; m -j vulkan.default
Change-Id: I98384d2b4f09c3b5de2ed04417880f366828234e
We support these in Soong now, so stop marking them in soong_to_convert.
Test: Compare soong_to_convert.txt
Change-Id: I89befdfd8c4eaa04ab626d8be090765a1b879d8b
Building with LOCAL_STATIC_ANDROID_LIBARIES and LOCAL_USE_APPT2
causes a warning:
build/core/package_internal.mk:143: Empty argument supplied to find-subdir-assets
Only call find-subdir-assets if my_res_dir is not empty.
Also improve the warning message to make it easier to find the module
that caused it.
Test: m -j
Change-Id: I9a71162c7e2ed82f64d6844baca256968ac77317
There are no users left, so remove all of this.
Test: lunch aosp_arm-eng; m -j native
Test: build/tools/kati_all_products.sh on aosp and internal master
Change-Id: I32f5c8b470a43dd203d7e20c192167630e4e6888
Add BOARD_VNDK_VERSION and LOCAL_USE_VNDK to specify the version of the
VNDK that will be used globally, and whether to use the VNDK on a module
basis.
If the board is using the VNDK:
* LOCAL_COPY_HEADERS may only be used by modules defining LOCAL_USE_VNDK
* LOCAL_USE_VNDK modules will compile against the NDK headers and stub
libraries, but continue to use the platform libc++.
* LOCAL_USE_VNDK modules will not have the global includes like
system/core/include, but it will use device-specific kernel headers.
This change does not attempt to enforce any linking constraints, that
will come in a later patch.
Test: out/build-aosp_arm.ninja is identical before/after
Change-Id: Icce65d4974f085093d500b5b2516983788fe2905
The last user of this was the NDK, which has been using full paths in
binary.mk. So remove it.
Test: lunch aosp_arm-eng; m -j native
Change-Id: Ida2523a2d19131ee3ef005edb3e5bcf830710b11
We are investigating replacing make_ext4fs with the upstream tool mke2fs.
To mitigate the trouble that may arise if the new tool behave differently
compared to the old one, there will be a transition period.
Devices that want to use the new way of creating ext4 images can set the
variable "TARGET_USES_MKE2FS" to true in their BoardConfig.mk
By default, the build system will choose the old tool 'make_ext4fs'.
Test: m otapackage with TARGET_USES_MKE2FS={,false,true}
Change-Id: I282bcb9efe335a86c53986283090ca947d65c7f8
This make aidl-cpp write out a dep file that ninja parses correctly.
Test: Manually inspect ninja depfile
Change-Id: I4890a91eb29a6388e17b1ffac23a3dc0ffe6c212
