Make android_app modules a MultiTargets module, which means the
common variant will have a list of Targets that it needs to handle.
Collect JNI libraries for each Target, and package them into or
alongside the APK.
Bug: 80095087
Test: app_test.go
Change-Id: Iabd3921e1d4c4b4cfcc7e131a0b0d9ab83b0ebbb
Add -0 flag to convert files in a zip to stored instead of deflated.
Bug: 69500920
Test: zip2zip_test.go
Change-Id: I6c2b10f3b200a53a3339e3c97a78f65192b309ca
JNI testing will need to create basic native shared library
modules, export the minimum mutators and module types for
the required dependencies of a native shared library.
Bug: 80095087
Test: cc_test.go
Change-Id: Ibe7bc88b69cb0851291cb09a4c0c6cdb421b8651
Allow modules to specify that they will handle multiple targets in
the common variant. This will be used by android_app modules to
handle JNI libraries from multiple architectures.
Bug: 80095087
Test: m checkbuild
Change-Id: Iede3e9c23b64fb516341c3ae08074a322b511d40
Remove more GCC-specific flags, TARGET_(arm|thumb)_CFLAGS now contains
pre-filtered clang flags, and GLOBAL_CLANG_CFLAGS_NO_OVERRIDE now
contains all flags instead of just the clang additional flags.
Test: m
Change-Id: I61e4c5a92113d4ff3b95261783269dc1d226c57a
Instead, hardcode the ~dozen paths into build/soong/Android.bp, which
will unblock removing more GCC support.
Bug: 114286031
Test: m
Change-Id: I2508432e00b1469141f01e667f3c6a2fe30cd805
If Srcs contains only srcjar, sourcepath is null and an error occurs. So
if sourcepath is null, we will use the "-sourcepath "" ".
Test: m -j
Bug: 112397488
Change-Id: I03ac0074fc041203fa1b427d4b4a418af44e85e2
This reverts commit 77e62dbf59.
Reason for revert: caused boot failure on mirror-aosp-master-with-vendor
Change-Id: I594df35878616dba8990d044d7cd529458a601b6
The path and dir are used by apex to include prebuilt files in APEXs.
Bug: 112672359
Test: m apex.test
Change-Id: I780edc8f15e00e644c41fa2eb9dc73c25339c727
libgcc is kept behind libcrt.builtins to provide unwinder symbols and
any other missing symbols.
libc and libm are excluded from the switch due to some symbols being
hidden in libcrt.builtins but not in libgcc. These will be addressed in
a separate CL.
Bug: 29275768
Test: checkbuild and bionic native tests on arm/arm64/x86/x86_64
Change-Id: Icc85bf88513e989a4b1547564405a22cf847db8a
prefer32 needs to be set differently for app and native modules.
Make it use lambda provided by the module type instead of trying
to make archMutator figure it out.
Test: m checkbuild
Change-Id: Ibf8af35fdd3e1721725539d1f5452f4439d2125c
Use a variable for module.base() and separate the full list of
targets for the class from the filtered list of targets for the
module.
Test: m checkbuild
Change-Id: I32edd26bcbaff0baf433d0581f253b8a5c8b81f3
Make java_import, android_library and android_library_import modules
defaultable, and add some missing property structs to java_defaults.
Also use InitJavaModule everywhere.
Test: m checkbuild
Change-Id: I981875bbec5d1f87bae73f36a385e0b77c56c368
Bug: http://b/109759970
r335286 fixed this for 64-bit Windows. This remains broken for 32-bit
Windows.
Test: m native-host-cross
Change-Id: I4d11b83d51db055406c6723aad89efd1b86b9b7b
When sdk_version is equal to or less than 28, java_version is implied to
1.8 unless it is explicitly specified. It was 26 originally, and is now
updated to 28 which is the most recent API version.
Bug: 117069453
Test: EXPERIMENTAL_USE_OPENJDK9=true make apex_aidl_interface
Change-Id: I18ebe3ec51ada54103f6d203b9bfae8ec6ea5d90
Setting targetSdkVersion to '1' causes ManifestMerger to add
implicit permissions when merging to a higher targetSdkVersion.
It should really be unset, but ManifestMerger treats unset
targetSdkVersion as 'Q' if minSdkVersion is 'Q' (but not if
minSdkVersion is '28'). Set it to something low so that it
will be overriden by the main manifest, but high enough that
it doesn't cause implicit permissions grants.
Bug: 115415671
Bug: 117122200
Test: m checkbuild
Change-Id: I1d2d031a21314f6b55d8ea1cc7c4c8e3ecae7f06
compdb.go was incorrectly splitting compiler arguments. This could
cause empty strings to be included in compile_commands.json,
potentially confusing tools.
Bug: 117124308
Test: m checkbuild
Test: m SOONG_GEN_COMPDB=1 \
SOONG_GEN_COMPDB_DEBUG=1 \
SOONG_LINK_COMPDB_TO=$ANDROID_BUILD_TOP nothing
Examine $ANDROID_BUILD_TOP/compile_commands.json
Change-Id: I375baf255c50a1329cd644ac584d200aba9daa69
The double space is confusing compdb.go. compdb.go should be fixed,
but the double space is unnecessary so remove it.
Also make -isystem consistently followed by a space.
Bug: 117124308
Test: m checkbuild
Change-Id: I5ce7530d2ef66be8d8285e252d60a39299984a06
* changes:
Use file glob for zipping classes*.dex files
soong_zip: add --ignore_missing_files flag
Add a --symlinks argument to soong_zip
soong_zip: support globs in -f and -D arguments
* Add relativeFilePathFlag compiler flag category whose format must be
a=b/c/d, where "a" is the flag name and "b/c/d" is the flag value that
is a relative path
* During CMake conversion, "b/c/d" will be prefixed with ANDROID_ROOT
before being written to CMakeLists.txt
Bug: 112599284
Test: cd system/bt; mm -j40
croot; make -j40
Change-Id: I4bb039678482201db60f0af97ab280d5028e0463
Make the cleanspec & Android.mk runs of kati use much of the same code
and arguments.
Also renames 'Kati' to 'KatiBuild' in many cases, in the the expectation
that we'll have a 'KatiPackage' step in the future.
Use --no_ninja_prelude and move local_pool & _kati_always_build_ into
the combined ninja file. This will reduce the need to re-read makefiles
when Goma is enabled, and it allows us to include more than one
Kati-generated ninja file in the build graph.
Bug: 116968624
Test: build_test on downstream branches
Change-Id: Ibdac689b81f62dc293647fad917d84946f2c3cfa
Using a directory to zip the output of d8 and r8 was causing the
empty classes.dex.jar file in the output directory to get
zipped in. Use a file glob instead so that it only picks up the
classes*.dex files even if classes.dex.jar already exists.
Test: m checkbuild
Change-Id: I1f3c1cbc23439fc52756aad027ab8cac10213d1d
soong_zip builds a list of files to zip early and then starts
zipping them all. If a directory being zipped is concurrently
modified, a file that existed when soong_zip started may not
still exist. Add a flag that continues when an expected file
does not exist. Print a warning, since this should be rare
in normal usages but is a sign of a problem if it happens
regularly.
Test: zip_test.go
Test: m checkbuild
Test: m platform
Change-Id: I78426fe66fded8528ddd436c0f71a7442183cfeb
Add a --symlinks argument that defaults to true to soong_zip.
Passing --symlinks=false will cause it to follow symlinks instead
of storing them in the zip file.
Relands I4deb98daa9d4ba9f94e3d7670c117fe00381d2ba with tests.
Bug: 112843624
Test: glob_test.go
Test: zip_test.go
Test: m checkbuild
Change-Id: I0eff9c1f2dba79e873fda381ff585df55d5aaaad
-f and -D arguments can now take globs in the Soong format.
Also update the use of soong_zip that jars resources to escape the
globs in the arguments, and then shell-escape them when writing to
the rsp file so the glob escape are not intepreted by ReadRespFile.
Also remove an unused argument to the buildAAR rule that could
have contained values that needed escaping.
Relands I7f20bb169dc01f952d2a7681ec6ee9c05737ed37 with a fix for
trailing "\n" in list files, which causes a call to pathtools.Glob("")
that returns "./", which could then get incorrectly translated to
"../../../" in the zip file. Also adds tests.
Test: m checkbuild
Test: zip_test.go
Change-Id: I54b8eef9231875e6042a32c9f8bcc5c2f779922a
