These tests fail on sdk-sdk_mac-trunk_staging because this feature is
not intended to run outside of Linux. Let's skip these tests for now
until we figure out how to disable snapshot feature properly.
Bug: 339782737
Test: go test
Change-Id: Ibe70a0e3dc92262c78ce9e3a8af58f2da3bfeda2
android.ModuleBase already calls aconfigUpdateAndroidBuildActions()
that is the same with CollectDependencyAconfigFiles(). Remove the
CollectDependencyAconfigFiles() to avoid duplication with
aconfigUpdateAndroidBuildActions().
To make the aconfig information available in
GenerateAndroidBuildActions() of all modules, call
aconfigUpdateAndroidBuildActions() before calling
GenerateAndroidBuildActions() of each module.
Also, we don't need SetAconfigFileMkEntries(), which is a duplicate
of aconfigUpdateAndroidMkData()
Bug: 335363964
Test: diff `adb shell printflags` before and after the change.
Change-Id: I52808e442e9fed7db1eae7b7c5ed0b1c5ba74f5d
Background: if a lib has `sdk_version` set, it is mutated into two
variants: platform and sdk. The latter is for the inclusion into
unbundled APKs which can be installed on older platforms. The former is
what is installed to the platform.
So far, for unbundled app builds, (1) the platform variant was marked as
uninstallable, (2) while the sdk variant was left installable. (1) is
causing a problem (b/339262059) when an APEX containing a filesystem
module is built as unbundled. The filesystem skips installing the
platform variant. (2) is not causing a problem, but is unnecessary
because the sdk variant is not something to be installed (on the
platform). It is built into the APK via the jni_libs property. It's not
considered an installation.
This change fixes (1) and (2). For (1), the platform variant is always
marked installable even for unbundled build. For (2), the sdk variant is
always marked non-installable.. because, again, it's not designed to be
installed to the platform.
Bug: 339262059
Test: banchan com.android.virt aosp_arm64
UNBUNDLED_BUILD_SDKS_FROM_SOURCE=true m apps_only dist
check libc++.so, libcrypto.so are in the microdroid image in the APEX.
Change-Id: I8999b724926cd9ab6d56796133e218fefcc1910b
llndk.libraries.txt build fails from clang tools build, as there is no
valid target path for llndk.libraries.txt from the build. This change
skips llndk.libraries.txt build if target output path is empty.
Bug: 330100430
Test: prebuilts/clang-tools/build-prebuilts.sh passed
Change-Id: I867dc90dbf98e3d8c24b191a744686286740b230
This is necessary for cc_cmake_snapshot to snapshot and generate build
rules for actual source files (AIDL files) and not generated cpp
sources. Otherwise, this information is inaccessible for cc module after
aidl_interface pass.
Bug: 302718225
Test: Build aosp/2798713
Change-Id: I991e73f74d7551d21f337a7161cce55ba51fec71
To handle static inline functions, bindgen generates a C file which is
expected to be compiled and linked into dependents on the generated
bindgen source.
This CL adds support for cc modules ingesting this output and for
bindgen to produce it (and link it against the bindgen rust_library
variant as well).
Bug: 290347127
Test: m blueprint_tests
Test: Example module with static inline functions builds locally
Change-Id: I167a8356eb6e0059fc21169fd3bfc6bf4d9c812f
The sanitizer runtime logic was rewritten in aosp/2438959 to simplify
static sanitizer library selection. However, enabling ubsan diagnostics
for host tests would result in incorrectly adding an unnecessary
libclang_rt.asan.so dependency.
This check should have been removed as well -- the library dep
determination for c.sanitize.Properties.UbsanRuntimeDep is now handled
higher up in the logic.
(Some formatting fixes are included as well)
Bug: 323237381
Test: UBSanitized Host test with diagnostics runs correctly.
Change-Id: I97bd01344b7b676fccb9f250c0f801b8bc212a1a
NDK team requested not to generate the dumps by header ABI checker in
the finalization process.
Test: m out/target/product/generic_arm64/lsdump_paths.txt
Bug: 332959609
Change-Id: Ia04c6f266088968ff731c0fabbbfce26300109ae
This allows using select statements with it.
Bug: 323382414
Test: m nothing --no-skip-soong-tests
Change-Id: I6f3efaaa3d82505e38a91ee4ba0e18e404360191
Merged-In: If355d24506e3f117d27b21442a6c02bca3402dc7
This was already done for target.recovery, but was forgotten for
target.vendor and target.product.
Test: add exclude_header_libs to a target with export_header_lib_headers and build successfully
Tags:
Change-Id: Ib206f54b2176830e32af33b9b093b71d4f1351ac
Signed-off-by: Abhishek Gadewar <abhishekgadewar@meta.com>
Otherwise we cannot disable memtag_stack on first stage init, because it
statically links in a lot of libraries.
This was benign, but if we want to add a stack history buffer, we do not
want to emit these instructions for first stage init.
Bug: 309446520
Change-Id: Ifbdcbe5ba06eb46970bc1e9b64a26c1ca59ba5f2
Currently VNDK contains some logic for LLNDK libraries as they are
treated in a similar way with VNDK. However, those logics should stay from
VNDK deprecation. To keep the logic, this change moves LLNDK related
logic into llndk_library.
Bug: 330100430
Test: Soong tests passed
Test: llndk.libraries.txt did not change from CF build
Change-Id: I1d02a3c2a398f1b1060b4f2bdd23af32310503bb
Merged logtags files will be used for Soong built filesystem images.
Bug: 336189540
Test: m out/soong/.intermediates/all-event-log-tags.txt
Test: m out/target/common/obj/all-event-log-tags.txt
Change-Id: Ib590c2bc8073e9acee6b45ef08092768237cf9d3
This commit adds --include-symbol-tag and --symbol-tag-policy to
header-abi-linker commands. The arguments make header-abi-linker load
the symbols tagged with llndk/apex/systemapi from the version script.
It filters the function declarations by the symbols. The output dumps
will not contain unfinalized ABI.
Test: make findlsdumps
Bug: 333532038
Change-Id: I28234a3749f389b8a5a09ac84341b1fcd1ee88a6
This commit adds a rule that builds APEX ABI dumps separately from the
implementation libraries' ABI dumps. The rule takes the
export_include_dirs and the symbol_file of the stubs as parameters.
The dump paths are tagged with "APEX" in lsdump_paths.txt. The script
updating the prebuilt reference dumps can differentiate the APEX dumps
from the opt-in dumps tagged with "PLATFORM".
This commit also adds an ABI diff rule. Soong compares the APEX ABI
dumps with the reference dumps in version 35. It compares the
implementation dumps with the reference dumps in old versions.
Bug: 333532038
Test: make
Change-Id: I76902a8e3b7d0e96a5ad756f493924371cd7ad3c
This reverts commit 321dd23b34.
Breakage is fixed by https://r.android.com/3051002.
Change-Id: I70925c7aa01fb78ec701971ee97192993aef0edd
Test: abtd
Bug: 322868540
When testing test-only validations, I noticed that some cc_test modules
where not reporting that they were test-only.
This happened because test-per-src variations were short-circuiting
writing out the provider.
The `all_teams` target is using the first variation for a module, but
it possibly should see if any variation is test-only.
For now, just making this fix to cc.
Test: m all_teams && gqui from "flatten(out/soong/ownership/all_teams.pb, teams)" proto build/soong/android/team_proto/team.proto:AllTeams ' where teams.kind = "cc_test" and teams.target_name="libnativebridge-tests"'
Test: go test ./cc
Change-Id: I6d44a521f5f2457527049399509d979559d7dc17
The __BIONIC_NO_PAGE_SIZE_MACRO flag has to be passed to
the build system because there are C/C++ libraries using
__BIONIC_NO_PAGE_SIZE_MACRO to decide whether to use getpagesize()
or PAGE_SIZE macro.
Bug: 333973679
Test: Build manually
Change-Id: Ie5d71f4b2b256ab8f429ed4724b8d249b364d5da
As part of aosp/3022586 where we added the idea of "test-only" modules
and top_level_test_targets, this CL implements that for cc_ modules.
We let users set "test-only" on cc_library, but not on other modules
where the module kind is implicitly test-only, like cc_test.
Here the implementation, not the user decides it is test-only.
% gqui from "flatten(~/aosp-main-with-phones/out/soong/ownership/all_teams.pb, teams)" proto team.proto:AllTeams 'select teams.kind, count(*) where teams.top_level_target = true group by teams.kind' aosp_shiba[6:15:47]/0
+--------------+----------+
| teams.kind | count(*) |
+--------------+----------+
| art_cc_test | 56 |
| cc_benchmark | 68 |
| cc_fuzz | 515 |
| cc_test | 3518 |
| cc_test_host | 6 |
+--------------+----------+
% gqui from "flatten(~/aosp-main-with-phones/out/soong/ownership/all_teams.pb, teams)" proto team.proto:AllTeams 'select teams.kind, count(*) where teams.test_only = true group by teams.kind' aosp_shiba[6:16:26]/0
+--------------------------+----------+
| teams.kind | count(*) |
+--------------------------+----------+
| art_cc_test | 56 |
| art_cc_test_library | 13 |
| cc_benchmark | 68 |
| cc_fuzz | 515 |
| cc_test | 3518 |
| cc_test_host | 6 |
| cc_test_library | 484 |
+--------------------------+----------+
Bug: b/327280661
Test: m nothing --no-skip-soong-tests
Test: go test ./cc
Test: m all_teams
Change-Id: I344436c424a9dfbdcf27e10f42f5cebc3d2b1261
Move -Wno-error=format from noOverrideGlobalCflags to
commonGlobalCflags so that projects can re-enable it with
"-Werror=format"
Bug: 315250603
Test: Add -Werror=format to bluetooth_cflags
Change-Id: I677cee5373c670c1a3b0eceeea1b7c35d28a17c2
