This fixes two main sources of nondeterminism:
1. Fix a bug in the ConfigurationAxis comparator (which caused
ConfigurationAxis sorting to be nondeterministic)
2. Process C++ dependencies using the sorted ConfigurationAxis order. In
theory, the order in which dependencies are processed shouldn't
matter (as they should end up in different select stanzas). However,
in the case of InApex stubs, this is not the case; we now ensure
that lists are concatenated in a predictable order.
Added bonus: Some cleanup with SortConfigurationAxes which
makes use of go generics (this made it easier to debug this issue).
Will follow-up with regression tests.
Test: Manually verified that build.ninja checksum and BUILD.bazel checksums do not change after running `m nothing` 6 times in AOSP (with comment-only Android.bp changes in between each run)
Change-Id: I81168e45bdbbcd61ea95ff665cf6c4bc180aa4e0
Disable VNDK apexes when their version is below the minimum
supported API level for the primary architecture.
Test: TestVndkApexCurrent
Change-Id: Ie72a5beb9da432660d3fec5c520402224326f961
This piggybacks onto the ApexInfo cquery handler, so we're issuing a
single bazel query call that reads two providers in the starlark expr.
Also rename requiredDeps to makeModulesToInstall to differentiate it from
APEX's required/provided libs in the apex manifest.
Test: unit test
Test: mkdiff
Fixes: 263123189
Change-Id: Ib7e43f1586f29864eee8627dba3631bfaff27afa
There is no prebuilt libc++ for riscv64 provided by the NDK, build
it using the platform sources instead.
Test: lunch aosp_riscv64-userdebug && m -k
Change-Id: Ib9c99381e61331e10c136127a3f1b1626ce75ebe
Prevent depending on CRT variations that don't exist by raising
the CRT dependency api level to the minimum supported for the
architecture.
Test: lunch aosp_riscv64-userdebug && m -k
Change-Id: I575355569b3772f5d1fe2530161a1d45aa00a349
This API surface is provided by platform and mainline modules to other
mainline modules. Rename it to module-lib API surface to align it with
the terminology in go/android-api-types
Test: go test ./bp2build
Test: go test ./cc
Change-Id: Ieb9f3214e66366fc2ceb6f47e2d6623acb537827
To match the scudo allocator's always return zero allocation behavior,
set the max malloc fill size larger than the largest size class.
Currently, the largest is 65552, so set the max to 128KB to avoid
needing to change this in the future.
Test: Run test that verifies allocations returned are zero and it passes.
Change-Id: Ie9f0b0d54f846008b1c6d5dc6e55c7031262d29e
Bug: b/261642850
Test: Build and check warnings. Add two xfail tests in bionic and see the
results locally.
Change-Id: I61be649f935c05461bdd6c260627f3c72261a9e8
The adbd_test androidmk definition is missing some cc-related variables
for cc_test in mixed builds. These variables should be populated from
information from Bazel.
Bug: 265758350
Change-Id: I59d017e2eb2f139188ba3383c457cc0055372b61
Bug: b/261642850
Test: Build and check warnings. Add two xfail tests in bionic and see the
results locally.
Change-Id: I68fca0084787c329b6c49ce4dff6fd132f820735
Currently, Bazel sets __ANDROID_APEX_MIN_SDK_VERSION__ based on the cc library's min_sdk_version while Soong does based on apex's min_sdk_version. This causes a diff in the clang commands.
Digging further, we realized __ANDROID_APEX_MIN_SDK_VERSION__ isn't used anywhere so we might just delete it altogether instead of fixing it.
Bug: 265134933
Test: presubmit
Change-Id: I08e89dc8f6ef86669248ea5c47de6603e9d2ffbb
Currently we document the semantics of these two symbols. I would like
to add some documentation about why this granularity is necessary. The
majority of the details are contained in b/191371676
Test: N/A (documentation change)
Bug: 265485106
Change-Id: I752c921eabceff94f92f24a46c9a16dc28a2af19
This code block exists for LLDNK code path (m.UseVndk()...), but was
missing from NDK code path (m.IsSdkVariant())
Test: TH
Change-Id: Ib156d14a6f013c88499a0641dbb547c632357362
Most of the complexity in cflag_artifacts can be avoided
by using WriteFileRule.
Bug: 262629589
Test: m out/soong/cflags/module_cflags-Wall.txt
Change-Id: I5b0a678c24609f7b187c9fb8e5fcae7f89adf59c
* Remove `Bp2buildCcLibraryStaticOnlyList` it is never set
* Remove the rest of the sutff related to it
* Streamline MixedBuildsEnabled
* Rename BazelContext.BazelAllowlisted to BazelContext.NameFilter
to reflect its action
* Rename bazelContext to mixedBuildBazelContext
* Fix minor warnings from Go static analyzer
Test: treehugger
Change-Id: Ie245c992f94bbfd5c7c23dd06917026200a28263
Rust and cc binaries currently use $ORIGIN/lib64:$ORIGIN/../lib64 as the
rpath, and cc tests add $ORIGIN/../../lib64:$ORIGIN/../../../$ORIGIN:$ORIGIN.
This causes problems when a binary is included as test data in
out/host/linux-x86/testcases/<test dir>/<CPU>/<test>, as the
binaries can't find the libraries in out/host/linux-x86/lib64.
Use the same rpath for test and binaries, and for cc and rust.
Bug: 264604160
Test: m USE_HOST_MUSL=true out/host/linux-x86/testcases/acpi_tables_test_src_lib/x86_64/acpi_tables_test_src_lib && out/host/linux-x86/testcases/acpi_tables_test_src_lib/x86_64/acpi_tables_test_src_lib
Test: m USE_HOST_MUSL=true out/host/linux-x86/testcases/gen_sdk_test/x86_64/gen_sdk_test && out/host/linux-x86/testcases/gen_sdk_test/x86_64/toybox
Change-Id: I10fe5dc0de01d1f3c6aea8dbabbf60edab5989c3
