To generate ld.config.txt dynamically(b/123722631), each APEX should
provide some dependency information:
a) list of libraries which other APEXes(or system) can use from this apex
b) list of libraries which this apex uses from other APEXes(or system)
This change puts dependency information in apex_manifest.json at
build-time with two additional keys:
a) provideNativeLibs
b) requireNativeLibs
Bug: 138695532
Test: m (runs soong tests)
Test: find $OUT/apex -name apex_manifest.json -exec cat {} \;
(shows contents of apex_manifest.json files)
Change-Id: Iaad12c8c35454222ad177ce923cce76ef12a8a5a
Before, the values were often simply ignored. If bad values are provided
now, we should be able to catch them.
Bug: 119771576
Test: relying on this for AIDL/HIDL
Change-Id: I8d2ff26da0b2d01ebe7f78c26d69c7b618a65367
For smart terminals, default to using the status table, using 1/4 of the
terminal height (with a min of 1 and a max of 10).
This behavior can still be overriden to a specific height with
SOONG_UI_TABLE_HEIGHT, and turned off by setting SOONG_UI_TABLE_HEIGHT
to 0.
Test: m <adjust terminal height, see it shrink when it gets too short>
Test: SOONG_UI_TABLE_HEIGHT=20 m
Test: SOONG_UI_TABLE_HEIGHT=0 m
Change-Id: I224348a29a6e07f168c92cf5514a94bd27e32618
To build the platform for ASAN, we do
`m && SANITIZE_TARGET='addresss' m`
However, at the end of the second build, the system partition could have
conflicting APEXes; prebuilt APEXes from the first build and
source-built APEXes from the second build. Since the file names for the
prebuilt and the source-built are different (e.g.
com.google.android.media.apex v.s. com.android.media.apex), we end up
having two files for the same APEX. This is confusing apexd at runtime
and the device fails to boot.
To fix this, when building a non-prebuilt APEX, the prebuilt APEX might
have been installed by the previous build is deleted.
Bug: 138146044
Test: lunch aosp_cf_x86_pasan; m && SANITIZE_TARGET='address' m
check that out/target/product/vsoc_x86/system/apex has
com.android.*.apex only.
Change-Id: Ib5a021a297cf0173ea5a3b50e9398b1cf295c558
The code still needs to exist, but for buildspec.mk files, not the
PRODUCT-* arguments.
Test: none
Change-Id: I60c6044ab2d1627af7e6e2e8831b77d9db12aa55
If sort_bss_symbols_by_size is true, a shared library is built twice.
The first build generates an unsorted output file, which is used to
generate the symbol ordering file. The output of the second build is
a shared library with its bss symbols sorted by their size.
With this, the only user of symbol_ordering_file, libc, is migrated to
use the new property, so we remove symbol_ordering_file support as well.
Bug: 135754984
Test: Build and check the resulting libc.so has its bss symbols sorted.
Change-Id: I5c892b44d82eb99cbc070cfa2c680be3087f3364
In `apex.apexBundle.GenerateAndroidBuildActions`, we used to pass the
"all tests" ("") module as `module` for all `apexFile` objects created
from a test module using `test_per_src: true`. An immediate issue of
this situation was that the "" module is hidden from Make, which made
all the generated `apexFile` objects hidden from Make too. This would
break the construction of flattened APEXes, as they rely on Make logic
to install their files.
Instead of collecting `test_per_src` test variations' output files in
`cc.Module.GenerateAndroidBuildActions` and using them in
`apex.apexBundle.GenerateAndroidBuildActions` as part of handling the
"" variation as a direct dependency of an `apexBundle`, process them
as indirect dependencies (and do nothing for the "" variation direct
dependency).
Adjust the indirect dependency logic in
`apex.apexBundle.GenerateAndroidBuildActions` to allow not only
shared/runtime native libraries as indirect dependencies of an
`apexBundle`, but also `test_per_src` tests.
Test: m (`apex/apex_test.go` amended)
Bug: 129534335
Change-Id: I845e0f0dd3a98d61d0b7118c5eaf61f3e5335724
As sdk_check.mk checks, soong starts to check every app.
If sdk_version is empty, platform_apis must be true.
If sdk_version is not empty, platform_apis must be false.
Test: soong test
Test: m
Bug: 132780927
Change-Id: I7ba702d616404d155f8ac40cd008828663ad1488
Reduce boot image space taken from 28803072B -> 9572352B (-20MB).
No reason to have this be configurable yet.
Test: ls -l $OUT/system/framework/*/apex*.art | awk '{s += $5} END {print s}'
Bug: 77863440
Change-Id: I280b740dfbd001109ce1e288e395595f21567aa6
Use Kythe (https://kythe.io) to build cross reference for the Android
source code. ~generate the input for it during the build. This is done
on demand: if XREF_CORPUS environment variable is set, build emits a
Ninja rule to generate Kythe input for each compilation rule. It
also emits two consolidation rules (`xref_cxx` and `xref_java`),
that depend on all Kythe input generation rules for C++ and Java.
The value of the XREF_CORPUS environment variable is recorded in the
generated files and thus passed to Kythe. For the AOSP master branch it is
`android.googlesource.com/platform/superproject`, so the command to build
all input for Kythe on that branch is:
```
XREF_CORPUS=android.googlesource.com/platform/superproject m xref_cxx xref_java
```
Each Kythe input generation rule generates a single file with .kzip
extension. Individual .kzip files have a lot of common information, so
there will be a post-build consolidation step run to combine them.
The consolidated .kzip file is then passed to Kythe backend.
The tools to generate .kzip files are provided by Kythe (it calls them
'extractors'). We are going to build them in toolbuilding branches
(clang-tools and build-tools) and check them in as binaries into master
and other PDK branches:
For C++, `prebuilts/clang-tools/linux-x86/bin/cxx_extractor`
for Java, `prebuilts/build-tools/common/framework/javac_extractor.jar`
Bug: 121267023
Test: 1) When XREF_CORPUS is set, build generates Ninja rules to create
.kzip files; 2) When XREF_CORPUS is set, building
`xref_cxx`/`xref_java` creates .kzip files; 3) Unless XREF_CORPUS is
set, build generates the same Ninja rules as before
Change-Id: If957b35d7abc82dbfbb3665980e7c34afe7c789e
* changes:
Add defaults_visibility support
Refactor visibility to support visibility on defaults modules
Add DefaultsModule interface
Improve documentation of defaults mechanism
Clang derives the value of __ANDROID_API__ from the triple these days. In a
future version of clang I plan to start making the behaviour of the HWASAN pass
dependent on the API level in the triple, so it's going to need to be accurate.
Test: walleye-userdebug boots
Change-Id: Ie5e36b5c8f6dcda084cc12b1160abbdf94765174
mm is now doing the same thing as mma. Now print a deprecation message
for other tools that may be setting ONE_SHOT_MAKEFILE directly (or users
that are still using a very old envsetup.sh).
Test: mm # does not print
Test: ONE_SHOT_MAKEFILE=bionic m
Change-Id: I98221657024d824a61c0d721d5c7dd1a2ea2fb1e
Existing modules, either general one or package ones have a single
visibility property, called visibility in general, and
default_visibility on package, that controls access to that module, or
in the case of package sets the default visibility of all modules in
that package. The property is checked and gathered during the similarly
named phases of visibility processing.
The defaults module will be different as it will have two properties.
The first, visibility, will not affect the visibility of the module, it
only affects the visibility of modules that 'extend' the defaults. So,
it will need checking but not parsing. The second property,
defaults_visibility, will affect the visibility of the module and so
will need both checking and parsing.
The current implementation does not handle those cases because:
1) It does not differentiate between the property that affects the
module and those that do not. It checks and gathers all of them with
the last property gathered overriding the rules for the previous
properties.
2) It relies on overriding methods in MethodBase in order to change the
default behavior for the package module. That works because
packageModule embeds ModuleBase but will not work for
DefaultsModuleBase as it does not embed ModuleBase and instead is
embedded alongside it so attempting to override a method in
MethodBase leads to ambiguity.
This change addresses the issues as follows:
1) It adds a new visibility() []string method to get access to the
primary visibility rules, i.e. the ones that affect the module.
2) It adds two fields, 'visibilityPropertyInfo []visibilityProperty'
to provide information about all the properties that need checking,
and 'primaryVisibilityProperty visibilityProperty' to specify the
property that affects the module.
The PackageFactory() and InitAndroidModule(Module) functions are
modified to initialize the fields. The override of the
visibilityProperties() method for packageModule is removed and the
default implementations of visibilityProperties() and visibility()
on ModuleBase return information from the two new fields.
The InitDefaultsModule is updated to also initialize the two new
fields. It uses nil for primaryVisibilityProperty for now but that
will be changed to return defaults_visibility. It also uses the
commonProperties structure created for the defaults directly instead
of having to search for it through properties().
Changed the visibilityProperty to take a pointer to the property that
can be used to retrieve the value rather than a lambda function.
Bug: 130796911
Test: m nothing
Change-Id: Icadd470a5f692a48ec61de02bf3dfde3e2eea2ef
Changes the InitDefaultsModule method from taking a DefaultableModule
to taking a DefaultsModule in preparation for adding visibility support
to defaults modules.
Bug: 130796911
Test: m nothing
Change-Id: I1d459da1017ae7f2654e7eb275cb424e52d85730
Also, removes unnecessary cast from DefaultableModule to Defaultable.
Bug: 130796911
Test: m nothing
Change-Id: I01d8f5186927215a1aa6b7431558041f427d7381
fuzzers require linking to libclang_rt which does not support
linux_bionic. Therefore always disable this target.
Test: ./art/tools/build_linux_bionic_tests.sh
Bug: 138307504
Bug: 118058804
Change-Id: I12816f302d32d5ee846f90c62814744ae35c49bb
