This reverts commit 27f601f5d9.
Reason for revert: Droidmonitor triggered revert due to build breakage in b/348090986. Will be verifying through ABTD before submission.
Change-Id: I55e0880741e543309a7b991d49e59e4fa1f62e53
Many places in the codebase have checks like this:
if shipping level > X: do cool thing Y
This is great because it reduces the cost to upgrade and test
old Android devices or refactor their code. However, many
targets, such as the NDK, the SDK, mainline modules, and so on
do not set a shipping level, so they don't get to do the cool
thing Y.
In order to resolve this, we could modify every check to check
the default case. However, this is an invasive change, and it
is not maintainable. Instead, consider non-shipping products to
always be in the future.
In general, Android features are required to be backwards
compatible, so this should always work for things like mainline.
Also note, this means when someone adds a new feature Y like this,
they'll clearly see the impact of it being added everywhere, rather
than on such a small selection of newly shipping devices. This
avoids a risk where a change needs to be run on mainline modules (or
other targets) but is not tested in this configuration from the
start.
Future work: many `ifndef` checks can be cleaned up after this,
or if this approach makes sense, they can be cleaned up in this
CL.
Bug: 339026799
Bug: 279808973
Bug: 333908433
Test: builds of:
- errorprone-trunk_staging
- mainline_modules_sdks-trunk_staging-userdebug
- mainline_modules_x86_64-trunk_staging-userdebug
- ndk
- sdk-trunk_staging-userdebug
Test: too large number gives:
build/make/core/product_config.mk:598: error: integer greater than 10001 is not supported!.
Change-Id: I17c34267f774ea8b9265e1d798a67af7838715c5
zucchini failure in OTA is observed. To unblock, disable
zucchini first. After the investigation are done and determined
zucchini is safe to enable, we will turn it back on.
Test: th
Bug: 342523982
Change-Id: I79c245356a356389328c788b3b7977b224e1df6a
Go 1.22 is more strict about GOPROXY=off, so we need to update the
go.work file to point at local copies of dependencies.
Bug: 330574836
Test: Presubmits
Change-Id: I37dbc821f138fe8977636316ebe3e7986ccdfa0e
This was copied over from when Soong required us to define it in the
properties file, but this is the default and we also don't care.
Bug: None
Test: None
Change-Id: I35311093c83fddcb7c413535fbb50a126b615036
Update storage read api to not find storage file location from a pb
file, instead directly read from /metadata copy. Previously for
package.map and flag.map, we are reading from the respective RO
partition. Now we are reading from /metadata/maps dir. This has a few
advantages:
1, early flag availability, since /metadata can be mounted much earlier
than mainline modules, so it would make mainline flags availabile even
before mainline modules are mounted.
2, we no longer need to read from a pb file to find where package.map
and flag.map are. Thus the read api can be further simplified and
downsized. With this change, we are able to shrink the cc flag read api
lib size from 171k to 120k.
Bug: b/321077378
Test atest -c
Change-Id: Ic9086fe4c49c139a4d1c66a8d472023a88c9dd17
When parsing API signature files, check-flagged-apis relies on
ClassItem.superClass to get the parent class of a class or interface.
That method always returns null for interfaces.
When generating api-versions.xml, metalava marks interface classes as
inheriting from java/lang/Object:
<class name="android/os/Parcelable" since="1">
<extends name="java/lang/Object"/>
[...]
</class>
This confuses check-flagged-apis when comparing data parsed from both
sources, as the symbol signatures will be identical, but the superclass
entries differ. Work around this by explicitly marking all interfaces
as inheriting from java/lang/Object when parsing API signature files.
Bug: 334870672
Test: atest --host check-flagged-apis-test
Change-Id: Icbb8f7d4c3d4232a083289a778b347e33a0856ab
Currently, both stdout and stderr are redirected to a file. We want
stderr to be visible on the terminal in case the release config
fails.
Test: m nothing
Change-Id: I1337718fe8f9394be3017cc8b6d76f8dc1a94ef0
Fix the case when call a tool with run_tool_with_logging with a single
arg start with -- (e.g. adb --help), make sure when we pass that arg
to the logger, it treat it as the value for the --tool_args option
instead of a separate option
Test: atest run_tool_with_logging_test
Test: manually run source build/envsetup.sh and run adb --help and check
the event log is sent to clearcut. Tested with both bash and zsh
Bug: 341382247
Change-Id: I1e09907f267b453cb62876e171064daa021e3d91
This is meant for platinum tests that require device
signing. They will be migrated out of device-tests.zip
Test: presubmit
Bug: 341719739
Change-Id: Ie30c4bedc2edf4f537deb49099d0e23b4eec2626
Certain applications, like tracking the dependency chain of modules,
knowing the "required" dependencies is required.
In this context, "required" here means the `required` field in
Android.bp files.
Bug: 339193439
Test: Inspect module-info.json for the `required` field
Change-Id: I09977bf71a0a0ae46ea9268ce1faa5c90673bb7a
Apprently target-specific variables are only valid for the body of
the target, not its dependency list, so the dependency on the zip file
wasn't being added.
Bug: 340392832
Test: m, rm file_list.txt and the zip file, m again, and ensure we don't get an error that the zip file doesn't exist
Change-Id: Idaca6606ff24055aa98c78ca1206b587cc1df4d2
