A container is software which is always built in its entirety using the
same build environment. In particular, all of its parts are built using
the same build-time default flag values. In addition, containers are
always installed as a single unit.
Bug: 312696545
Test: atest aconfig.test && m all_aconfig_declarations
Change-Id: I2ef3db836c4456f4f4fb5c066edf9094e38f89cc
This variable is always false or empty now, so clean up remaining
references to the variable.
Bug: 241346584
Test: Presubmit
Change-Id: I5dd3d356729d0bd2639633580bbb85f8b03582cf
Add a flag to aconfig dump that will allow identical flags to be merged
without erroring. This will allow merging the aconfig cache files from
dependencies for each module, which requires passing less data to Make
from Soong, and thus reduces the percentage of builds require Kati
analysis.
Bug: 313698230
Test: aconfig.test
Change-Id: Id2d9b78225720ce01defad4b13ea9395f6ab1033
This commit adds a third codegen mode, _exported_, in addition to
the existing modes, production and test.
When codegen mode is _exported_, getters are generated _only_ for
flags marked as exported as well. Also the getters always look
up DeviceConfig values at runtime, and have a default value of
false.
This only implements exported mode for Java codegen, follow-up CLs
will support Rust and C++.
Test: atest aconfig.test
Bug: 311152507
Change-Id: Ie39379b40de072180e05d84c76361b24cc0e0d83
Expose this option at `merge_builds.py` so that this script
can also resolve the conflict AVB rollback index location
while building the vbmeta image.
Bug: 289368551
Test: TreeHugger
Change-Id: I404160ff5d7591cb6a81ed2d2267de1b0e6ac65f
It was a temp folder created for check_target_files_vintf.py.
But now it is wrongly archived by create_target_files_archive
in merge_target_files.py.
And then such package will cause CheckVintf exception in
ota_from_target_files.py because the APEX folder is created
again but already exists.
So use MakeTempDir to make it removed after check finished.
Test: merge_target_files & ota_from_target_files
Change-Id: Iab54afff2c346a582d5ee5228103fe091e016cd0
We no longer build GKIs from the platform tree.
These build commands were neither used nor maintained anymore, so clean
them up to reduce maintenance effort.
Bug: 229701033
Fix: 229701033
Test: presubmit
Change-Id: Ie882fccd864920289e48366e99a4ebd67e784d0d
F2FS has been expanded to allow for different block sizes. Currently the
block size must be the same as page size, but this may change in the
future. This gives support for 16K while not restricting the ability to
choose a block size that isn't based on the page size later on.
Bug: 279820706
Test: Boot 16k F2FS device
Change-Id: I4e025d83a5e5ad03d89bc80b02fea390c2cbad69
When a flag is exported, it can be read by code built in other containers. By
default flags are not exported. Set this to true if your flag is gating
a public API which may be called from other containers (typically, a
flag used in @FlaggedAPI annotations).
Test: atest aconfig.test
Bug: 311152500
Change-Id: I53e3ed18b96c1518f04172d5933ef96b41ccda7e
provider_ object relies on cache_ vector to be alive, but the order
of destruction between these objects at exit are not guaranteed.
This can lead to crash at the exit especially on ASAN build.
By putting cache_ as a member it is guaranteed that cache_ is not
destructed until provider_ is destructed.
Test: m ; m AconfigDemoActivity
Test: atest aconfig.test
Test: observe no crash on _hwasan build
Change-Id: If9e23e99c501bf8f06fcab003622948b9e730352
The flag `disabled_rw_2` is used to verify that aconfig can
auto-generate code for flags in different namespaces. Rename the flag to
`disabled_rw_in_other_namespace` to highlight that it belongs to a
different namespace than the other flags.
This CL is a semantic change only.
Bug: None
Test: atest aconfig.test
Change-Id: Ib82fdce0ca3f7cd56b9b3f80e44fe424a03a2ae4
This reverts commit efda207c09.
Reason for revert: this unrevert contains the fix for b/311187402
Test: cargo test
Change-Id: I210aebd30edd864a7c141ede336c12aebf4f1fcd
Build from text stub feature was initially implemented to improve build
performance and may be unstable for api finalization, as the stubs
generated from api signature files miss some methods, which may cause
backward compatibility issues. Therefore, this change disables building
from text stub for api finalization.
Test: build/tools/finalization/finalize-aidl-vndk-sdk-resources.sh and inspect generated ninja file to verify that the stubs are generated from source Java files.
Bug: 310279899
Change-Id: I9d374e973788bc97514815f64eeb949a94a97647
Move OTA signing flags to payload_signer.py so that we can re-use these
flags in multiple binaries.
Test: th
Bug: 293313353
Change-Id: I44f9910cee37c449397e174a5784f747ec9fb0d6
There are certain options which we need to share in multiple binaries,
for example, the signer options. Current options parsing function only
accepts 1 extra option handler, which is inflexible. Extend it to take a
list of extra option handlers.
Currently, to add a new CLI flag, caller must append the flag name to
`extra_long_opts`, then pass an extra option handler which can handle
that option. Define a new dataclass which contains both the CLI flag
name and the code to handle that flag for better composition.
Test: th
Bug: 293313353
Change-Id: I758db66dfd95934f5b2701454d97bfe7d37dc16d
By caching flag values directly in member variables instead of caching
a HashMap and accessing that, flag reads avoid `hashCode()`, map
lookup, and Boolean.parse runtime costs. Flag reads are turning out
to have performance problems in hot paths, so this should help to
alleviate that.
Bug: 309625014
Test: m
Change-Id: I923bf6af2ae3fcbbf2fee7126b492a47cd6049ad
