Change https://r.android.com/1335521 added tag property to the Dist
struct so that it could be used to select one of a number of different
output files to copy to the dist instead of the single file that the
module type made available for dist. The output files were selected
by passing the tag to OutputFiles(tag).
Module types that wanted to support this new approach had to explicitly
set AndroidMkEntries.DistFiles = GenerateTaggedDistFiles(module).
Unfortunately, doing that had a side effect of changing the behavior of
dist entries without a tag.
That was because the change treated a tag that was not specified, as
being the same as "". So, prior to the change no tag meant use the
default dist file but after it meant use the paths returned by
OutputFiles(""). That changed the behavior of the java.Library type
which affected the behavior of the android_app module type.
Prior to the change the java_library would make the
Library.outputFile available for dist when no tag was specified. After
that change it would make Library.outputFile plus
Library.extraOutputFiles. The latter is usually empty except for
android_app which adds some extra files into there which will now be
copied to the dist. That change may have been intentional but there
was no mention of it in the change or the bug. Even if it wasn't
intentional it may still be beneficial.
Any module type that wants to add support for tags in dist runs the
risk of introducing similar changes in behavior. This change
differentiates between the tag not being set and the tag being set to
"" to avoid that possibility and to make the default behavior
explicit for those module types that have switched.
It does so as follows:
* Adds a DefaultDistTag constant that is used when the tag is not set.
It is a string that is unlikely to be used as an actual tag as it
does not start with a . and uses some special characters.
* The DefaultDistTag is used in MakeDefaultDistFiles(paths) to indicate
that the supplied paths are the default ones and and also in
GenerateTaggedDistFiles() for Dist structures that have no tag
property set.
* The DefaultDistTag is passed to OutputFiles(tag) just in case the
module type has explicitly defined the paths to associate with that
tag in there. If it has then it overrides the legacy behavior. If it
has not then it is just ignored and falls back to using the previous
behavior.
* The java.Library.OutputFiles(tag) method explicitly handles the
DefaultDistTag and returns Library.outputFile for it which restores
the behavior from before the change that added dist.tag support.
* Similar change was made to apexBundle.OutputFiles(tag) in order to
preserve its previous behaviour.
* The customModule used by TestGetDistContributions has been modified
to also preserve its previous behavior after this change.
Test: m nothing
m dist sdk - before and after this change, compare result to
make sure that there are no significant differences.
Test the effect on the apex by following instructions in
http://b/172951145
Bug: 174226317
Change-Id: Ib8f0d9307751cc2ed34e3d9a5538d3c144666f6d
This change adds some additional tests to cover the different
combinations of setting OutputFile and DistFiles properties in the
AndroidMkEntry.
These tests will prevent follow up changes from inadvertently changing
the behavior that some modules rely on.
Test: m nothing
Bug: 174226317
Change-Id: I6ab6f2711a5fa416ea2b84a1e0b8754d525d7a0c
Previously, only the dist property's nested properties were checked for
correctness. This change also checks the dists property's nested dist
structures and adds some tests to verify that the checks are run and
correctly report the location of the incorrect property even when it is
within a slice of dist structs.
Test: m nothing
Bug: 174226317
Change-Id: If5a19360e1e4c98ee3b5afc813e35349d1fc6f6f
A previous change duplicated the test cases from TestGetDistForGoals()
to test the getDistContributions() method. This change removes the
duplicate tests and leaves a single test to verify that
GetDistForGoals() uses getDistContributions() and
generateDistContributionsForMake() correctly.
Test: m nothing
Bug: 174226317
Change-Id: I545501016b932f9af0d5fda2f452a3a09932d3fb
Previously, the GetDistForGoals(Module) func combined the processing
of the dist properties with generating the make specific rules for
generating that dist. That has a couple of problems:
1. It combines two pieces of functionality into one method which is
bad practice.
2. It makes it hard to test because the make specific output ends up
containing absolute paths to temporary directories created by the
test.
3. It makes switching to a non-make output difficult and fragile as
changing the output will also require changing the tests.
This change adds an intermediate data structure to contain the result
of the dist processing. That processing is done by the new method
getDistContributions(Module) which returns the new intermediate
structure. It also adds generateDistContributionsForMake(..) to
generate the make output. The GetDistForGoals(Module) func uses them to
implement the previous behavior.
It adds identical tests to those in TestGetDistForGoals() but leaves
those tests alone to show that this refactoring does not change the
behavior. Follow up changes will clean up TestGetDistForGoals(). It
also adds a test for generateDistContributionsForMake(..).
Bug: 174226317
Test: m nothing
m dist sdk - before and after this change, compare result to
make sure that there are no significant differences.
Change-Id: I458b7c8e4485bf66d3498f50df85a8d65fc2ee00
By default, nanopb outputs a timestamp into the .c and .h files it generates.
This is usually considered bening. However, when using remote execution or
remote caching, this poisons all actions that have these files as inputs, and
also appears to affect transitive actions (actions depending on actions that
have these files as inputs).
This change adds the -T flag to suppress the timestamp, which in turn makes
the output files deterministic, which in turn allows the remote cache to
return previous results rather than requiring a re-execution of the
corresponding actions.
Test: ran the build a few times to check that the output is deterministic
Change-Id: Iecfa143dbf12a9bb7703dd0bc3f1b3e3fd56317c
The hardlink optimization attempts to save I/O by hardlinking input
files into the sandbox instead of copying them, but it never works on
a Linux build due to nsjail always making out and the source tree into
separate bind-mounted filesystems. If the optimization did work it
would actually cause build failures when it hardlinked the relative
bionic/libc/fs_config_generator.py symlink into the sandbox directory
without the target. Just remove the optimization for now since it
never works.
Test: m checkbuild
Change-Id: I96e8c0c145e7c99958639594edf8a93b69ae90eb
Previously, TestGetDistForGoals tested multiple test cases within a
single test so when it failed it was difficult to determine which test
case was the cause. This change runs each test case as its own
nested test.
It also corrects the order of expectedLine and line format parameters
to match the order in the message.
Test: m nothing
Bug: 174226317
Change-Id: I1408ec4125afc5c0b392cd7643dd3f630fe468e5
Source-generated library modules will have a dependency on the source
variant. This creates a cycle in the dependency graph which triggers a
warning from rust-analyzer. Ignore this type of dependency.
Test: m nothing
Change-Id: I13365093ebb88b00f6a72734b01114ec3e9a320e
This patch reworks the approach introduced in
https://r.android.com/1450819. That patch based the decision which
libraries should be added to the manifest <uses-library> tags by the
manifest_fixer on the "shared" status of the library.
That approach is incorrect for two reasons:
- It doesn't make sense to have a non-shared library in class loader
context ("shared" libraries are those specified in
frameworks/base/data/etc/platform.xml, and they are the only ones
that PackageManager knows about).
- It doesn't make sense to add anything but the top-level of the
class loader context tree to the manifest, because this part of the
tree is flattened to a sequence, and PackageManager cannot restore
it to the previous tree shape (there is an information loss).
This patch removes the "shared" bit of information from class loader
context elements and assumes that all libraries that end up in class
loader context are shared. Consequently, only the top-level libraries
should be passed to manifest_fixer.
Test: lunch aosp_cf_x86_phone-userdebug && m
Bug: 132357300
Bug: 168686456
Change-Id: I902690f0f38f1047fa79cf6ccbe956077eceaab0
Paul Duffin