In addition to reducing the amount of boilerplate we need to generate
whenever a new header is added, this also improves the existing tests
by running them in separate Lit tests (so they can be parallelized).
This also creates separate translation units for most header tests,
which is what we really should have done from the start since it
isolates each header we're testing.
Details
- Reviewers
jdoerfert philnik Mordante - Group Reviewers
Restricted Project - Commits
- rGc0f4b12da415: [libc++] Transition several remaining generated tests to the new .gen format
Diff Detail
- Repository
- rG LLVM Github Monorepo
Unit Tests
Event Timeline
LGTM with green CI.
libcxx/test/libcxx/clang_tidy.gen.py | ||
---|---|---|
19 | Given that we will most likely need this in (almost) every gen test, maybe we want to just make this a utility variable for gen tests? IDK whether we want to add a few utility functions/variables like this, but will definitely need at least some variables anyways. |
libcxx/test/libcxx/clang_tidy.gen.py | ||
---|---|---|
19 | I think this will eventually transition for XFAIL instead of #ifdef in the future now that we have one test per header, so I would wait until we make that change to see whether that's worth refactoring. |
Did we add documentation for this new feature?
LGTM!
libcxx/include/__memory/allocator.h | ||
---|---|---|
101 | Can this be landed separately? |
It is documented in format.py like we do for other types of tests.
libcxx/include/__memory/allocator.h | ||
---|---|---|
101 | Sure, will do! |
Drop changes to transitive_includes tests. Those are failing on Windows and I'll tackle them in a separate patch cause I want to land this and it's taking longer than expected.
Can this be landed separately?