[Lldb-commits] [PATCH] D81697: Add support for batch-testing to the LLDB testsuite.

[Benson Li via Phabricator via lldb-commits]

bbli added a comment.

- Ohh ok, guess I should have gotten the point about catching all exceptions recursively clarified before I began.

- From looking at the code for `subTest`, the error catching functionality comes from `Outcome_object.testPartExecutor` context manager, and the rest seems to be subTest specific. The repo's version of unittest2 has not defined the `Outcome` class, however. Should I port over just this class then? In regards to updating to the newer framework, I personally am not confident in myself doing so as I do not know the exact specifics of why the fork was made, and from what I have been told, the version of unittest2 we have in the repo is about 10 years old.

- Also, `testPartExecutor` will catch all exceptions, but I believe we want to exit early and propagate non `self.failureException` errors all the way up for `runMethod` to handle, correct?

- Finally, by `expect_expr`, are you referring to the "expect*" methods that are defined in lldbtest.py? And are you suggesting to have the error handling context manager to be called inside these functions? Because wouldn't that lead to the same problem of having a function continue in the situation where the function has multiple `expect_expr` commands?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D81697/new/

https://reviews.llvm.org/D81697