Similar to regular output of LIT tests:
https://github.com/llvm/llvm-project/blob/c162f086ba632ffaedfe92d63bf21571bc8ae4da/llvm/utils/lit/lit/TestRunner.py#L1569
Details
Diff Detail
- Repository
- rG LLVM Github Monorepo
Unit Tests
Event Timeline
The patch seems reasonable. I have a minor nit about the use of the % operator but other than that it seems fine.
llvm/utils/lit/lit/formats/googletest.py | ||
---|---|---|
125 | Minor nit. This patch is mixing the old python string interpolation (% operator) and the format function. Given that Python 3.6 is the minimum maybe we should just use format strings in both places e.g. header = f"Script:\n--\n{' '.join(cmd)}\n--\n" ... return (lit.test.TIMEOUT, f"{header}Reached timeout of {litConfig.maxIndividualTestTime} seconds") |
I see test failures on Windows that appear to be related to CRLF handling. I think the CHECK-NEXT checks aren't working because something is doubling up the newlines in the test output. That seems unrelated, so I suggest dropping the -NEXTs as necessary to get it past the presubmit testing.
it's just rebase so far, windows is not fixed
it does not reproduce for me so I will wait for another pre-merge
llvm/utils/lit/lit/formats/googletest.py | ||
---|---|---|
138–139 | I assume we don't need it here |
Issue is in the path separator.
lines like FAIL: googletest-upstream-format :: [[PATH:[Dd]ummy[Ss]ub[Dd]ir/]][[FILE:OneTest\.py]]/[[TEST:FirstTest\.subTestB]]
always have /
and command line is \
Minor nit. This patch is mixing the old python string interpolation (% operator) and the format function.
Given that Python 3.6 is the minimum maybe we should just use format strings in both places e.g.