[PATCH] D154130: [lit] Avoid os.path.realpath on Windows due to MAX_PATH limitations

Fri Jul 7 12:15:31 PDT 2023

aaron.ballman added reviewers: jdenny, jhenderson.
aaron.ballman added a comment.

Adding a few more folks who are interested in lit changes to try to get the review unstuck.

FWIW, I worry about the subtlety of the `>` change because it's not entirely clear to me when I'd need to use `%>t` in a test. I worry code reviewers will miss this sort of thing and we'll only find out there's an issue when the test fails for someone with a problematic path. Is there a rule of thumb we should be following for its use?

In D154130#4480946 <https://reviews.llvm.org/D154130#4480946>, @MrTrillian wrote:

> All premerge build failures seem like flukes.
>
> - `x64 windows` failed 1/3 times
> - `x64 debian` failed 2/3 times with a timeout (passes locally)
> - `libcxx` seems to be failing for everyone: https://buildkite.com/llvm-project/libcxx-ci

FWIW, I agree.

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D154130/new/

https://reviews.llvm.org/D154130