rupprecht wrote: > Though maybe you intend this to break builds, that way obscure builders that > we don't know about will still hear about this? >
Yes, but I agree with your suggestion to land this with a default of `ON`, which makes this change NFC. I can then switch it to `OFF` once known buildbots are updated. I'll update the PR later today. > I can understand that angle, in which case do we want to make sure all known > public builders have pexpect installed before this commit goes in? I will > sort out Linaro's bots this week. Thanks! > Also are there known compatible/incompatible versions of pexpect? Is `pip > install pexpect` going to install something with significant API differences > to the vendored copy? - Our vendored version is at 4.6, with two local patches for timeouts: - a8419b1f2767c7fd5c0d0696b76d17efb2a5b418 - 97c6ef4ea678ef9a69e1feaf9d77a0880bca09ba - For my local machine (where tests are passing :) ), `python3-pexpect` is 4.8. Looks like that was released in 2020, and probably the version most people have. - The page https://pypi.org/project/pexpect/ lists trunk as version 4.9.0. We have some internal builds using that version which also work. - https://pexpect.readthedocs.io/en/stable/history.html doesn't seem to list anything significant recently btw, I see many tests w/ comments like this: `@skipIfWindows # llvm.org/pr22274: need a pexpect replacement for windows`. As of pexpect-4.0, there is (experimental) windows support. So if you're interested in Windows support, it'd be a good idea to see if those tests work now. https://github.com/llvm/llvm-project/pull/83191 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
