David

I see, I think that I probably deleted the transformations that kept this working as these aren't tested on CI anymore from the TF point of view, but I'll try...

This should be working now. Example here: https://github.com/openxla/xla/pull/2532 . I'm not sure there will be a commit to refer to (modification only touches internal files), but I'll update this comment...

There isn't much we can do in that case - I'm planning on making a wider push to ensure all regular contributors to XLA have their github accounts registered so...

That's a good point - I'm actually a bit surprised that the gardener is still so high. Part of that is due to LLVM integrates, but there are still many...

This is a good point - we will probably avoid using GitHub's codeowners, but something along these lines would make sense. Some document which records who is a good person...

Apologies for not responding to this earlier at @bhack, when you originally commented about this I looked into it briefly but couldn't find anything. It seems like this is due...

Made a couple extra changes to fix configure_test on ARM, should merge later today or Monday. Sorry for the delay!

FYI - I'm reverting the changes in gpu_compiler.cc for now because of some errors. Repro instructions should come shortly. Sorry about this!

I'll have to investigate what we do internally, an explicit goal I have for XLA is that the CI maps directly to what we do internally as often as possible...

TF and JAX behave differently in this regard, TF may have that pylint job on PRs, but internally the status of the pylint on the PR is irrelevant - only...