Are we still doing a dev release today? The fixes I needed landed.

apparently i caused a pretty significant perf regression with #10441… sigh.

No worries. The regression is that the inference is expensive to do?

yea. guessing that it’s because none of the AncestorFiles requests are memoized because they’re all for unique sets

What would be the remedy, then?

either one request per file, or a recursive-upward implementation.

inference by default we could use the module mapping too

This might make a lot of sense. And then it’s easier to convince people to turn on import inference A key detail for me is that you can still turn off all inference if you want to avoid any perf hit.

ok, i’d like to get https://github.com/pantsbuild/pants/pull/10491 in and then cut the release. if folks could hold off on landing things that don’t block their other work, it would be appreciated.

Sounds good. I wanted to get

--use-first-matching-interpreter

if possible, but it won’t be ready in time. Have to do the Pex release right now, and then land the Pants PR. We shouldn’t block the release more

ok, this landed: feel free to merge!