Determine symmetric linear op fro CG from abstract output dtype

[Joshuaalbert]

Address #23403

[Joshuaalbert]

@jakevdp want to review this?

[jakevdp]

Lint is failing due to an unused import

[jakevdp]

The CI failures look to be related to this change

[jakevdp]

There's some issue with github actions currently, so for some reason I can't trigger the tests. I'd suggest running them locally via pytest -n auto tests to see if there's any remaining issues

[Joshuaalbert]

I don't see an easy way to squash since I've already created the PR. @jakevdp

[Joshuaalbert]

I squashed on another PR. https://github.com/Joshuaalbert/jax/tree/update-cg

[jakevdp]

You can squash the commits and force-push to the branch this PR was made from, and it will update here. Please let me know if you'd like me to walk you through it – thanks!

[Joshuaalbert]

@jakevdp figured it out. There was some issue with it dropping my branch after interactive rebase, which made force push fail, but found that git rebase --continue solved it.

[jakevdp]

Thanks - but it looks like there are now conflicts with respect to the main branch. Can you rebase against the updated main branch, and then make sure your branch only contains one commit on top of main? Thanks!

[Joshuaalbert]

@jakevdp There you go

[jakevdp]

Something went wrong here: you now have almost 100 commits on your branch. It looks like you probably merged an updated main branch, and then rebased against an older version of the main branch.

Can you please rebase against the most recent main branch?

[Joshuaalbert]

Probably because I forked and then modified the main branch. I'll see what I can do with this mess.

[Joshuaalbert]

Mess resolved?

[Joshuaalbert]

@jakevdp sorry was off, but remembered to come back to finish it off. Tests added

[Joshuaalbert]

@jakevdp back to you

[Joshuaalbert]

@jakevdp

[Joshuaalbert]

@jakevdp