[mlir][python] normalize paths in AddMLIRPython.cmake

[makslevental]

No description provided.

[TiborGY]

@makslevental If there is no reason for this PR to remain open as a draft, please close it. Thank you.

[makslevental]

This is the second time someone has left this comment (maybe it was you both times). I'm just curious - who are you and where does it say in the community guidelines (or whatever) that one can't leave PRs up?

[TiborGY]

This is the second time someone has left this comment (maybe it was you both times). I'm just curious - who are you and where does it say in the community guidelines (or whatever) that one can't leave PRs up?

It was me, and as far as I am aware, there is no guideline against leaving PRs up. I merely got mildly annoyed when I tried to search for open PRs on certain topics (to see if anyone is working on the issue), and got a bunch of stale-looking PRs to sift through. So I took it on myself to chip away at what I saw as a problem, and look for 1-2 year old PRs that seem to have fallen through the cracks for no apparent reason.

I do not mean to be nuisance. If you think this is disruptive I can certainly stop.

[makslevental]

I leave PRs up as a reminder to myself about current progress on various projects. Obviously not this one but several others that I'm sure would be considered stale by you (that you might've commented on eventually).

I merely got mildly annoyed when I tried to search for open PRs on certain topics (to see if anyone is working on the issue), and got a bunch of stale-looking PRs to sift through.

I think it's ironic because existing PRs, stale or not, serve as a resource of code/technique (and I often refer people to such stalled/stale PRs). My advice: if some one has a stale PR that works on something you want to work on, just work on it - no one can "claim" a project/issue as their own.

[TiborGY]

I leave PRs up as a reminder to myself about current progress on various projects. Obviously not this one but several others that I'm sure would be considered stale by you (that you might've commented on eventually).

I merely got mildly annoyed when I tried to search for open PRs on certain topics (to see if anyone is working on the issue), and got a bunch of stale-looking PRs to sift through.

I think it's ironic because existing PRs, stale or not, serve as a resource of code/technique (and I often refer people to such stalled/stale PRs). My advice: if some one has a stale PR that works on something you want to work on, just work on it - no one can "claim" a project/issue as their own.

Thanks for the perspective and advice. I do limit myself to those that are old (inactive since at least 2023) and either

1. look ready for review, especially if GH indicates no merge conflicts
2. have positive reviews but no merge (or apparent reason for the stall)
3. drafts with not much in them (like this one)