Finally got a chance to work on a reproduction repo for performance issues: https://github.com/jriddy/clam-diggers/pull/1 This adds a huge number of dummy files that are generated by obfuscating my monorepo at $current_job. It doesn't quite succeed at making Pants plod the way our work repo does, but I don't think we're far off. I expect that by resolving some of the unowned dependency issues and adding a few plugins, and potentially dummy python content we can really push this thing to stress Pants the way it needs to be. Please provide any feedback, or if you have any suggestions on how we can make this nastier, please let me know

Thanks for doing this! Critical step in improving perf. Now we have an objective benchmark to work with.

It's not quite finished yet. As I said, I'm trying to get it to really lumber like it can on our main repo at work. But it seems like this method could allow us to reproduce it at least if I keep adding things to it.

thanks for doing this. i’m currently focused on the call-by-name changes, but will be back on this issue this month sometime.

I'm also going to give at try soon to pants 2.17, since we got a memory leak on that one at work that consumed 20+ GiB of memory in 2-3 min and sent our CI node into swap hell, no idea why yet 🙈