Bas Zalmstra

I have been trying to reproduce this but have so far not been able to. Would you be able to share your complete `pixi.toml`?

Mm Im having a hard time recreating this from your gist because all the revisions that I could find already included a `pandas-only` environment. The first revision of the gist...

Thanks I will take a closer look!

Perhaps we can make a simpler integration test by loading our sample repodata.json, sorting sprcific packages like libuuid and pytorch and creating some snapshots out of that?

That is testing whether we solve similar which indeed requires sorting but doesnt necessarily hit these edge cases. Besides that I think it would be nice to have a test...

One more final thing: Could you also run the rattler_solve benchmarks to see if we didn't accidentally introduce a performance regression?

Implemented in https://github.com/conda/rattler/pull/1701

Thanks for these! I can confirm that these are horribly slow and instant with libsolv. We will investigate!

I've found a performance improvement that fixes these cases (they resolve instantly) but makes all other solves about 50% slower... Need more investigation. 😅

We have landed some pretty [significant](https://github.com/mamba-org/resolvo/pull/66) [improvements](https://github.com/mamba-org/resolvo/pull/67) to the solver in the latest version (py-rattler 0.7.1). | spec | time to solve | |-----|--| | mlflow=2.12.2=py39hb3b8efb_0 | 102 ms |...