Pavel Dikov

Hello! We're experiencing this problem in our organisation as well. We tend to use our Python interpreters as absolute paths, as opposed to system-wide installations. We have tried various ways...

PR #3311 opened. On a related note, it transpires that I will need to make a similar change to `tox-uv` as well. The PR works fine with the default `virtualenv`-backed...

btw, the merged PR only fixes it for `tox-uv`. The vanilla `virtualenv`-backed runner still suffers from this issue. I had raised another PR (which is still open), but for consistency's...

> If I preinstall a .pth file in the venv, pointing to the base install (or some shared libs folder), would uv respect that? Just tried that -- it does...

I think catching ENOENT and friends (at arbitrary depth?) and re-caching the distribution from scratch should be robust enough? I am trying to think of the various types of race...

Heads up, I am seeing possible regressions on this one. I wonder if it may be worth adding an integration test for this scenario: 1. do a `uv pip install`...

Ok, I tried to repro by hand by chaos-monkeying with my `archive-v0` directory and I found two things so far: * if I remove a file from the package itself...

It's a synthetic reproduction of a real-life scenario, whereby a daemon process (e.g. `tmpwatch`) prunes files from a cache directory based on some coarse-grained criterion such as mtime/ctime. I would...

Ok, I had a deeper dive with the code and the problem is complicated by `tmpwatch`'s behaviour in that it will delete files but not directories. It is the 'deleting...

+1 for this. Being able to make multi-Python-version constraints/lock files is super useful, especially when ensuring build/test reproducibility on a library project that is being tested across a matrix of...