I’ve been mostly a poetry guy but have tested out uv a bit lately. Two main advantages I see are being able to install Python (I relied on pyenv before) and it’s waaay faster at solving/installing dependencies.

[–] [email protected] 2 points 1 month ago (2 children)

Yeah, it certainly looks nice, but my problems are:

everything runs in a docker container locally, so I don't think the caching is going to be a huge win
we have a half-dozen teams and a dozen repositories or so, across three time zones, so big changes require a fair amount of effort
we just got through porting to poetry to split into dependency groups, and going back to not having that is a tough sell

So for me, it needs to at least have feature parity w/ poetry to seriously consider.

[–] [email protected] 3 points 1 month ago (1 children)

You should be using dockers cache mounts

https://docs.docker.com/build/cache/optimize/#use-cache-mounts

[–] [email protected] 3 points 1 month ago

Good call. We have some other tech debt related to our docker usage, so I'll add this to the list.

[–] [email protected] 3 points 1 month ago* (last edited 1 month ago) (1 children)

uv is still faster with a cold cache

and uv does have dep groups

about the second problem, there's an issue open on writing a migration guide, but migrating manually is not too difficult.

[–] [email protected] 1 points 1 month ago

I'm not really worried about the migration work, from what I can tell it's basically just moving a few things around. I'm more worried about losing features the team likes largely for performance reasons.

Our primary use cases are:

dev tools - standardize versions of tools like black, pylint, etc; not necessary if we move to ruff, we'll just standardize on a version of that (like we do with poetry today)
tests - extra deps for CI/CD for things like coverage reports

I like the syntax poetry has, but I'd be willing to use something else, like in PEP 735.

One thing we also need is a way to define additional package repos since we use an internal repo. I didn't see that called out in the PEP, and I haven't looked at uv enough to know what their plan is, but this issue seems to be intended to fix it. We specify a specific repo for a handful of packages in each project, and we need that to work as well.

I'm currently looking to use ruff to replace some of our dev tools, and I'll look back at uv in another release or two to see what the progress is on our blockers.