[kubav027]

I am pretty happy with poetry for near future. I prefer using python interpreters installed by linux package manager. In cloud I use python docker. Poetry recently added option to install python too if I changed my mind.

I have already setup CI/CD pipelines for programs and python libraries. Using uv would probably save some time on dependency updates but it would require changing my workflow and CI/CD. I do not think it is worth the time right now.

But if you use older environments without proper lock file I would recommend switching immediately. Poetry v2 supports pyproject.toml close to format used by uv so I can switch anytime when it would look more appealing.

Another thing to consider in long term is how astral tooling would change when they will need to make money.

[js2]

> I prefer using python interpreters installed by linux package manager.

uv will defer to any python it finds in PATH as long as it satisfies your version requirements (if any):

https://docs.astral.sh/uv/concepts/python-versions/

It also respects any virtual environment you've already created, so you can also do something like this:

   /usr/bin/python3 -m venv .venv
   .venv/bin/pip install uv
   .venv/bin/uv install -r requirements.txt # or
   .venv/bin/uv run script ...

It's a very flexible and well thought out tool and somehow it manages to do what I think it ought to do. I rarely need to go to its documentation.

> Using uv would probably save some time on dependency updates but it would require changing my workflow and CI/CD.

I found it very straightforward to switch to uv. It accommodates most existing workflows.

[irjustin]

I'm pretty much with you and still trying to figure out why I want to switch away from pyenv+poetry.

I get that uv does both, but I'm very happy with pyenv+poetry combo.

Old baggage, but I came from the rvm world which attempted to do exactly what uv does, but rvm was an absolute mess in 2013. rbenv+bundler solved so many problems for me and the experience was so bad that when I saw uv my gut reaction was to say "never again".

But this thread has so many praises for it so one day maybe i'll give it a try.

[armanckeser]

Uv dependency solving is light years faster than poetry. If you are working on actual projects with many dependencies, poetry is a liability

[kubav027]

You are right that it is faster but how often you are running dependency update? It will take more time to ensure that new dependencies did not break anything than doing upgrade.

[armanckeser]

It should, but if you are using poetry and say you forgot to pin boto to a specific version, your entire day might be spent waiting for poetry to solve.

[WhyNotHugo]

Yeah, using the package manager is the logical choice and usually the most likely one to work.

IIRC, uv downloads dynamically linked builds of Python, which may or may not work depending on your distribution and whether linked libraries are locally available or not. Not sure if things have changed in recent times.