RE: https://mamot.fr/@denissalem/116727062417331210

Lately I was complaining about the removal of #pipx from official #gentoo repository. It obviously doesn't disallow me to use #pipx by installing it through #pip …
🤡

The removal was certainly justified.

What struck me here is that it is now advised to use #astral_uv in replacement.

Which may be a great software … BUT

But now I need to compile an entire #Rust compiler which is itself a non trivial package for a source based distro …

Does #devs are aware of the concept of #minimalism …?

Kinda late but... Whoa, that's harsh. Maybe it's fair. I don't know. What I know is that python package management is still a mess in 2026. This is painful for devs and end users ...

The fact that the current miracle solution comes from #rust is a bit sad since python ecosystem is still unable to have a proper, unified and official package management ...

#gentoo #python #pipx #astral_uv

#javascript #nodejs peeps - how do you manage your nodejs tools that install binaries?

On the #python side @charliermarsh and co at Astral have given us #astral_uv that nicely bundles each 'app' into its own automatically upgradeable virtualenv.

I feel like I'm all over the place with installed node tools, is there any analog?

If anyone is installing Pyrolite (the excellent python-based geochemical data tool), it currently doesn't install with Python 3.13 and above due to a MatPlotLib compatibility issue. If you're using a managed virtual environment (such as uv), it installs fine with Python 3.12:
% uv init --python 3.12.13
% uv add pyrolite

@metasomite #Pyrolite #Geochemistry #astral_uv

Is there a best practice around setting up a modern Python module repo with pyproject.toml, uv, linting, tox, pytest, coverage, GitHub actions ...? There are so many different approaches, I'm struggling to bring it all together. #Python #GitHub #astral_uv #tox

*: unless, of course, your *other* dependencies have incorrect or missing lower bounds themselves, but at least now it’s easy-ish to find out ! (also, a great time to *fix* it)

#pytest #python #testing #softwaredevelopment #astral_uv

#pytest 9.0.0 was released yesterday. 🤗

It has a couple of exciting features like native toml support and a new umbrella strict mode, but the one thing I’m *really* excited about is that it’s the first version ever that has complete, correct and functioning lower bounds on its own dependencies… which means that it’s finally possible to test #python libraries with UV_RESOLUTION=lowest and get working, reproducible builds for no additional cost* !

#testing #softwaredevelopment #astral_uv