Wait. Wat. Someone finally brought python to the 21st century? I may be able to unban python from my toolkit for scripts and programs? Fuck. Well. about damn time I guess
https://astral.sh/blog/uv-unified-python-packaging
=> More informations about this toot | More toots from Di4na@hachyderm.io
@Di4na I've used Poetry for this purpose for several years now, I'm interested to see how uv
shapes up over time.
=> More informations about this toot | More toots from Conan_Kudo@fosstodon.org
@Conan_Kudo I have had... problems... with poetry... but yeah. Maybe things will finally get to a real place.
=> More informations about this toot | More toots from Di4na@hachyderm.io
@Di4na i interviewed for them and it was a lock until i asked about their monetization strategy
=> More informations about this toot | More toots from hipsterelectron@circumstances.run
@hipsterelectron How I have ... well. Let's say I have no doubt that part will be bad. But that is ok, monetization strategy for these kind of tools are always bad because there is none that survive our current reality
=> More informations about this toot | More toots from Di4na@hachyderm.io
@Di4na i'm just mad because i suspect financial incentives are part of why python tooling sucks
=> More informations about this toot | More toots from hipsterelectron@circumstances.run
@Di4na not trying to harsh your vibe but after that experience i only became even more strongly motivated to make pip light speed
=> More informations about this toot | More toots from hipsterelectron@circumstances.run
@Di4na obv doesn't address the stuff pip doesn't do, i'm extending spack for that
=> More informations about this toot | More toots from hipsterelectron@circumstances.run
@hipsterelectron I mean fair. That said, it is not the speed my problem. It is the UX. The UX of python tooling is uh... well abysmal? At the very least.
As I keep saying, as long as "venv" is something I need to know to do python, any kind of python, then it is a non starter.
=> More informations about this toot | More toots from Di4na@hachyderm.io
@Di4na yes absolutely. i think spack starts off from a better spot than any single-language package manager because it already has to handle architectures and language interop as well as bootstrapping binaries and caching. it doesn't solve a couple things relating to deployment (including deployment of spack itself) so that's what i'm integrating scie-jump from pants for https://github.com/spack/spack/discussions/44714 and hoping to achieve an extremely similar result to what's described in this post but on a much more solid foundation
=> More informations about this toot | More toots from hipsterelectron@circumstances.run
@hipsterelectron I doubt that tbh. But I am not the bearer of the only truth :D
I think the problem is simply that nearly all dev tools suck, we have just been lucky that a few people in a couple of visible community decided to lose their sanity and health trying to fix them and managed to do it.
=> More informations about this toot | More toots from Di4na@hachyderm.io
@Di4na i definitely think it's an easy pitch to VCs to make good tooling and a harder pitch to corps to let their engineers work on open source tooling vs enriching the corp and those incentives are part of why tooling always sucks even if nobody's trying to be actively evil
=> More informations about this toot | More toots from hipsterelectron@circumstances.run
@Di4na andrew gallant writes great code and docs and i really wish i could have worked with him and i'm upset that when i tried to frame to charlie how i can sell his open source tool to open source programmers that he interpreted it as me trying to undermine his goals
=> More informations about this toot | More toots from hipsterelectron@circumstances.run
@hipsterelectron I mean, totally fair. My point is usually that it needs really long term investment. I consider 10 years the minimum investment to start to get a viable tool that does a little bit of impact.
=> More informations about this toot | More toots from Di4na@hachyderm.io
@Di4na luckily pants and spack have both passed that limit already
=> More informations about this toot | More toots from hipsterelectron@circumstances.run
@Di4na i'm thinking about a lot of stuff today, appreciate the discussion and please don't let me stop you from using uv, i will be interested to read your thoughts on it
=> More informations about this toot | More toots from hipsterelectron@circumstances.run
@hipsterelectron @Di4na may I ask what is their monetization strategy? I was trying to figure out but I couldn't find information about it. Thanks
=> More informations about this toot | More toots from andreagrandi@mastodon.social
@andreagrandi @Di4na adding easy integrations with paid services similar to axo dot dev (which i have a rant about in one of my pinned tweets) but charlie was very cagey about it i wish i had realized that might upset him because i think their work is super cool
=> More informations about this toot | More toots from hipsterelectron@circumstances.run
@Di4na I may be a weirdo but I wish as much effort went into making python packages work fully with rpm and deb rather than building yet more elaborate virtualenv tools, so you could have one package manager with visibility into all the software dependencies rather than just the individual native language ones
=> More informations about this toot | More toots from raven667@hachyderm.io
@raven667 I mean, my answer is that deb and RPM are so obsolete compared to what we need for a packaging tool ecosystem at this point that I doubt they can be saved.
=> More informations about this toot | More toots from Di4na@hachyderm.io This content has been proxied by September (ba2dc).Proxy Information
text/gemini