@csantosb I would say that niche scientific packages should definitely go to Guix-Science.
Also, contributing to Guix-Science is usually deemed easier, first because it’s on a forge of the kind most people are familiar with, with a clear feedback loop, and second because it’s a small repo without the ./bootstrap && ./configure && make step.
@daviwil As one of the folks trying to keep Guix infra afloat, I would rather reduce its scope than increase it right now.
But one thing that I find important is plausible self-hosting. And indeed that’s something we get with Forgejo, much less with GitLab (which is much more complex and cannot realistically be packaged in Guix).
@Guillawme That’s something for the community to decide but I’m willing to propose it once I have enough confidence, hopefully soon (maybe that’ll be the first RFC?).
Just now I tried ‘fj.el’ by the amazing @mousebot. There are rough edges (I didn’t manage to view the body of messages, almost!) but it looks promising: simple to use, convenient, and with apparently good coverage of the API.
I should say I’m pretty happy with the workflow we have on Codeberg and the many tiny features that overall make it easier to keep track of pull requests and ensure they reach the right people.
It’s routine for anyone who’s been using a “modern forge” but it’s good to see concretely how it could play out for a project like Guix.
Always amazed that when yt-dlp doesn’t work (it was unable to fetch an audio track in my case), you can run ‘guix shell --with-latest=yt-dlp yt-dlp -- yt-dlp …’ and it just works.
In a recent stream, I started using the Fibers library for Guile Scheme to learn more about how it enables Go-style concurrency in Scheme programs.
Here's a clip from that stream where I explain what Fibers is for, read through an example snippet using channels, and set up a basic HTTP request handler that uses Fibers' own server backend.
@lxo All I’m saying is that an org cannot claim to represent a movement while at the same time explaining that it’ll remain forever a club of a dozen pals.
@joeyh I also wonder what pushed RMS^W the five co-founders to publish this; what made them feel even more besieged than usual?
As to whether it’s still an “activist org”, I wonder. For an activist org, it’s much less active that many others in this domain. Much less relevant, too.