Gitlab is not willing to make a team around federation
Gitlab is not willing to make a team around federation
Note: I think this is a duplicate but I can't find the other issue where this was previously discussed. It is related to other discussions on federation. I think useful to have a discussion where notes on how ActivityPub is progressing in GitLab can be recorded. --- @oelmekki started work to in...
The external developer who started the work and was highly praised by Gitlab offered to work for them if they made a team around federation --> nothing.
A group of French universities are now considering making a group in order to work on it themselves and contribute back to Gitlab.
Gitlab will most likely use it as a big selling point once all the work has been done by externals with little to no cost to Gitlab.
Not surprising at all. Federation is not a feature that their paying customers likely have much interest in... same for Gitea btw. the developers interested in forge federation AFAIK all went to Forgejo.
Personally I think Gitlab also needs a community driven fork. Maybe a more ambitious https://heptapod.net/ could be a good base to start that.
Considering that git already support email based collaboration, I agree that a federated forge is really niche. It's more of a frontend bonus. Which, considering the amount of company still using IRC, is not really a priority in the commercial world.
Gitlab will most likely use it as a big selling point once all the work has been done by externals with little to no cost to Gitlab.
I don't think so. It'd/'ll be a nice feature, and be listed as such. But it's not one of their primary selling points or marketing targets. Federation will be niche. Most useful in the FOSS space that pays little anyway.
I think you missed the point where 63 universities are using gitlab and actively interested in having federation. And that's just in France.
So what's stopping them? Universities have internship programs and internal projects. In a university team of 4 people doing projects, 63x4 252 students could be assigned to a project to build this.
But
The french open science committee (CoSO) is indeed interested in the ActivityPub implementation in GitLab
Good phrasing. They are "interested in the ActivityPub implementation" not "interested in the implementing ActivityPub" - so who gives a shit what a bunch of universities are interested in
I would not do any work for Gitlab nor run any software on Gitlab due to how poor they are doing at software security. https://arstechnica.com/security/2024/05/0-click-gitlab-hijacking-flaw-under-active-exploit-with-thousands-still-unpatched/
Also while trying to look that up there's a another one that is making the news from just the last couple of days? https://www.bleepingcomputer.com/news/security/gitlab-warns-of-critical-bug-that-lets-attackers-run-pipelines-as-an-arbitrary-user/
Good luck to anyone attempting to federate on top of a foundation like that.
I interviewed for them once, got rejected because I didn’t know some tiny corner of ruby on rails syntax despite working on it for 3 years. Huge bullet dodged..
Gitlab has always been too bloated for my liking. It's fine, but I prefer gitea. Edit: spelling
I host/manage it in my workplace. Only for groups, repositories, and merge requests with reviews. It's super bloated for such a [simple] use case. Every time I upgrade I especially see how loaded it is.
Their promotion is targeting a full-featureset devops and delivery pipeline with stats tracking and managed target environment.
Back then it was the better alternative to Phabricator, which we used before. We (I) may have chosen something different today.
I moved from GitLab last year, was a great decision for productivity
Federation is a bad fit anyway. Gitlab should just state that and reiterate focus on core functionality and sustainability, so we don't end up with another Gitorious situation.