Moving away from GitHub

[ddevault]

Strongly considering moving this project to Gitlab. Leave your opinions here. Advantages:

• Gitlab is open source
• Gitlab has free, well integrated CI including support for self-hosted workers
• I prefer Gitlab's politics

[Luminarys]

I agree with all of your points, but you do have to consider the primary disadvantage of Gitlab having a much smaller user base. The good, and perhaps also bad, thing about Github is its ubiquity, which results in many users being able to find your project and contribute to it. But if you're ok with the decreased visibility, then I'd say go for it.

Edit: Perhaps the project could be primarily developed on Gitlab while also being mirrored on Github.

[ddevault]

Needs more discussion, folks. Speak up.

[onny]

Lets stay on Github until atleast one stable release is made. After that, the project might be "mature" enough to move on GitLab ... but now, we'll need as much contribution as possible

[ddevault]

I'm not sure we do need a ton of contributions. We've already got enough consistent contributors to build an i3 clone in a fairly short amount of time.

[ddevault]

I've found that most of the time, even projects with lots of contributors and exposure will be written almost entirely by one or two people.

[ghost]

Gitlab sounds great, but I'm a little confused; is Gitlab self hosted or only on the web?

[ddevault]

It's both. There's a public gitlab running on gitlab.com, but the code is open source and you can deploy it on your own infrastructure if you want. We'd be using gitlab.com.

[ghost]

I'd say go for it 👍

[ghost]

Because its an alternative to Github, I guess they've added migration options.

[ddevault]

Eh, for now let's just stay here.

[jansuchomel]

Maybe it's time to reconsider the move given that Microsoft is about to acquire Github?
https://www.bloomberg.com/news/articles/2018-06-03/microsoft-is-said-to-have-agreed-to-acquire-coding-site-github?

[RyanDwyer]

Yeah, it's time to reconsider this. I support a move to GitLab, even if Microsoft isn't buying them out. But I'd probably wait a while to make sure GitLab has sorted out any infrastructure issues that arise.

[Ongy]

I'll give my 2ct here, mainly as wlroots consumer, and small time contributer, since I highly expect wlroots to move with sway if it does move.

I'm personally indifferent to github. I have been until now and will be for now. I think it's one of the more fleshed out platforms, but I dislike how it centralises everything at a single company.
I personally don't expect a microsoft takeover to change anything in the near future.

Afaik the long term infrastructure for sway and related projects is sr.ht, which makes the current situation a bit more complicated.
IMO sr.ht isn't ready for a big project yet. It can host, but there's missing infrastructure for public review, and public tracking of current ongoing work.
For this reason I'm not currently in favor of moving there.

On the other hand, moving to gitlab at this point will probably require another move during the next couple of years, once sr.ht is at the point where it's suitable for sway.
Since moving projects leaves dangling pointers (i.e. old hpyerlinks) and is generally an annoying effort (migrating accounts, issues) I don't think it's a good idea to move now and then again later without a pressing need.

I propose to continue as before, long term plan to migrate but don't do anything rash.
Should there be any change to the platform after a buyout that reinforces reasons to leave, I'll gladly help out to do so.

[emersion]

As a contributor who probably wouldn't have started contributing if the project was using git-send-email on an obscure mailing list, I'd like to highlight that we need contributors and that lowering the barrier to entry is important. GitHub does a good job at this - I haven't really been using GitLab but it seems good enough.

Yeah, the plan is to eventually migrate to sr.ht - but it's not ready yet.

In any case, I don't think we should rush into migrating right now. As @Ongy, I don't want to migrate to anything if it's to migrate again after that.

[ascent12]

This is a discussion for the sway team, and has no bearing on end users. The opinions of those who don't have some stake in sway/wlroots aren't helpful at all.

[ddevault]

The plan is to move sway and wlroots to sr.ht - this has been the plan for longer than there have been talks of Microsoft buying GitHub. However, in light of the recent news, I intend to speed up work on sr.ht so that people leaving GitHub have another choice. To this end, I want to work on bringing sway+wlroots over sooner rather than later.

I know that it's not ready yet and that it would be disruptive to the project to move now. Can I get some feedback from core contributors on what you think sr.ht needs before it's suitable for sway+wlroots? If you don't already have an account please reach out and ask for one.

The biggest one is obviously public code review tools, ideally which don't require git-send-email to use.

[emersion]

Another proposal is to move to FreeDesktop.Org's GitLab instance, alongside the rest of the FDO projects. A blocker for this is proper builds.sr.ht integration in GitLab (not possible right now because of GitLab MR limitations).

[emersion]

I've been working on a small project to overcome the GitLab API limitations: https://git.sr.ht/~emersion/dalligi

It allows exposing builds.sr.ht as a GitLab runner. This should bridge the gap.

It doesn't yet support multiple build manifests, nor does it mirror build logs. But that's planned.

[emersion]

https://gitlab.freedesktop.org/freedesktop/freedesktop/-/issues/473

[ddevault]

Dunno what this project is but maybe you guys should consider SourceHut Codeberg