Alternatively, move the few things we do maintain on GitHub to be authored through our own Gerrit instance, and leave the GitHub mirrors be?

We should probably discuss this first. Either via a TechCom RfC or MediaWiki RfC.

Noting most of the GitHub "code hosting" is code mirrors. I don't see any problem with having the visibility of the repos in more places.

For repos/projects where GitHub is the canonical repo... Well, I'm not sure adding a third thing into the mix is the right thing. As has been said before, they apparently use GitHub to get more contributors (whether this actually happens is a different story), but like most of the rest of Wikimedia projects, they should probably be using gerrit like nearly everyone else

I believe there's been discussions of mirroring to other projects before, but never really got anywhere (bitbucket for one, maybe?)

In T247203#5952250, @Aklapper wrote:

Please clarify which specific code hosting you mean when writing "Move all code hosting from GitHub". Or how this relates to Wikimedia Gerrit / Git.

I thought I was clear but: don't use Microsoft's service, do use the FSF's (when it exists). Yes, it's not an option now, this is just a to-do, just like [[[ https://phabricator.wikimedia.org/T157529 | using GNU Taler ]] for donations.

@Koavf: Is this about mirroring to Github? Or is this about code repositories that have their canonical hosting on Github? Or both?
And again, how does this relate to Wikimedia Gerrit / Git?

In T247203#5952255, @Koavf wrote:

I thought I was clear but: don't use Microsoft's service, do use the FSF's (when it exists).

Ownership feels quite irrelevant here.

Yes, it's not an option now, this is just a to-do, just like [[[ https://phabricator.wikimedia.org/T157529 | using GNU Taler ]] for donations.

GNU Taler exists, "FSF Forge" does not...

Ditch the Microsoft-run code hosting and move to one run by the FSF, focus on free culture

Because ideologies and who produces that hammer matters more than whether that hammer actually exists... epic.
I'm not a Microsoft fan (rather the opposite) but it does not matter 2 dimes that they run GitHub. Moving repositories is a tremendous disruption. Moving them to /dev/null is a showstopper.

I would understand this idea if moving to GitLab would be proposed, as most of the anti-Microsoft agenda people did, which has a similar interface and functionality, reliability and it's proven.
Furthermore, WMF could run a GitLab instance. That has the potential of replacing the outdated gerrit, thus boosting the WMF into this decade. That's worth considering.

In T247203#5952301, @Demian wrote:

Furthermore, WMF could run a GitLab instance. That has the potential of replacing the outdated gerrit, thus boosting the WMF into this decade. That's worth considering.

Discussed, declined and offtopic - T167547: Replace Gerrit & GitHub with GitLab

I'm not sure why you're calling gerrit "outdated" either...

In T247203#5952303, @Reedy wrote:

In T247203#5952301, @Demian wrote:

Furthermore, WMF could run a GitLab instance. That has the potential of replacing the outdated gerrit, thus boosting the WMF into this decade. That's worth considering.

Discussed, declined and offtopic - T167547: Replace Gerrit & GitHub with GitLab

Discussed in 3 comments, declined by one person. 3 years ago. As time passes on and better, more efficient tools become available it's time to revisit that question, but not here, GitLab would be just one option to address the initial wish.

I'm not sure why you're calling gerrit "outdated" either...

... ...

In T247203#5952313, @Demian wrote:

In T247203#5952303, @Reedy wrote:

In T247203#5952301, @Demian wrote:

Furthermore, WMF could run a GitLab instance. That has the potential of replacing the outdated gerrit, thus boosting the WMF into this decade. That's worth considering.

Discussed, declined and offtopic - T167547: Replace Gerrit & GitHub with GitLab

Discussed in 3 comments, declined by one person. 3 years ago. As time passes on and better, more efficient tools become available it's time to revisit that question, but not here, GitLab would be just one option to address the initial wish.

I'm not sure why you're calling gerrit "outdated" either...

... ...

And Open Core software probably doesn't fit the bill. Which is one reason GitLab was declined. There are also more than 3 comments on the task.

Please don't make statements if you're not going to even attempt to back them up with a reason (personal or otherwise), or evidence. Is Gerrit outdated because we're not using the latest version? Is it outdated because it doesn't have some functionality that others do? Something else? ... ... is not helpful to anyone

As above, it's off topic anyway.

In T247203#5952316, @Reedy wrote:

There are also more than 3 comments on the task.

https://phabricator.wikimedia.org/T167547#3336421
Fifth comment: "Declined" 12 minutes after task was created. 1 comment by the same user, 3 comments by other users. Rock solid consensus.

Please don't make statements if you're not going to even attempt to back them up with a reason (personal or otherwise), or evidence. Is Gerrit outdated because we're not using the latest version? Is it outdated because it doesn't have some functionality that others do? Something else? ... ... is not helpful to anyone

That needs explanation? Honestly, I did not expect that question.
The answer is: UX. Productivity.

As above, it's off topic anyway.

When the perceived problem is the owner of GitHub, the proposed solution is to move, then GitLab, where most projects moved is hardly off topic. But true it wasn't the focus until you highlighted it in 2 comments.

@Demian: Please stop adding off-topic comments here. If you want to discuss Gitlab, then please go to the Gitlab task. This task is about "FSF Forge" - see its task summary. Thanks for your understanding.

edit: arg I just refreshed on comment submission and saw Andre's post. never mind

@Krenair I did not wish to discuss that topic either here, nor on T167547. I've just answered Reedy's comment to point out 3 comments in 12 minutes is not a consensus, then the topic blew up. I'm not interested in continuing this discussion in this manner.