Your argument in relation to GitHub does not take in the reality of the effort involved with migrating to a different platform, effort that is likely unpaid, has no logistical upside and stalls the development efforts of a project,

forgejo can automatically import issues, PRs, Wiki articles, and automatic pull/push mirrors can be set up to keep the repo up to date at other places.

the CI/CD system is almost the same.

all the usual features are the same or very similar, including the whole user interface.

not to mention breaking every single source code repository link across the wider internet, links that represent publicity and community engagement.

who said they need to delete the repo from github?

I think it's admirable that a project is asking its community if it should stay or move

it indeed would be admirable if the communication did not contain offensive tones at multiple places.

and also don't forget that they already migrated once. but in the meantime it seems they have gone back to github for reasons unknown to me.

You're right. I don't mean to minimize the effort required. The effort required is a big part of the argument in favor of moving, or at least aspiring to move to a platform with more open and interoperable values. I can't imagine MS will make that transition any easier as time goes on despite forgejo and others best efforts. I've no problem with an OSS projects using GH but I'd hope they'd take the risk more seriously in a discussion about it.

Edit: I also don't think the effort is wasted or insurmountable. Regarding broken links, I've stumbled across many projects that have changed their GH repo to a mirror and link to their new platform. And RE logistical v philosophical reasons, I consider avoiding vender lock-in to be risk management and part of a project's long-term logistics.

Certainly it was a mistake to use Github in the first place.

Maybe, but if so, I bet it's negligible. When it comes to discovery, there's so many places I'd look for FOSS projects before going GH. Except maybe to check awesome-lists, but you don't have to be on GH to be linked on one (and I've seen them popping up on Codeberg). GH's design in general doesn't seem to promote stumbling across new projects. Even if I'm wrong, one could always mirror on GH.

As for contributing, if someone is willing to go though the trouble to contribute, I'd hope they'd go through the trouble of signing up on a new platform. Maybe there's a non-zero number of contributors who would not, and that's an unacceptable for some projects. There's also potential for more contributors if they trust a project is living FOSS principles and less at-risk of vender lock-in. The fosstodon thread shows people care about where a project lives. The arguments in favor of staying on GH seemed mostly inertia-based.