Rutherther <[email protected]> writes: > Hi Simon, > > first of all sorry for this, this commit was from me and it shouldn't > have ever made it to Guix as my key file was wrongly added in keyring > branch and Guix couldn't authenticate it.
To clarify, I didn't intend to blame anyone, just collectively ponder if there is anything to improve here. >> I suppose I'll just force-push to (hopefully) intended commit from >> codeberg and things will be alright again, but I worry if this is a >> common pattern that there ought to be some policy around. > > How are you managing this mirror? It is normal GitLab.com server-side git mirroring. > Git already has a facility for > properly managing mirrors. Specifically, the repo is cloned with `git > clone --mirror`, then for updates `git remote update` and for updating > the mirror, `git push --mirror <url of the mirror>`. This way all refs > are updated, it does a force push of every ref so that the mirror is > the same as upstream, no matter what happens to it, be it force pushes, > removals of refs, additions of refs etc. GitLab.com has an option to mirror via force push, which maybe make sense. Although unless this happens frequently, maybe it is better for me to manually notice this. It could be one way to notice if anything less intentional happened. /Simon
signature.asc
Description: PGP signature
