Force pushes disallowed to all branches on @haskell

[andreasabel]

It seems that all force-pushes are disallowed on @haskell hosted repos.

old-time │ ghc-9.14 ⇣1⇡3 *1 ?1  git push --force     
...
remote: Resolving deltas: 100% (6/6), completed with 4 local objects.
remote: error: GH003: Sorry, force-pushing to ghc-9.14 is not allowed.
To github.com:haskell/old-time.git
 ! [remote rejected] ghc-9.14 -> ghc-9.14 (pre-receive hook declined)
error: failed to push some refs to 'github.com:haskell/old-time.git'

This is hampering my workflows when I prepare PRs for packages I maintain here.

Can we lift this restriction so it applies only to the main and master and other default branches?

Ping @Bodigrim

[Bodigrim]

It's not a global setting and if some maintainers prefer to prohibit force pushes I don't feel like overriding them is necessary.

That said, many repos were protected from force pushes by GitHub admins long time ago (I don't know why, presumably there was a security incident or suspicious behaviour). Let me ask GitHub support to unblock old-time. Feel free to ping if there are other repositories.

[davean]

I'd like to also comment that force pushes break mirroring.

…

On Tue, Jan 6, 2026 at 3:42 PM ˌbodʲɪˈɡrʲim ***@***.***> wrote: *Bodigrim* left a comment (haskell/meta#14) <#14 (comment)> It's not a global setting and if some maintainers prefer to prohibit force pushes I don't feel like overriding them is necessary. That said, many repos were protected from force pushes by GitHub admins long time ago (I don't know why, presumably there was a security incident or suspicious behaviour). Let me ask GitHub support to unblock old-time. Feel free to ping if there are other repositories. — Reply to this email directly, view it on GitHub <#14 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABZLSZXM3V3LCEX47HSQSL4FQM3XAVCNFSM6AAAAACQZ4G2AOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTOMJWGI3DCOJQHA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[andreasabel]

@davean

I'd like to also comment that force pushes break mirroring.

This statement is not self-evident to me. Why can't one mirror forced-push commits?
Do we have a mirror of the haskell organisation?