bug#81520: Add pre-receive hook to Savannah to reject LLM-encumbered commits

> From: Sean Whitton <[email protected]>
> Cc: [email protected], [email protected]
> Date: Fri, 31 Jul 2026 13:32:48 +0100
> 
> Eli Zaretskii [31/Jul  2:45pm +03] wrote:
> > My point is that when I have several commits, and then pushing them
> > all fails because one of them, going back and redoing the one that
> > caused the fail is a terrible nuisance with Git.  So I want the
> > rejection happen when I commit, so I could fix it right away.
> >
> > I don't mind of push fails for a reason whose fix is simple.  For
> > example, if someone pushed in-between, all I need to do is "git pull"
> > followed by another push, and that's fine by me.  By contrast, fixing
> > a commit in the middle of a large number is much harder, especially
> > when they are not your commits.  Think about merging from a release
> > branch, or "git am" of someone who submitted a series of several
> > commits.
> 
> Just to note that if it's commit message related, then for Git you can
> type 'e' in 'C-x v L' to fix it, these days.

Sorry, I don't trust these advanced features enough not to screw up
the Emacs Git repository, especially with my ancient version of Git.

What does it do? move all the commits after the offending one to a
branch, then back?

(Btw, what this does and when it can/cannot be used is almost
completely undocumented.  Both the doc string and the manual are
silent about that, and both are one-liners.  How is a user supposed to
learn about these capabilities and their limitations?)

> > What are the problems with rejecting individual commits when they are
> > done?
> 
> It makes it difficult to make commits you never intend to push, e.g.
> for local-only checkpointing.

On the public branches?  Why do you do that?  It's inherently
dangerous.