Matt

That's a good point. For security reasons, git can't run hooks unless you manually copy them to the `.git/hooks` directory ([docs](https://git-scm.com/book/en/v2/Customizing-Git-Git-Hooks)). Could we consider something similar where we require a...

> And then the daemon re-opens `$EDITOR` with something like a `jj describe @- --add-footer 'BUG='` My point here was that by the point the validation had failed, the CLI...

> `jj fix` will be a standalone command and that has been decided since run was designed. I couldn't find any issues for fix when I looked. Could you link...

As you said, this is a rather niche case. IMO we would be better off fixing this by implementing the `jj gerrit download / patch` command that @martinvonz and I...

> The latter seems workable, but I still can't help but feel like this is ultimately a bug in `jj gerrit upload`. i.e., it seems that a commit created in...

I think it's a wrong assumption to assume that the commit uploaded will be the same as the one stored in the remote branch. I recently submitted [this CL](https://chromium-review.googlesource.com/c/chromium/src/+/7017050) on...

The sync.py script does solve that problem, but I wrote that for chromium specifically and it does a lot of chromium-specific things. It's way overkill for your use case. In...

> If something (either JJ or gerrit) adds the ChangeId footer to the commit on upload, could the local change be updated (as if you ran `jj desc`) amended such...

I have watchman enabled, but `watchman.register-snapshot-trigger = false` set.

I've added Yuja as a reviewer since he's the one I strictly need approval from, but FYI @martinvonz and @PhilipMetzger in case you'd like to review as well.