New revision on publish?

One big disadvantage of the current approach is that the original draft revision message is replaced with the generic 'Bulk operation publish revision' log message and you lose the information about what author originally made the actual edit.

I'd definitely advocate for retaining the information about who authored the changes and why.

agreed, open to reviewing merge request(s) or patch(es) for adjustment(s).

I created an MR with my suggestion.

If this looks like a good approach, I could add the same logic in the unpublishing actions.

This works for me :)