On Wed, Aug 27, 2008 at 08:44:05AM -0700, Dan Kegel wrote:
It would be hard to recognize later versions of the same patch, even with the '(try 2)' convention, since they often get split and munged while being beaten into submission. So Patchwatcher would have to have a way for developers to manually mark old patchsets as obsolete.
Most developers probably wouldn't bother unless this was really, really easy. I can't imagine that happening unless, say, we start requiring patches to be submitted via a secure web interface that verifies that the author has signed the Wine Contributors Agreement etc. At which point it wouldn't be too hard to coax developers into clicking a box saying "obsoletes this old patchset". But that'd be a pretty big change in wine's workflow.
- Dan
If it could handle working with GPG signatures, would that be a viable way to determine if the new patch is from the same author? I'm sure that most people could setup their mail clients to automatically sign their emails.
With that supported patchwatcher could limit the ability to obsolete patches to the originating author for the general case. And a special case of allowing people with certain trusted GPG signatures to obsolete patches.
When I say obsoleting patches, I'm thinking of the supporting commands embedded into the email: i.e. "!PatchWatcher patch mark obsolete <hash/id/etc>"