gps rr on a patch series not picking up the branch name from the first patch

[Alizter]

One thing that I haven't been able to carry over from the gps branch workflow that I had going on is the ability to quickly do gps branch 0 get a review, reviewer wants changes, create a commit with those changes and then do gps branch 0 1. This worked before and in the new workflow the equivalent is gps rr 0 and then gps rr 0-1 however the second one requires a branch name due to the patch stack. I think in this situation the branch name should obviously be the one created during gps rr 0 and it becomes annoying to have to write out the branch.

I understand this is supposed to work, so it may be a bug in the current implementation.

[drewdeponte]

I remember writing code that would look at all of the patches in a branch and find all the associated branches, and if there was only one branch that was associated then that branch would be used. I don't remember exactly where this code is used. But I remember writing it.

I also remember deciding at some point that we were going to require a branch name be provided for a patch series. I don't remember all the reasoning behind that and if it conflicts with the idea above.

We need to do some digging into the commit history to review and figure out the purpose for the respective changes described above and figure out what makes the most sense.