Resolving a rejected patch crashes the workbench
Created originally on Bitbucket by DeeHants (Deanna Earley)
Hi.
Trying to resolve a patch that has rejected chunks causes a win32 access violation when I click yes to edit them.
In this case, the files being created by the patch already existed in the working dir but were NOT under mq/hg control yet.
This is reproducable in a test repo and the crash/patch failure does not occur if the conflicting file doesn't exist. If the rejected hunk only makes a change to a file that has failed, then the edit works fine.
I have a crash/memory dump if required.