Details
-
Bug
-
Status: Closed
-
Major
-
Resolution: Fixed
-
1.5.x
Description
As part of the investigation of issue #3067, code examination led me to wonder how the merge algorithm handles a particular scenario. Namely, when a subtree that needs special attention (due to having different mergeinfo than its parent, typically) has a rename in its history in the middle of the requested merge range. I suspected the logic would suffer the same sorts of problems noticed in issue #3067, and I think I was right. I'll attach a recipe to demonstrate.