Stefan Monnier writes: > I pushed a patch to `master` which should handle those cases correctly. > Please confirm that it also fixes the problem on your side, Yes, thanks a lot Stefan. The fix looks good to me! I found one unhandled case. When a diff changes a file mode, like in the following: diff --git a/argos/session.py b/argos/session.py old mode 100755 new mode 100644 Note that it becomes "problematic" when the diff cumulates file modification and mode change for the same file. I understand that the "old mode" and "new mode" lines are the culprit since they're not handled by the regexp used by diff--font-lock-prettify: