Edmundo Carmona Antoranz <eantor...@gmail.com> writes: > One question that I have is if it makes sense to set option_commit > to 0 if the user didn't specify --commit when using --squash, so > that the current behavior of git is not broken.
If you mean that "git merge --squash <other args but not --[no-]commit>" should behave identically with or without your patch, then I think the answer is definitely yes. > Like you run merge --squash, git will stop as it currently > does... but it would be possible to run with --squash --commit so > that the revision is created if there are no issues to take care > of (currently impossible, you would see that message saying "You > cannot combine --squash with --commit."). That is exactly a safe way to extend the system by adding a new mode of operation in a backward compatible fashion. Good thinking.
