On Wed, Apr 12, 2017 at 2:26 PM, Paul Eggert wrote: >> >> Perhaps we should change the hook so that it doesn't complain about >> problems from merges? > > Something like that might make sense, yes, for people in Alan's situation. > When I do a merge, though, I'd rather see problems from the other side (so > that I can fix them). How about an environment variable that captures the > user's preference? Since git only runs the hooks on the merged changes when you pass --no-commit to 'git merge', defaulting to non-blind merge doesn't really work anyway. There I propose the opposite default, as in the attached (also you had a copy-pasto "; done" in your patch).