On Tue, May 16, 2017 at 9:47 AM, Jeff King <peff@xxxxxxxx> wrote: >> I think we want to behave consistently for shell builtins and for >> exported functions --- they are different sides of the same problem, >> and behaving differently between the two feels confusing. > > Yes, I think ideally we'd handle it all transparently. Although I'd also > point out that Git the behavior under discussion dates back to 2009 and > this is (as far as I can recall) the first report. So documenting the > current behavior might not be that bad an option. This is on Arch Linux which is usually pretty aggressive with their Git upgrades and I first saw the problem this week. The script that hits this case runs constantly on my machine so my guess is that if anyone relies on the old behavior, they're just starting to see a new-enough Git to have a problem.