On Sat, Nov 24, 2018 at 04:47:19PM +0900, Junio C Hamano wrote: > I do agree that forcing the parent to wait, like you described in > the comment, would be far more preferrable, [...] Stray processes can sometimes have funny effects on an outer test harness, too. E.g., I think I've seen hangs running t5562 under prove, because some process is holding open a pipe descriptor. This would probably fix that, too. > but until that happens,[...] But if we can't do that immediately for some reason, I do agree with everything else you said here. ;) -Peff
