From: Junio C Hamano Date: Tue, 31 May 2011 17:36:12 GMT Subject: Re: Re* diff: --quiet does not imply --exit-code if --diff-filter is present Message-ID: <7v8vtmkc1f.fsf@alter.siamese.dyndns.org> In-Reply-To: <20110531171401.GA12466@sigill.intra.peff.net> Jeff King writes: > I guess you did it this way to better share the discard-the-result > codepath. No, I did it as a hack because many places already do: if (... sub helper function that eventually call_callback ... < 0) break; // or return; // or goto fail_return; // or whatever to exit recursion and loop and obviously it was too much pain to change everybody to also pay attention to the new flag. A possibly cleaner way would be to designate a single negative value that is not -1 as "early return but not failure" without using an extra bit, but that also needs full vetting of the existing callchain, which I didn't want to do just to write a "it would be as little as this" patch.