On Wed, Nov 28, 2012 at 2:54 AM, Junio C Hamano <gitster@xxxxxxxxx> wrote: > Felipe Contreras <felipe.contreras@xxxxxxxxx> writes: > >>> This is not just "just shuffle the die to make it explicit" but it >>> does change the semantics; earlier ref->deletion was perfectly fine >>> as long as data->refspecs is not given, but the new code always >>> dies. >>> >>> If this semantic change is a good thing, please explain why it is so >>> in the log message. If the change is "it does not matter because >>> when data->refspecs is not given and ref->deletion is set, we die >>> later elsewhere in the code anyway", then it needs to be described. >> >> refspecs are optional, but when they are not present the code doesn't >> work at all. This patch changes the behavior that was totally broken >> anyway. > > In case it was not clear, I did not request/expect responses in the > discussion thread, but a rerolled series with updated description. An updated description that is irrelevant; the stuff is totally broken, that's my point. But I'm tired of explaining it, and showing it with test cases, and patches that fix those test cases, it's not my itch, and "the other camp" doesn't even bother to acknowledge that indeed remote helpers without marks just don't work, or even utter a single word replying to one of these patches, nothing. Not blaming you, just saying that I don't think anybody cares, clearly nobody is exercising this code. Cheers. -- Felipe Contreras -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html