On Sun, Feb 14, 2010 at 02:32:29AM -0800, Junio C Hamano wrote: > Jeff King <peff@xxxxxxxx> writes: > > > It feels ugly that we need to re-resolve the ref when we simply threw > > that information away earlier from when we called resolve_ref in > > get_ref_dir... > > It is true and I indeed thought about it while doing the patch, but "how > often would we see symref in the entire refs list, and what percentage of > callers would benefit from seeing this information?" stopped me from going > further in that direction. It is really only two code-paths (the one I mentioned, and the new one you are adding). I was thinking that we had to pay the resolve_ref cost for _every_ ref, symref or not, but that is not the case. Because we pass along the flag, we can re-resolve only the few that need it. So there is even less efficiency to be gained than I at first thought. Let's just drop it, as it is such an invasive change for no good reason. -Peff -- 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