On Sun, Aug 31, 2014 at 11:30:54AM -0400, Jeff King wrote: > On Sun, Aug 31, 2014 at 01:24:48AM +0300, Max Kirillov wrote: > >> for_each_ref() does not include it itself, and without >> the hash the detached HEAD may be missed by some >> frontends (like gitk). >> >> Add test which verifies the head is returned >> >> Update test t6018-rev-list-glob.sh which relied on exact >> list of returned hashes. > > I think the missing bit of the justification here is that > "--all" _does_ include HEAD in other contexts (like in > git-log), and rev-parse should probably match it. > > This is probably the right thing to do. It's possible that > some caller of rev-parse really depends on "--all" meaning > "just the refs", but I kind of doubt it. Being in sync > with the revision.c parser seems saner. Actually, yes, this is a bit incompatible change, and while I'm pretty sure that rev-parse returning hashes should include detached HEAD, returning HEAD when it's called with something like "--symbolic" might be questioned. It could depend on the output mode (add HEAD only if printing hashes) but this kind of logic does not look good. So maybe some more opinions should be asked for. btw, manpage for git-rev-parse says "Show all refs found in refs/.", should it also be changed? -- Max -- 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