Strange reachability inconsistency (apparently, at least...)

I have a (fsck-clean) git tree in which for 2 commits A and B:

* "git merge-base --is-ancestor A B" returns 0
* "git log B..A" returns a non-empty set of commits

I get this behaviour with 2.3.0 as well as with 2.1.3 and 1.7.12.

Is that a real bug or am I just misinterpreting something ?

On Thu, Feb 19, 2015 at 7:03 AM,  <ydirson@free.fr> wrote:

Sounds strange. Is it possible to share the repository (or an anonymised copy of it)?

Junio C Hamano <gitster@pobox.com> writes:

Actually it does not sound all that strange.

I suspect that B is indeed a proper descendant of A, but you have commits with screwed-up committer timestamps between these two commits that break the cut-off heuristic used by log and reb-list.

Grab the output from "git rev-list B..A". These are supposed to be ahead of B (i.e. should not be reachable from B). Then look at the output from "git log B" (no lower bound) and see if these commits appear there, perhaps like:

    $ git rev-list B..A | sed -e 's/^/^/' >phantoms
    $ git log --format='%H %ct' B | grep -C2 -f phantoms

Hits from the latter "grep" would indicate that what the traversal B..A claimed that are not reachable from B contained some commits that are indeed reachable from B.