Re: Bad objects error since upgrading GitHub servers to 1.6.1

From: Jeff King <peff@peff.net>
Date: Jan 28, 2009, 08:17 UTC
Message-ID: <20090128081745.GA2172@coredump.intra.peff.net>
In-Reply-To: <7vfxj3vos2.fsf@gitster.siamese.dyndns.org>

On Wed, Jan 28, 2009 at 12:05:33AM -0800, Junio C Hamano wrote:

Show 23 quoted lines

> Jeff King <peff@peff.net> writes:
> 
> > But in other cases, it silently gives you the wrong answer.  For
> > example, consider a history like:
> >
> >        C--D
> >       /
> >   A--B
> >       \
> >        E--F
> >
> > now let's suppose I have everything except 'E'. If I ask for
> >
> >   git rev-list F..D
> >
> > then it will not realize that A and B are uninteresting, and I will get
> > A-B-C-D. I think it is much better for git to complain loudly that it
> > could not compute the correct answer.
> 
> Fair enough.  I think we can resurrect the conditional and the traversal
> option revs->ignore_missing_negative only for this hunk in my [2/2] patch
> to support that use case.
> [ hunk handling parent lookup]

Don't the other changes have similar parallel use cases? [2/2] also deals with tag lookup. Wouldn't you also expect, if you had a tag "T" pointing to "E" in the above scenario that "git rev-list T..D" would barf? I really think you don't want to ignore missing negations _ever_ unless the caller knows that such a miss is really only about optimization and not correctness.

Side note:

As you described, we expect to reach this situation from a partial transfer. Which means that you don't actually have a _ref_ for "T" (or "F"). So it is unlikely to come up in normal use (you would have to manually specify the sha1 of a broken portion of the graph).

But what is more important is that your repository _is_ corrupted, I think we are losing an important method by which the user finds out. Git is usually very good at informing you of a problem in the repo early, and I think unconditionally ignoring missing objects would lose that.

-Peff