Re: [PATCH 0/3] "-x" tracing option for tests

From: Jeff King <peff@peff.net>
Date: Oct 14, 2014, 13:44 UTC
Message-ID: <20141014134415.GA20675@peff.net>
In-Reply-To: <543CE44E.5050009@alum.mit.edu>

On Tue, Oct 14, 2014 at 10:52:30AM +0200, Michael Haggerty wrote:

Show 18 quoted lines

> >> Is your plan to reroll the prune-mtime stuff on top of these?  The
> >> additional safety those patches would give us is valuable and they
> >> are pretty straight-forward---I was hoping to have them in the 2.2
> >> release.
> > 
> > Yes, I've delayed while thinking about the issues that Michael raised.
> > There are basically two paths I see:
> > 
> >   1. These do not solve all problems/races, but are a solid base and
> >      sensible path forward for further changes which we can worry about
> >      later.
> > 
> >   2. There is a better way to provide prune safety, and these patches
> >      will get in the way of doing that.
> > 
> > I wanted to make sure we are on path (1) and not path (2). :)
> 
> FWIW I think we are on path (1).

Good. :)

I was preparing this to re-send, but I realized there is one snag. I mentioned that we should probably be ignoring already-broken links from recent objects to missing objects. For the traversal in pack-objects, we can use revs->ignore_missing_links for this. But for the one in git-prune itself, we use mark_reachable, which does not respect that option.

I think mark_reachable's traversal is essentially the same as the one in list-objects.c, and the two can be merged. I'll look into that, but I ran out of time for it tonight (er, this morning. Oops).

-Peff