Re: regression in multi-threaded git-pack-index

[Jeff King <peff@peff.net>]

On Fri, Mar 15, 2013 at 03:42:40PM -0700, Stefan Zager wrote:

> We have uncovered a regression in this commit:
> 
> b8a2486f1524947f232f657e9f2ebf44e3e7a243
> 
> The symptom is that 'git fetch' dies with:
> 
> error: index-pack died of signal 10
> fatal: index-pack failed
> 
> I have only been able to reproduce it on a Mac thus far; will try
> ubuntu next.  We can make it go away by running:
> 
> git config pack.threads 1

I couldn't reproduce the problem on Linux with the instructions you
gave. I did try running it under valgrind and it produced:

  ==2380== Conditional jump or move depends on uninitialised value(s)
  ==2380==    at 0x441631: resolve_delta (index-pack.c:837)
  ==2380==    by 0x4419D6: find_unresolved_deltas_1 (index-pack.c:898)
  ==2380==    by 0x441A45: find_unresolved_deltas (index-pack.c:914)
  ==2380==    by 0x4427CA: fix_unresolved_deltas (index-pack.c:1232)
  ==2380==    by 0x4421F5: conclude_pack (index-pack.c:1111)
  ==2380==    by 0x443A5C: cmd_index_pack (index-pack.c:1604)
  ==2380==    by 0x4058A2: run_builtin (git.c:281)
  ==2380==    by 0x405A35: handle_internal_command (git.c:443)
  ==2380==    by 0x405C01: main (git.c:532)

but the line in question is:

  if (deepest_delta < delta_obj->delta_depth)

And in the debugger, both of those variables appear to have sane values
(nor should either impacted by the patch you bisected to). On top of
that, running with pack.threads=1 produces the same error. So I think it
may be a false positive from valgrind, and unrelated to your issue.

Other than that, it seems to run fine for me.

-Peff