Regular repository corruption -- help needed.

Sat Dec 22 22:47:31 UTC 2012

On 12/22/2012 11:33 PM, Bryan O'Sullivan wrote:
> On Thu, Dec 20, 2012 at 12:37 PM, Alexander Krauss <krauss at in.tum.de
>     I see. Is this analysis consistent with the fact that
>     already-existing changesets get corrupted by the push?
>
> Are you sure that's what is happening? Because that is not what I see in
> the repo tarball you posted. In that case, the corruption is very
> definitely just at the end of the changelog file.

What was pushed in that attempt was just the very last changeset 
(f490917c959c). What got corrupted (according to hg verify) were a 
couple of the earlier ones as well.

Alex