repository integrity(?) problem

Benoit Boissinot benoit.boissinot at ens-lyon.org
Tue Sep 15 04:23:41 CDT 2009


On Tue, Sep 15, 2009 at 08:41:42AM +0200, Berkes Adam wrote:
> Ok, I got the fun in it, but that is too easily reproducable, and I 
> already tried couple of different methods from different machine (I 
> tried first bundle source - unbundle in destination). I already over ~15 
> tries to get rid of this problem by different builds and transfer 
> methods (I've written to this list as a 'last chance').
> 
> I run hg verify on both machines: no problem.
> MD5 checksums for 00manifest.* are equal on both machine (in filesystem 
> level they only differs in creation date).
> 
> Last idea is that maybe encoding... But really not understand either as 
> ~5200 changesets are traversed without problem, and then this error.
> 
> 1 more difference on the machines: on ubuntu python 2.5.2 used while in 
> SuSE source builded 2.6.

Can you try tip ? There was some recent changes on the manifest code, it
might fix it.

regards,

Benoit

-- 
:wq


More information about the Mercurial-devel mailing list