win32text extension processing 00changelog.i? (was Re: error: pretxnchangegroup.crlf hook raised an exception: integrity check failed on 00changelog.i:12)

Ringo De Smet ringo.desmet at gmail.com
Tue Oct 6 04:15:29 CDT 2009


Benoit,

2009/9/30 Benoit Boissinot <benoit.boissinot at ens-lyon.org>:
>
> No it shouldn't make a difference in this case. It would be very helpful
> if you try to get a "unbundle" setup very close to what mercurial-server
> does. And then reduce the test case from, it would make the debugging
> easier. (I know it's not always easy to do that, I hope you can find the
> time)

I found quite some time to try to get this failure scenario
reproduced. When I got it reproduced, it wasn't reproducible
systematically. I mean that when the error occurred, if I retried it
worked. So I failed to make it 100% reproducible. The only thing I can
tell now is that the scenario is most reproducible when running behind
mercurial-server and when one tries to push multiple changesets in a
single push.

I will contact the mercurial-server author to check if he has any
ideas to get this debugged...

Thanks anyway for your time.

Ringo


More information about the Mercurial mailing list