Mercurial bug? Core hg dev input requested??? (was Re: win32text extension processing 00changelog.i?)

[Benoit Boissinot]

On Tue, Oct 13, 2009 at 11:26 AM, Ringo De Smet <ringo.desmet at gmail.com> wrote:
> Hello all,
>
> 2009/10/6 Paul Crowley <paul at lshift.net>:
>> I don't really have any great ideas I'm afraid.  All I can think to suggest
>> is: try cutting down the "hg-ssh" script to do almost nothing, so that it
>> just executes what it's given with no preparation or tests, and see if you
>> can reproduce the fault that way.  If it does then it's not my fault :-) If
>> not, add back the things that "hg-ssh" does until the fault comes back, and
>> we'll know a lot more about what's responsible.
>
> I haven't had the time yet to try and cut down on the mercurial-server
> script. However, I can eliminate the win32text extension from the
> problem domain. That hook is deactivated, but now I get the changelog
> problem in another hook I recently activated, one that only allows
> merges in a certain direction.
>
> I am tempted to say that there is a bug of some sort in the Mercurial
> internals, but it's very hard to track, as the problem is still not
> consistently reproducible. Input from a core Mercurial developer would
> be of great help!

Just an idea, did you check it's not a problem with your setup (hardware
failures, corruption in virtual machine, etc.)?

(without a way to reproduce it's really hard to help you, I'm sorry.
There doesn't seem to be anything obviously wrong in mercurial-server)

regards,

Benoit