pretxnchangegroup hook fails because it can't acquire lock

[Simon King]

On Tue, Nov 5, 2013 at 10:21 PM, Mick Jordan <mick.jordan at oracle.com> wrote:
> On 11/5/13 2:19 PM, Simon King wrote:
>>
>> [Re-CCing the list]
>>
>> On Tue, Nov 5, 2013 at 10:14 PM, Mick Jordan <mick.jordan at oracle.com>
>> wrote:
>>>
>>> On 11/5/13 1:56 PM, Simon King wrote:
>>>>
>>>> WSGI applications don't communicate with the web server over stdin and
>>>> stdout, so it doesn't matter if the app happens to write to stdout.
>>>
>>>
>>> Thanks, I'll get to WSGI asap.
>>>
>>> Meanwhile even a simple client clone, which shouldn't run any hooks, is
>>> failing with a similar error:
>>>
>>> [Tue Nov 05 14:06:41 2013] [error] [client ...] malformed header from
>>> script. Bad header=2 changesets found: hgweb.cgi
>>>
>>> Am I getting some hg logging somehow and if so how to turn it off?
>>>
>>> Mick
>>>
>> A wild guess: do you perhaps have an hgrc file somewhere with
>> ui.verbose or ui.debug enabled?
>
> Not AFAIK
>
>>
>> Are you able to use the hgweb interface in a web browser, or is that
>> broken as well?
>>
>>
> No that works (just tried it for the first time)
>

I think the message "2 changesets found" is coming from here:

  http://selenic.com/hg/file/85a358df5bbb/mercurial/localrepo.py#l1760

...but I don't understand the code well enough to know why it is happening.

Before pursuing it *too* much further, you might want to verify that
the same thing happens with a more recent mercurial release.

Simon