Re: [documancer] Failed to lock the lock file

[email protected] Mon, 16 Feb 2004 22:16:07 +0100 (CET)
Newsgroups gmane.text.documancer.user
Message-ID <[email protected]>
Hi Vaclav,

On Thu, 12 Feb 2004, Vaclav Slavik wrote:

> See attached patch, it adds "cachedir" attribute to books which lets
> you change the directory to /var/lib/documancer/bookname or whatever
> else you want. No read-only checks yet.

Many thanks  - I applied it and moved
the indizes for all the documents for our course - it works perfectly.

> >     and merging with
> >        documancer -c python_config.ini -c c_config.ini
>
> wxConfig can handle only two configuration files: global one (read
> only) and local one (where changes are written).
>
> >     (I don't think that a special --merge option would be necessary
> > ) - CVS: app.py is still ... 1.4  ;-(

Well, then presumably a simple -p documancer_profile_file
option to specify the local one
(and maybe -P documancer_global_profile_file for the global one)
is enough...

> If you're going to write some code, as you seem to want ;), then I can
> give you write access to the repository. Unlike anoncvs, this one is
> still fast.

Many thanks for the offer - if I get beyond the
point of "speaking is cheap" I will get back to your offer...

 > - occasional error message: I will ask the student to
> >   look into that problem - maybe he can track
> >   this down with enough playing ... ;-)
>
> Thanks!

by accident I realised that this strange "from-time-to-time"
error also comes up in the following context:
- start one documancer instance
- try to start another one. An error box "Another documancer instance
is running, exiting".
- Sometimes (1 in 20 or even less) I then get the following

OnInit returned FALSE, exiting...
Unhandled exception in thread started by <bound method
PyDocServerThread.__bootstrap of
<PyDocServerThread(Thread-1, stopped daemon)>>
Traceback (most recent call last):
  File "/home/python/PYTHON_TST/lib/python2.3/threading.py", line 451, in
__boot
strap
    self.__stop()
  File "/home/python/PYTHON_TST/lib/python2.3/threading.py", line 460, in
__stop
    self.__block.notifyAll()
  File "/home/python/PYTHON_TST/lib/python2.3/threading.py", line 256, in
notify
All
    self.notify(len(self.__waiters))
  File "/home/python/PYTHON_TST/lib/python2.3/threading.py", line 238, in
notify
    currentThread() # for side-effect
TypeError: 'NoneType' object is not callable

At least this is faster to test than clicking around and exiting ;-).
Maybe this gives you (me definitively not ;-)
an idea where/why this happens.
Somehow it seems to me that while exiting
the PyDocServerThread is gone but an attempt
to execute self.__stop()  is made,
but self is already "NoneType" ?


> > I can only _speculate_ what is going on here:
> > could it be that on your box  when it gets to wxLogWarning
> > the GUI is not yet there and thus stderr is the default?
>
> Yes, certainly. The strange thing is that we are running same app with
> same wxPython version so we should both see the same thing -- either
> the app is in GUI at this point or not.

maybe this is because the "remote" machine (the one with
NFS mounts) sometimes reacts "slow" - but this is just a further
speculation and I have no idea how to test this ...

Best,

Arnd