matplotlib

Torsten Bronger <bronger@...> writes:

> Thus, this is the trailing part which is missing:
> http://www-users.rwth-aachen.de/torsten.bronger/pds_missing.txt
>
> Does anybody has an idea at which point and why Matplotlib stops
> working?

The missing parts are the xref table and the trailer, which are written
by functions called by close():

    def close(self):
        # End the content stream and write out the various deferred
        # objects
        self.endStream()
  ...
        self.writeImages()
        self.writeMarkers()
        self.writeXref()
        self.writeTrailer()
        if not self.passed_in_file_object:
            self.fh.close()

So it looks like writeMarkers() has been called but writeXref() has not. 

-- 
Jouni K. Seppänen
http://www.iki.fi/jks

Jouni K. Seppänen <jks@...> writes:

> Torsten Bronger <bronger@...> writes:
>
>> Does anybody has an idea at which point and why Matplotlib stops
>> working?
...
>         if not self.passed_in_file_object:
>             self.fh.close()

I wonder if we should flush a file object that was passed in... Could
you check if the following patch (committed on the trunk) helps with
your problem?

Hallöchen!

Jouni K. Seppänen writes:

> Jouni K. Seppänen <jks@...> writes:
>
>> Torsten Bronger <bronger@...> writes:
>>
>>> Does anybody has an idea at which point and why Matplotlib stops
>>> working?
>
>> ...
>>         if not self.passed_in_file_object:
>>             self.fh.close()
>
> I wonder if we should flush a file object that was passed in... Could
> you check if the following patch (committed on the trunk) helps with
> your problem?
>
> [...]

Yes, it solves my problem.

Apparently, if you wait long enough, the rest of the file is indeed
written.  However, with your patch, I see the whole file
immediately, which is much better in the multi-processes environment
of an Apache server.  Thank you!

Tschö,
Torsten.

-- 
Torsten Bronger, aquisgrana, europa vetus
                   Jabber ID: torsten.bronger@...