Re: [Pan-users] Re: Reproducible crash retrieving headers in million-header group

[K. Haley]

On 7/30/2010 8:51 PM, Wayne E. Nail wrote:
> Pan died again, attempting the same newsgroup's 400-day header
> retrieval with
> the same conditions. I include this only because this time, I know that the
> relevant debug packages were installed *before* the Pan run.
>
> http://pastebin.org/434882
>   
The location is slightly different but it's still failing to allocate
some memory, about 3.6MB.
> If there are other switches, package installs, environment settings, etc. I
> might try to elucidate this problem further, I am ready and willing to 
> entertain them.
>
> To make sure... I have built this Pan instance by:
>
>  # cd /usr/local
>  # git clone git://github.com/lostcoder/pan2.git
>  # git pull
>   
...
>  # ./autogen.sh
>  # make
>  # make install
>
> Then, as my regular user, I ran gdb and pan from my ~ directory
>   
For testing I would suggest doing the build as your regular user and
leaving of the install.  Just run pan from the build directory.  In fact
I haven't actually installed Pan in ages, to many compile & run
sequences while programming.  If you haven't already tried it I would
suggest getting the headers without opening the group just to remove
rebuilding the header list from the equation.  Although it shouldn't
matter in this case you could also try the testing branch.

Oh, and I just pulled in a little over 1 million headers from
gmane.linux.kernel on windows and Pan used less than 500MB.

signature.asc
Description: OpenPGP digital signature