Re: [Pan-users] Leaks in 0.13.93

[Charles Kerr]

On Tue, Apr 01, 2003 at 07:22:58PM +0100, Dave Hill wrote:
> On Tue, 1 Apr 2003 04:57:37 -0800
> Charles Kerr <address@hidden> wrote:
> 
> > On Mon, Mar 31, 2003 at 04:12:09PM -0800, Steven Ellis wrote:
> > > 
> > > No sign of the memory leaks being fixed is there. I
> > > was running 0.13.92 last night and performed a load of
> > > some of the larger Binary groups and saw Pan's memory
> > > usage get to above 600 Mb before it was killed by the
> > > kernel.
> > 
> > Valgrind logs detailing the leaks are welcomed.
> > The pan-users mail archive has valgrind leak-finding
> > instructions.
> > 
> > I have already spent more than a day trying to track down
> > memory leaks, which never seem to come about on my system,
> > and am hesitant to spend more time on it unaided.
> > 
> > The last time I suggested a valgrind run, the pan-users
> > response was, IIRC, "it made Pan run really slow so I'm
> > not going to do it".  So for now that's also my response
> > to this reported memory leak.
> > 
> 
> OK, I just did "valgrind --num-callers=32 --logfile-fd=9
> --leak-check=yes --leak-resolution=high pan 9>logfile.txt", then I did a
> "get new headers from all subscribed groups" and then quit. Valgrind
> says it definitely lost 49 bytes in 5 blocks and possibly lost 800 bytes
> in 20 blocks. Is this the sort of thing you're looking for or is that
> just "in the noise".

Go ahead and mail the leak output offlist at address@hidden,
but we're talking about less than 1K, so none of these are the
source of your memory problems.