[mythtv-users] mythbackend still eats memory: the current status

[Gavin Hurlbut]

> On 2010-11-28 09:38, Jean-Yves Avenard wrote:
> The ps output gets updated once a day.
> The trend was already clear, so maybe you want a graph?

Yes it was pretty clear.  For 24 data points straight, the memory use
was constant.  Look at the third column of your data.  That is the
only one that matters.  The fourth is how much of it is actually in
physical memory, and that is not a result of memory leakage, but
rather the virtual memory management system in the Linux kernel.  Your
own stats are disproving your theory.  24 days with absolutely no
change in VSZ indicates no memory leak whatsoever for 24 days.

After that, it grew a bit, likely due to different schedule data or something...

If you are going to claim that we are leaking memory, you need to
prove it.  So far, what you've shown this time around pretty much
proves to me that there is no memory leak.  As does the valgrind runs
I've done in the past, which indicate we may be leaking a few
kilobytes, which really isn't terribly significant.

> So I am looking for a spec file to build 0.24 from svn.

It's in svn.  I'm pretty sure that the spec file in there is
reasonably up-to-date.