[mythtv] Profiling mythtv Channel Changes...

[Linuxguy123]

On Wed, 2008-10-22 at 21:05 +0200, Bert Haverkamp wrote:
> Do you have a script that we can run to benchmark channel changes? I
> can run this on 2 machines, one is analog, the other dvb-c. Currently
> I see ~5-7 sec channel change times overall. But we need to drill down
> to see where the delay is coming from.

I don't have a script.   I was thinking that I would download the code
and add debug like statements to each of the procedures that are used in
the channel changing sequence for my own private use. 

The debug statements would write timing info to a file and the file
could be analyzed later. This is what I did with an application under
Windows about 10 years ago.  Not pretty, but it worked.

I know that Linux/gcc supports a few profiling tools.  I've never used
them. 

The thing about using a single file to log all the calls is that one
could log info from both the front and back ends into the same file to
create a single log of the whole process. 

Is there a better way to profile the mythtv code ? 

I wasn't going to share the logging code.  I didn't anticipate that
anyone else would be interested in it.   If I do share it, how should I
do that ?  Should I create a new branch in svn ?   Should I "dirty" the
exiting code ? Should I create a private repository ?

Thanks