KMail stats

[Josef Weidendorfer]

On Tuesday 23 March 2004 03:03, George Staikos wrote:
> progress stuff.  Calltree shows a very large number of signals emitted, and
> a huge amount of time spent drawing progress bars and the folder tree. 

If painting is triggered by a timer (as seems to be the case here), calltree 
data will be missleading, as the timer will be triggered much more often in 
the simulation than in reality. The problem is that calltree can't provide a 
slowdowned wall clock time (I think that nanosleep() is slowdowned in a 
correct way by valgrind, but that's of no help here).

So the timeout value for the timer should be multipled by the slowdown 
(50-100) before calltree runs. If somebody has ideas how to handle this 
better, these are welcomed.

Josef