On Sun, Sep 14, 2014 at 09:22:56PM -0400, Peter G wrote: > > > > can you make me a new log with the attached events.c please? I'm > > only interested in the part with the many ...NAME notifications, > > before and after you kill the application. > > > > sure, attached. > > around line 246 i skip a lot of the same events... > > then after line 511 i kill maple.
I assume you mean line 411? > ..things now come back quickly - seems you're "flushing" the left over > events. Hrm, I just added another flush call in the case XGetGeometry() fails. I wonder why that catches 26k events per atom, when before the application was killed it weeded out only between 1 and 50 events per call. When you run "top", you see only fvwm using a noticeable amount of cpu, and not the application? Tomorrow I'll write my own application that generates these events. I hope with that I can debug the situation myself. Ciao Dominik ^_^ ^_^ -- Dominik Vogt
