From: Guido S. <__g...@we...> - 2004-10-21 17:33:51
|
Am 21.10.2004 10:47:51 schrieb(en) Stephen Watson: > >>> Guido Schimmels <__g...@we...> 20/Oct/2004 17:48:02 >>> > >Am 20.10.2004 12:29:30 schrieb(en) Stephen Watson: > >> Just got another crash in the same place. It looks more like the > >> pointer is getting corrupted somewhere. > > > So something must be stepping on the toe > >somewhere. I don't know how to debug that. Probably must set watch- > >points. > > Got it. The clue is the value 0x1, or rather TRUE. Look at client.c > line 519 and client.h lines 62 to 64. No, I checked this. That's ok. I have found the real cause. handleButtonPress(): win = get_mouse_window(dpy, c->frame) -> win = ev->subwindow Previously OroboROX handleButtonPress() only operated on the frame, and bailed out early else. Now clicking inside the app-window could cause a race or something. get_mouse_window() does a XQueryPointer(), which is silly when the value we are looking for is right there at our finger-tips. So this fix brings a small optimization also. |