Michael Hanni <[EMAIL PROTECTED]> writes: > Just a historical note: when I first worked on the menu code many moons ago I > used override_redirect windows for the menu windows. I never had this race > problem back then... Of course, using override_redirect windows caused some > other problems, I just don't remember what they were... sorry. :-)
I just tested the use of the override_redirect flag. Problems: * With WindowMaker, they are present on all desktops. * Somehow placement is messed after hiding. * The X ICCCM documentation explicitly states that using the override_redirect flag is NOT a good idea when you are not grabbing the cursor at the same time. After a short time of experimenting I have found a solution. However the solution is such a hack that I hardly dare proposing it. Solution: -------- * Just before mapping in a NSBorderLessWindow (**) change the window attribute override_redirect to True. * After receiving the MapNotify event, change the window attribute override_redirect to value back to False. [**] We should restrict this as much as possible, so it is probably better to add an additional style flag like GSBypassWMWhenOrderIn or something along these lines. Note: It is NOT a good idea to use this approach for all windows because this will mess up the normal window manager placement behaviour. I have it implemented and it seems to work nicely. But it is such a terrible hack that I hope that someone else can solve it better. Please? At the moment I have a terrible time accessing CVS so I can not really make a nice patch out of this. Wim Oudshoorn. _______________________________________________ Bug-gnustep mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/bug-gnustep
