On Tue, 1 Jan 2008 11:04:42 -0600 Luis Rodrigo Gallardo Cruz <[EMAIL PROTECTED]> wrote:
> On Tue, Jan 01, 2008 at 10:11:13AM -0200, Thadeu Lima de Souza > Cascardo wrote: > > Hello. > > > > I've spent last day investigating this bug. The problem lies whithin > > Xlib and the integer format in librep, but the short-term solution > > goes in sawfish. I would like to reopen this bug as soon as it is > > closed and reassing it to Xlib. > > Thank you very much! This is one of the bugs I've never been even > close to understanding! > > I'll be testing your patch for regressions over the next days. Since I > don't have a system that exhibits the problems, I'll have to rely on > others to test it further. I uploaded your patch to the sawfish wiki > and will be forwarding your message to the mailing list, to see if we > get more testers. > Please, tell me if they find any problems or do any suggestions for a better solution. > > As a bonus, I attach my XCB code that will trigger the problem. > > Please, try it and the fix in 32-bit too, so we can be sure there > > will be no regression. > > I'm not quite sure what the code is supposed to do ;) In my system, it > just maps a little square window and prints > > Root window has ID 63 > New window has ID 2600000 > > The window is mapped in the current desktop, with normal decorations. Well, it will trigger the problem described in both bugs, if you are using amd64. That is, a message like Bad argument: #<subr make-vector>, 73786976294838206464, 1 will appear in standard error and if you click the middle button on the desktop, no menu will appear. If you use a platform where long is 32-bit, you won't see any of these problemas. But no other problems should appear. In amd64, with my patch applied, you will no longer see any of these problems. However, in a working system (there is, a 32-bit with or without my patch and a 64-bit with my patch), you should see that the window will be in every workspace, not only the current one, just like a panel or dock should appear. Since gnome-panel does not use _NET_WM_DESKTOP, it does not show these problems, but fbpanel does, and I assume gkrelmm, xmms and evolution do something like that too, in some condition or another. Regards, Thadeu Cascardo. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
