Version: (using KDE KDE 3.5.5) Installed from: Gentoo Packages Compiler: gcc version 3.4.6 (Gentoo Hardened 3.4.6-r2, HTB-3.4.4-1.00, ssp-3.4.6-1.0, pie-8.7.9) OS: Linux I have a non-rectangular merged-fb (radeon - it simulates xinerama) setup on a laptop; 1024x768 on the built-in panel, 1280x1024 on the external monitor arranged thus: ._________._______. | | | | | | |..........___2___| | x x | |___x1___x| Normally I place the panel at the bottom of the larger display (at 1) and everything works fine. However if I place the panel at the bottom of the smaller screen (at 2), there's a "dead" area below the bottom of the smaller screen on the larger screen (marked with 'x's), such that when moving windows around, the refuse to move into the dead area. Windows can expand downwards into the dead area, but can't move into it completely. To be accurate, the title bar of the window refuses to move below the level of the panel on the smaller screen. It seems that the placement algorithm is bounding the window movement by limiting the base of the title bar to the top of the panel, rather than the base of the screen in which the window is placed. Attachments to follow: xprop -root (panel on larger screen) xprop -root (panel on smaller screen) xprop (window with panel on larger screen) xprop (window with panel on smaller screen)
Created attachment 18683 [details] xprop -root; panel on larger screen
Created attachment 18684 [details] xprop -root; panel on smaller screen
Created attachment 18685 [details] xprop; <window> - panel on larger screen
Created attachment 18686 [details] xprop; <window> - panel on smaller screen
I see a similar problem with two identical screens; if the panel on screen 1 is at the top, windows cannot be placed in the corresponding area on screen 2.
Clarification: if I select "Allow other windows to cover the panel" (Panel->Configure Panel->Hiding), the problem goes away. So, this is probably a panel bug, not a kwin bug.
It would be wonderful to know if it is happening with plasma or no
*** This bug has been marked as a duplicate of bug 160068 ***