Created attachment 3737 Patch I have two 1680x1050 monitors side by side with NVidia TwinView. I have a panel set to fill 100% of Monitor 1, docked at the top. Using xprop I get the following struts info for the panel: _NET_WM_STRUT_PARTIAL(CARDINAL) = 0, 0, 29, 0, 0, 0, 0, 0, 0, 1680, 0, 0 So top_end_x is 1680. But I think this should be 1679. The reason it matters for me is that I am using xmonad as my window manager, and with the current strut behaviour of the panel xmonad reserves 29 pixels along the top edge of my *second* monitor as well as the first, presumably because it thinks there is a strut butting 1 pixel into the second monitor. Attaching a patch which fixes this problem for me.
The 1680th pixel is still on the first monitor. Doesn't hurt tho, so I think its safe to apply this patch.
It works as expected in xfwm4 and I'm sure the prop is correct, so report the bug against xmonad instead.
(In reply to comment #1) > The 1680th pixel is still on the first monitor. Right, but the 1680th pixel is pixel 1679. :-) Pixel 1680 is the 1681st pixel, on the second monitor. I guess the question here is whether a top_end_x value of 1680 means that the strut ends *on* pixel 1680, or *before* pixel 1680. I don't think the EWMH spec makes this clear at all. Xmonad seems to be taking the former interpretation, but you're saying XFCE is taking the latter interpretation?
Yeah I guess that the difference here. Olivier enlighten me ;-).
For what it's worth, the panel in GNOME 2 agrees with xmonad here. A buddy of mine has a similar setup to what I described, but with a 1600x900 monitor on the left. He gets: _NET_WM_STRUT_PARTIAL(CARDINAL) = 0, 0, 25, 0, 0, 0, 0, 0, 0, 1599, 0, 0
*** Bug 7703 has been marked as a duplicate of this bug. ***
Applied in dbc73e9.