On Sun, Apr 09, 2023 at 08:10:43PM -0400 I heard the voice of
Stefan Monnier, and lo! it spake thus:
>
> For me it prints:
>
> Layout: [monitors=[len=2 cap=2 [x=0 y=335 w=1600 h=1200] [x=1600 y=0
> w=1200 h=1920]]
> horiz=[len=3 cap=3 [x=0 y=335 w=2800 h=1200] [x=1600 y=0 w=1200 h=335]
> [x=1600 y=1535 w=1200 h=385]]
> vert=[len=2 cap=2 [x=1600 y=0 w=1200 h=1920] [x=0 y=335 w=1600 h=1200]]]
>
> That looks very much like yours :-(
I looked over this quickly last year, but didn't see anything much. I
spent some more time poking at it today, but still couldn't reproduce
anything surprising.
Using a Xephyr window at 1365x768, and a MonitorLayout to split in
into similarly shaped pieces,
MonitorLayout
{
"ShortL:682x600+0+84"
"TallR:683x768+682+0"
}
Overridden layout: [monitors=[len=2 cap=4 [x=0 y=84 w=682 h=600] [x=682 y=0
w=683 h=768]]
horiz=[len=3 cap=4 [x=0 y=84 w=1365 h=600] [x=682 y=0 w=683 h=84] [x=682 y=684
w=683 h=84]]
vert=[len=2 cap=4 [x=682 y=0 w=683 h=768] [x=0 y=84 w=682 h=600]]]
the zooms all seem to DTRT. Sticking an xterm on the right side gets
full height on f.zoom, covers the whole side on fullscreening, etc
(well, technically, it's a little short on the bottom on zooming,
since xterm clips itself down to full line heights, but it goes all
the way to the top just fine).
> > Do you use Xrandr rotation for HDMI-1?
> > I didn't succeed to test monitor rotation using virtualbox...
>
> Yes.
This initially makes me suspicious of a connection, since we don't
even look at any such information (and we'd need whole separate calls
to even get it). But, it doesn't really make sense that it would
_matter_ to us at all anyway. As far as we care, the right screen is
1200x1920+1600+0, we do things at x coordinates between 1600 and 3520
(0 and 1200 + 1600), y coordinates between 0 and 1920... the rotation
only matters down at the X server level where it's rotating things
onto the monitor.
So that seems simultaneously the only odd thing that might affect
something, and a detail far below us that couldn't affect anything.
Maybe you could trying providing a MonitorLayout of exactly what XRR
is giving you already. Though it's what's in the Layout that matters,
and that looks just like we'd expect it to, so I guess if that _did_
fix something, we'd be even _more_ confused...
I did see a few _other_ odd behaviors from the differing sizes. Like
when having DontMoveOff/MoveOffResistance set, having a window hanging
off the bottom of the left screen, and moving it sideways it does some
weird L/R jumps at the moment it starts impinging on the other
monitor. And the default +0+0 IconManagerGeometry inexplicably shows
up on the _right_ side at x+0 y+(left side's top). The universe
obviously needs to just give us more square monitors so we can ignore
this sorta thing.
--
Matthew Fuller (MF4839) | fulle...@over-yonder.net
Systems/Network Administrator | http://www.over-yonder.net/~fullermd/
On the Internet, nobody can hear you scream.
