Panel layout bugs with multiple displays

Duncan 1i5t5.duncan at cox.net
Wed Nov 11 09:00:05 GMT 2015


Ian Pilcher posted on Tue, 10 Nov 2015 10:17:46 -0600 as excerpted:

> On 11/09/2015 01:40 PM, Ian Pilcher wrote:
>> These panel overlap bugs are pretty bad.  It's really hard to get the
>> panel set up in a usable fashion.
> 
> So it turns out that the panel layout bugs that I'm seeing (panel
> widgets overlapping, running off the end of the panel, etc.) are in fact
> related to using multiple displays.
> 
> If I disable the second display, everything seems to work properly;
> re-enabling the second display causes the widgets to move back to the
> incorrect positions.

Good catch!

FWIW, I don't know if it's reasonable in your config or not, but at least 
based on my kde4/plasma1 experience, a bug of that type (positioning-
related, multi-monitor triggered) is often one-dimensional in regard to 
the orientation of your multi-monitor layout.

For example, if you have a side-by-side multi-monitor layout and the 
positioning bug occurs on a horizontally oriented panel (as a top or 
bottom panel would be), try changing one or the other, either switching 
the monitor layout to stacked, or the panel to a side instead of top or 
bottom.

In some cases this sort of layout workaround strategy can even work to 
overcome bugs when the software components aren't related, but the 
positioning just happens to be.  Here, for example, I run xorg's xf86-
input-mtrack multi-touch pointer driver, which at least back when I first 
installed it after getting my touchpad recognized, had an off-by-one-
pixel pointer logic error, that only triggered in the vertical, not 
horizontal, so I couldn't actually reach the bottom row of pixels on the 
display when running that driver.  In most cases that'd be just fine, 
since buttons, etc, tend to be more than 1 px high, but it broke plasma's 
panel popup if the panel was set at the bottom of the screen, hidden 
until the pointer reached it, since the pointer could never reach the 
bottom row of pixels required to trigger the popup, due to that
off-by-one pointer driver bug.

I had a small popup panel situated to the right at the bottom of the 
screen when I hit that bug, so my plasma panel was affected even tho it 
was the input driver with the bug.

Of course the most direct solution (other than a patch to the driver 
itself if I was a dev...) would be to give up on that driver and try say 
the much more mature but lower featured synaptics driver
(xf86-input-synaptics), but I really wanted access to the extra gesture 
recognition that the mtrack driver had.

Alternatively, at least in theory I could have tried a switch to a
top-panel instead of a bottom-panel, but my muscle-memory was already set 
for it at bottom-right, and besides, I already had dedicated my top 
monitor (of three in stacked layout) to some other task (a superkaramba 
system-monitor dashboard theme) and didn't want to disturb that.

But as it happened, I didn't have anything on it that really /needed/ to 
be horizontally oriented, so I was able to move the panel from the right 
side of the bottom of the screen, to the bottom end of the left of the 
screen, adjust its height and width somewhat so the plasmoids fit the new 
orientation a bit better, and continue with very little change to my 
actual behavior, since now instead of aiming for just left of the bottom 
right corner, I aimed just above that bottom right corner, to have the 
panel popup.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

___________________________________________________
This message is from the kde mailing list.
Account management:  https://mail.kde.org/mailman/listinfo/kde.
Archives: http://lists.kde.org/.
More info: http://www.kde.org/faq.html.


More information about the kde mailing list