I have Apache OpenOffice (AOO) built with Xorg VCL.
I don't see that behaviour in LibreOffice built
with GTK (but then it's perceptibly sluggish,
see below why it matters). And I have nothing
else big/biggish readily available and built
with Xorg visual libs.
I start one instance of AOO (no need to open
files) and switch it between various
maximisation modes. It's easy to see when it
misbehaves - the navigation panel in the center
of the app's window stops recentering or window
stops redrawing.
So there MIGHT exist two code execution paths
that get broken I suppose. Not a developer, me,
though.
It doesn't matter if the switching is done with
mouse menu clicks or keyboard. I have long ago
defined keyboard shortcuts for fullmax, LH-max
and RH-max - M1+F10, M1+F11, M1+F12.
The behaviour can break on the very start of
AOO, if WM was remembering AOO's window being
set to half-screen.
But usually the broken behaviour is achievable
on at least the first maximisation mode change,
and sometimes it starts on the second or third
max mode change (I don't think I remember it
getting up later than that).
Could it be dependent on timings/race conditions
just a little?
P.S. Doing mode changes quickly does not cause
misbehaviour in the well-behaving snapshot of WM.
On 28/03/2026 00:35,
david.m...@gmail.com wrote:
> On Friday, 27 March 2026 at 10:40:51 UTC-4 Yury wrote:
> It was commit
> c82e6dad5cd2b9813f5df982fe9710429c24e81e
> put in on 'Fri Jan 23 23:38:32 2026 -0500'
> that broke things.