https://bugzilla.gnome.org/show_bug.cgi?id=755947
--- Comment #7 from Simon McVittie <[email protected]> ---
(In reply to Olivier Fourdan from comment #5)
> So, FWIW, I agree with this patch but it depends on bug 766860 on Wayland.
Olivier's patches to signal a simple boolean tiling state through Wayland have
now been applied.
(In reply to Matthias Clasen from comment #6)
> I agree that this patch will improve things under gnome-shell. But I'd
> caution that 'tiled' does not necessarily imply 'fixed size' - Jasper had a
> prototype for a tiling system that lets you resize tiled windows
Can we apply this one for now, and cross that bridge when we come to it? The
protocol by which GTK on Wayland learns about tiled windows is currently
GTK/Mutter-specific, so this isn't going to hurt anything on non-Mutter
compositors; and it's a rather annoying "papercut" when using half-screen
terminals under Wayland.
If someone adds per-edge "snapped" flags to the Wayland protocol later, we'd
want to change the logic to something like this pseudocode:
if (maximized || fullscreen)
flags |= (EXACT_WIDTH | EXACT_HEIGHT);
if (snapped_left && snapped_right)
flags |= EXACT_WIDTH;
if (snapped_top && snapped_bottom)
flags |= EXACT_HEIGHT;
...
if (flags & EXACT_WIDTH)
use the actual width requested
else
round width down to the next character-cell step
if (flags & EXACT_HEIGHT)
use the actual height requested
else
round height down to the next character-cell step
(That would also require splitting up GDK_HINT_RESIZE_INC into horizontal and
vertical parts, so that gdk_window_constrain_size() can do the right thing.)
--
You are receiving this mail because:
You are on the CC list for the bug._______________________________________________
wayland-bugs mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/wayland-bugs
