bug#32977: 26.1.50; Emacs (on gnome 3, wayland, debian buster) calculates incorrect horizontal positions

[Robert Pluim <rpluim@gmail.com>]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> (This bug report unfortunately got no response at the time.)
>
> I can reproduce this bug in Emacs 28:
>
> (set-frame-position (selected-frame) 0 0)
> => t
> (frame-position (selected-frame))
> => (-10 . 66)
>
> If I do this, I can clearly see that the frame isn't outside the screen
> (i.e., I can see other windows between the Emacs frame and the border of
> the screen):
>
> (set-frame-position (selected-frame) 9 0)
> => t
>
> But still:
>
> (frame-position (selected-frame))
> => (-1 . 66)
>
> This is a Debian bullseye system with Gnome Shell.
>
> Anybody got an idea what's going on here?

This is under X, right? GTK/GDK's idea of window coordinates and
Emacs's idea sometimes disagree, especially when using Gnome Shell.

The answer is pure gtk, but in the pure gtk branch under Wayland,
frame-position will return (0 . 0) however the frame will not actually
move when you call set-frame-position.

pure gtk under X moves the frame, *and* returns (0 . 0), so I guess
this should be filed as a bug against the pure gtk branch.

Robert