Re: PGTK-related misconceptions

[Po Lu]

Akira Kyle <akira@akirakyle.com> writes:

> Thank you for this patch! I recently updated and ran into this
> issue. From the discussion in this thread, I really don't understand
> why this change was made. Is this fixing some issue for some other
> non-wlroots based compositor?

Yes, see below.

> If it isn't, but was just made "to comply with the GDK docs" this
> would seem to be just be making life more difficult for those of use
> on wlroots based compositors. Emacs has many other places where it
> tries to work around quirks and other issues with GTK for the benefit
> of the user, so why would this be any different?

It is to prevent errors when Mod4 is _not_ the super key.  Very common
on X (where Mod4 is commonly Hyper), and in fact it is caused by Mod4
not being part of the Super virtual modifier in the compositor's XKB
configuration.

The conclusion I eventually came to is that the problem lies in the XKB
configuration of the Wayland compositor in question (namely, Sway), and
not Emacs nor GDK.  It works with GNOME Shell and kwin because they set
up their keyboards with the correct virtual modifiers by default.  So
what I'd recommend is to start an X server where the Super key does
work, run:

  xkbcomp $DISPLAY name/of/xkb/file.xkb

and then use that file with Sway.