[Wayland-bugs] [Bug 754110] inconsistent interpretation of preserved modifiers with xkbcommon

[gtk+ (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=754110

--- Comment #29 from Daniel Stone <daniel at fooishbar.org> ---
Yeah, looking at things just as they stand, I agree that Carlos's patch is
probably fine, in that it fixes things for just the CTRL+ALT type. However,
it's also pretty nasty, and I can nearly guarantee it will have unforseen
issues later on down. We now have a bunch of pretty weird small keyboards which
rely heavily on multiple modifiers for chording - the full 80s Emacs experience
- so I worry that encoding this will just cause trouble later on down the line.

I really like Owen's conclusion here:
>  - consumed_modifiers contains modifiers that effect the translation
>   of the key *that are found in event->state*

although I would perhaps reword as:
> - consumed_modifiers contains modifiers that directly affected the translation of the last keypress

and I think that would be something good to encode generically through
xkbcommon. Whether that's something we do by default (i.e. break current
interface and also legacy XKB's definition of 'consumed'), or something we
require people to opt in through an xkb_context flag for is up for discussion.
My suggestion though, would be doing it by default, with an xkb_context flag to
_disable_ (i.e. force old behaviour), which happens to work fine as we don't
currently check for unsupported flags in xkb_context_new. :\

If you're fine with that, I can get the patch together and test with GTK+,
since this hits me every day.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/wayland-bugs/attachments/20160121/2b926865/attachment.html>