Keygrabbing not working for some keysyms (correction)

[Owen Taylor]

On Tue, 2006-01-31 at 07:45 -0500, Matthias Clasen wrote:
> On Tue, 2006-01-31 at 13:26 +0100, Pierre Ossman wrote:
> 
> > Ahhh... That seems to be it. Print shares the button with SysRq and
> > XKeysymToKeycode returns SysRq:s keycode (92), but the keyboard sends
> > Print's (111).
> > 
> > So what can I have done to bork this? (It's "correct" on my other machines)
> > Also, what would be the API for getting all mappings. XKeysymToKeycode
> > only gives a single keycode.
> 
> The Print<>SysRq issue is a known bug in the XKB data files shipped with
> xorg: 

I'm a little confused here. The XKB mapping data only deals with the
mapping of keycode to keysymbol. (*) The original problem (was there
a regression here?) was that the wrong keycodes were being generated,
which had to be fixed at the driver layer. (**)

Regards,
						Owen

(*) Well, it's not that simple ... there also is physical key to
   keycode translation in the XKB layer, but I don't think it can
   be used to fix this issue, and isn't what is discussed in the
   Red Hat bug report.

(**) PC hardware sends two different scan codes for Print Screen/Sysrq
    and for Pause/Break. These need to be mapped back together to 
    match the X keyboard model.