Re: [PATCH] Detect key modifier status in 'sleep --interruptible'

[Colin Watson]

On Mon, Aug 24, 2009 at 12:27:45AM +0200, Robert Millan wrote:
> On Wed, Aug 12, 2009 at 04:35:21PM +0100, Colin Watson wrote:
> > +static int
> > +grub_at_keyboard_keystatus (void)
> > +{
> > +  /* FIXME: I don't know if getting key modifier status is possible without
> > +   * BIOS help.  */
> > +  return 0;
> > +}
> > +
> 
> What we do in these cases is leave the function pointer undefined, which
> means NULL value (and check this in the caller of course).

OK.

> But in at_keyboard it's definitely possible (check how we handle e.g. CTRL
> or SHIFT keys).

No, that code only spots make scan codes arriving after GRUB's terminal
starts up. AFAICS it has no way to tell whether e.g. Shift was held down
already when GRUB started (except through the vagaries of key repeat),
which is important here.

-- 
Colin Watson                                       address@hidden