> > > > I wrote it off as a hardware problem (still do). > > > hmmm. ... it may be ... is there a standard showing what data to expect from a >keyboard ?
> > I have tried to ask the same thing two times. Of course > > it is a hardware problem but how can you prove it and I don't believe this is necessarily a hardware problem ! Whilst I agree that a piece of harware behaving differently from what we expect is causing the problem this does not mean that the keyboard is operating outside specifications. It could that invalid assumptions on the part of the programmer are responsible (nothing personal to whoever wrote the keyboard code, I'm pretty sure this is a trap I would easily fall into myself). ie. Obviously some keys need to be received without intervening keypresses (EG. Ctrl-A, Ctrl-E, ...). It seems plausable that the programmers of this code just thought it was impossible for someone to press the same key twice without releasing it first and therefore didn't test for a key release. Also even if it is a hardware problem should we check that the input from the hardware is valid ie. conforms to the contract ? Where do we draw the line here ? Another thing that interests me is: Why doesn't this happen outside X ? What is the difference between the way the kernel handles key presses and the way XFree86 handles them ? Anyone know ? NB. There was talk by Andrae Pitis of putting a reject multiple press of the same key into the kernel code but AFAIK Linus rejected it (Andrae originally had ridiculous delays of 250ms which made it hard to type things like 'look', 'sleep', ...). However the Kernel Release statement for 2.4 lists an 'm68k keyboard delay/repeat' problem being patched so maybe that's why this only happens in X ie. I'm using 2.4.2 (although I'm using x86 ????) Maybe someone can confirm this ? Wilson _______________________________________________ Xpert mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/xpert
