On Tue, Oct 19, 2010 at 08:39:05PM -0400, Kenneth R Westerback wrote:
quoted text
> On Tue, Oct 19, 2010 at 10:03:32PM +0200, Miod Vallat wrote:
> > > Index: pms.c
> > > ===================================================================
> > > RCS file: /cvs/src/sys/dev/pckbc/pms.c,v
> > > retrieving revision 1.10
> > > diff -u -p -r1.10 pms.c
> > > --- pms.c	17 Oct 2010 10:32:00 -0000	1.10
> > > +++ pms.c	17 Oct 2010 11:30:24 -0000
> > > @@ -134,21 +134,10 @@ pmsprobe(struct device *parent, void *ma
> > >  	/* reset the device */
> > >  	cmd[0] = PMS_RESET;
> > >  	res = pckbc_poll_cmd(pa->pa_tag, pa->pa_slot, cmd, 1, 2, resp, 1);
> > > -	if (res) {
> > > +	if (res || resp[0] != PMS_RSTDONE || resp[1] != 0) {
> > >  #ifdef DEBUG
> > > -		printf("pmsprobe: reset error %d\n", res);
> > > -#endif
> > > -		return (0);
> > > -	}
> > > -	if (resp[0] != PMS_RSTDONE) {
> > > -		printf("pmsprobe: reset response 0x%x\n", resp[0]);
> > > -		return (0);
> > > -	}
> > > -
> > > -	/* get type number (0 = mouse) */
> > > -	if (resp[1] != 0) {
> > > -#ifdef DEBUG
> > > -		printf("pmsprobe: type 0x%x\n", resp[1]);
> > > +		printf("pms: reset error %d (response 0x%02x, type 0x%02x)\n",
> > > +		    res, resp[0], resp[1]);
> > >  #endif
> > >  		return (0);
> > >  	}
> > 
> > I disagree with this change (yes, I know it has been commited). Either
> > you do not use option DEBUG, and your diff causes no change in the
> > generated code, or you do, and your diff makes the error message less
> > readable. I do not see any benefit in this.
> > 
> > Miod
> 
> I prefer the new error message and find the newer code easier to
> read, but obviously a matter of taste. If you have really strong
> feelings we can revert since it doesn't make functional changes.
> 
> .... Ken

Thanks, Ken. I even did not have to translate :)
This is about the same that I wanted to say.

-- 
Alexandr Shadchin