🚀 go-pugleaf

In article <1992Oct27.162538.10288@next.cambridge.ma.us>  
simsong@next.cambridge.ma.us (Simson L. Garfinkel) writes:
> The fact is that NeXT has a portable interface for the keyboard, and the  
> programmers at the corporations that are having problems with the keyboard  
just  
> didn't read the documentation.
> 
> (from event.h:)
> 
>     struct {    /* For key-down and key-up events */
>         short   reserved;
>         short   repeat;         /* for key-down: nonzero if really a repeat  
*/
>         unsigned short charSet; /* character set code */
>         unsigned short charCode;/* character code in that set */
>         unsigned short keyCode; /* device-dependent key number */
>         short      keyData;     /* device-dependent info */
>     } key;
> 
> If you used key.charSet and key.charCode, you win.
> If you used key.keyCode and key.keyData, you lose.
> 
> Its just that simple.  WordPerfect did it wrong.  Co-eXist did it wrong.   
> Executor did it wrong, but then they fixed it (an easy fix). SBook 2.0 did it  
> wrong, but SBook 2.05 does it right.
> 
> 
> In article <1992Oct26.221617.7372@wdl.loral.com> mdm@wdl1.wdl.loral.com (Mike  
D  
> Marchionna) writes:
[stuff deleted]
> > Write it off to NeXTSTEP development arrogance.
> > 
> 
> Write it off to not reading the documentation.
> -- 
> 
> .................................................................simson

Boy this is fun!

I just wanted to jump in here to explain some things about the ADB keyboard and  
co-Xist.  We were only aware of the ADB systems coming out after the last  
release of co-Xist (2.1.2) in April of '92.  However, we have always been aware  
that there was a device independent way to get the key that was pressed.  There  
is a problem with this, though.  Here it is:
	If any of you have worked with X servers, you know that X likes to get  
raw keycodes.  The platform initialization routines are responsible for mapping  
the raw keycodes into the appropriate character codes through a lookup table.   
We tried using a lookup table that used the character codes instead of the key  
codes.  This worked OK until you tried to an Alt with a key.  The character  
code would change, but X would interpret it incorrectly because it saw the Alt  
key down and tried to do its own conversion.  I'm sure there is a way to make  
this work properly, but it would probably require some significant work in the  
keyboard mapping code.  
	Another problem is that X likes to discern the difference between a left  
Shift or a right Shift, etc.  There is no device independent way to do this  
with the NXEvent struct.  The device independent flags only tell you if a Shift  
key is down, not which one is down.
	Of course, in co-Xist 3.0 this has been remedied with a keymap table  
dynamically configured at run time.  The keymap table has also been adjusted to  
allow xmodmap (X keymap modification) to change the keymap table to anything  
you want.
	We apologize for any inconvenience this has caused.  If any co-Xist user  
has any questions or concerns regarding ADB systems and co-Xist, please feel  
free to call or e-mail me.

Chris Chauvin
Product Manager - co-Xist
Pencom Software
chauvin@pencom.com
(512) 343-1111