On Wed, 10 Jan 1996, Jason Thorpe wrote:

> On Wed, 10 Jan 1996 00:11:18 -0800 
>  bwalker@musings.com (Brad Walker) wrote:
> 
>  > Could someone please explain to me why we are concerned if we maintain
>  > compatibility with K&R.. I must have missed that argument. And it's not
>  > a good enough argument to say that we need to maintain old code. I would
>  > like to hear the why argument and not the because argument.
>  > 
>  > Why not mandate all new code be ANSI.. And work on porting the old stuff.
> 
> We're beating a dead horse, but... :-)
> 
> Put quite simply, code that utilized K&R-style function decls is just 
> "more portable".  I quite often rip code out of NetBSD and compile it 
> under SunOS (with the native compiler). 
I guess that you meen yesteryears SunOS 4, that does not come with an
ISO/ANSI Std. c compiler, and whose libc does not comply with the 
Standard C Library ?!
The current SunOS (version 5, part of Solaris 2) does not even come with
a c compiler, but the optional c compiler product does have an ISO/ANSI
std. compiler. 
>                                         And there just might be a time 
> in the future when someone with a K&R compiler needs to bootstrap NetBSD 
> on some new platform...it could happen...
As mentioned before: Is this probability significantly greater than 0?
Is there any realistic posibility of porting NetBSD without porting
gcc first?
> 
> Of course, I don't really want to "mandate" anything.  I'd much rather 
> *encourage* writing clean, portable code.

Yes, I heartily and strongly agree.

But I consider ISO/ANSI C standardized code to be the most portable code. To
put it provocatively, ISO/ANSI Standard C enable you to have code that stays
portable in the future. Compatibility with K&R C is compatibility with the
past. 

Finally I will cite two small sections from ISO 9899, the C standard:

  6.9.4 Function declarators
  The use of function declarators with empty parentheses (not prototype-
  format parameter type declarators) is an obsolescent feature.

  6.9.5 Function definitions
  The use of function definitions with separate parameter identifier and
  declaration lists (not prototype-format parameter type and identifier
  declarators) is an obsolescent feature.

And as far as I recall from the discussions of the new revision of the C
standard (C9x), the K&R syntax for function declarations and definitions
may even be removed completely from the standard in its next revision.

regards
Erik Bertelsen