[ On Thursday, April 29, 1999 at 00:55:10 (-0700), Brian D Chase wrote: ]
> Subject: RE: CVSup collections for a NetBSD CVS tree
>
> A nice benefit of porting CVSup to C is that people don't have to get very
> deep into worrying about platform specific details (as they would with
> porting a Modula-3 compiler).  It also means that if `n' increases,
> there's no additional cost to having CVSup run on that platform.  It
> scales better.  

Porting CVSup to C would require maintaining two separate versions.
While this isn't such a difficult job if one person or a well
co-ordinated group of people maintain both implementations
simultaneously, or if the protocol specification is very static and
maintained by a third party.  However when one person controls both the
primary implemenation as well as the protocol specification, it takes a
lot more dedication and effort to maintain a separate implementation.

That's not to say this wouldn't be a good thing to do, just that you'd
better have either a) lots of energy and skills, or b) a tremendous and
overpowering need to do it, or c) both.  I don't see either poppping up
around here, never mind both....

As I understand it the protocol requires a multi-threaded server and
client too, so doing this in C in particular might not be very easy or
smart.  Any other language you pick is also called "NotC"!  ;-)

Use it if you can, don't if you can't, port M3 to your platform if you
must.  The free M3 compilers already work on the majority of the same
hardware platforms that NetBSD works on, and then some -- it's just a
matter of polishing and integrating NetBSD support.  Since the primary
free compiler comes from DEC there might even be an unsupported VAX
compiler in a back room somewhere....

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <gwoods@acm.org>      <robohack!woods>
Planix, Inc. <woods@planix.com>; Secrets of the Weird <woods@weird.com>