Especially for us poor souls in the slow-computer end of the netbsd 
world, i think it is definitely worth it to revive pcc in NetBSD/vax. 
Because of things like netbsd now being elf, does anybody know of any
reasons why pcc would not be easy to make work with the gnu binutils
for purposes of as, ld, etc? 
btw, it might be the case that pcc _is_ covered by the usl license: 
4.4BSD was the one that started using gcc, but still has pcc in there 
too. Net2 has gcc but does not have pcc. Weve got to do something about 
this though, it is really ridiculous to expect people to wait for four, 
five, ten, twelve hours to compile a kernel. Yes, there is more code. But 
if a uvax-II can't compile a kernel in 2 hours, there are serious 
problems.. 

Nobody maintains pcc, but it is still available if you have a unix source 
license, which is free from caldera now, from the PUPS archive 
on minnie.tuhs.org

isildur

On Wed, 29 May 2002, Chuck McManis wrote:

> At 09:56 AM 5/29/02, Chris Wareham wrote:
> I've often wondered why gcc replaced pcc in the *BSD's. I assumed
> >it was licensing reasons, or that gcc is more actively maintained.
> >Does anyone maintain pcc at the moment?
> 
> Because people were working on gcc and no one was working on pcc.
> Further gcc is a "suite" of tools, everything you need to compile
> and link, whereas pcc needed ld and as to make binaries.
> 
> Gcc is the FSF's way of proving that Microsoft isn't the only development 
> organization that can write huge bloated applications (HBAs). And not 
> surprisingly they are HBA's for the same reasons, backward compatibility, 
> thousands of corner cases, exceptionally wide scope, and waaaaaay to many 
> programmers.
> 
> The concept was great, code generator to "standard" intermediate form, 
> intermediate form to platform specific form, linked into a platform 
> specific binary. But the special cases and "features" have gotten so out of 
> hand as to be ridiculous.
> 
> --Chuck
> 
>