Ian Dall <Ian.Dall@dsto.defence.gov.au> wrote:

> Phil Budne <phil@ultimate.com> writes:
> 
>   > The other was that the pc532 ld.aout_so md.c generated warnings that
>   > killed the compilation.  Enclosed are diffs for the necessary changes.

Are these changes new?  (Changes being in -current, not your patch,
of course.)  I didn't have any problem with this when I built -current
from Nov 29 sources.  Admittedly, that's not quite -current anymore...

> Are new style function definitions allowed in NetBSD yet? I recall
> some debate about this, but don't recall the outcome.
> 
>   > What's the current word on using egcs to build stuff (esp. the
>   > kernel)?  I've been using the gcc version 2.7.2.2+myc1 that came with
>   > the 1.3 binaries.
> 
> I've been using egcs for some time. I have never rebuit the world, but I have
> rebuilt the kernel, usr.lib and all the programs which depend on libkvm.
> I'd regard egcs as fairly well tested, but my egcs is 1.03 I think, not
> whatever is in the current netbsd tree. If you could build workd with it
> it would be a good confidence builder.

I've been using it for a couple of months, and built the world
several times.  A few months ago, I couldn't get lint to do it's
thing, and I wasn't sure if that was a problem with -current, what
I was doing, or a bootstrap issue.  Whatever it was, it's gone now.
The only trouble I've had recently is the BIND/T_SLAVE thing.
Everything else seems to build and work just fine.

Anyone tried gas.new?  Any benefits to doing so, now or in the future?

Jon