> 
> hi,
> 
> just got netbsd up and running on an AXPpci board. no problems getting it up 
> and running, kudos to the team.
> 
> so, a few immediate apparencies. first: no telnet. second: skey is broke. the

Hmmm. telnet works for me on my AXPpci using not-so-current sources.
Don't know about skey.

> second of these two I can tackle ( and will :), but the first seems to be
> a symptom of a deeper-rooted problem. I hope everyone can forgive what might 
> be obvious, but I've tried. here goes.
> 
> gcc keeps complaining about, well a few things. like...
> 
> /usr/include/string.h:55: warning: conflicting types for built-in function `memc
> mp'
> /usr/include/string.h:56: warning: conflicting types for built-in function `memc
> py'

I don't see this problem. Are you using the tools that Chris distributes?
On my system...

[netbsd2(NetBSD):jb] gcc -v
Reading specs from /usr/local/lib/gcc-lib/alpha-unknown-netbsd1.0A./2.7-95q4/specs
gcc version 2.7-95q4

I haven't updated the tools lately, so I don't know what is the latest.

The only time I see "conflicting types for built-in function `memcmp'" is
when <string.h>  _hasn't_ been included.

> 
> basically, it thinks that the defs in the headers conflict with the builtins.
> I could use -fno-builtins, but that wouldn't get me much further. the defs in
> the header files seem to jibe with the defs in both versions of gcc source
> (cygnus and /usr/src/gnu), so I'm not sure why it's complaining. but it 
> bothers me that I don't know why.

Did you let 'make build' from /usr/src do the 'make includes' step with
a full source tree?

> 
> 
> the other, and more important problem gcc keeps barfing on is this:
> 
> commands.c: In function `tn':
> commands.c:2221: warning: address of register variable `__x' requested
> commands.c:2221: invalid `asm': invalid %%xn code
> 
> this shows up, from the looks of it, whereever any of the (htonl, htons, ntohl,
> ntohs) family are used. any clues?


I supped from the local NetBSD mirror this morning and my system has been
'make build'ing ever since. So far I've had only three problems:

1. libarch doesn't build.
2. The local mirror has its crypt files one directory level out w.r.t
   /usr/src/Makefile.
3. /usr/src/usr.bin/vi/common/pathnames.h needs '#define _PATH_MSGCAT    "./"'

The build has just done a make on telnet (but not an install yet), and
the compile was clean. Note that since I am outside the U.S. I don't
have /usr/src/domestic/usr.bin/telnet, so the directory that got built
was /usr/src/usr.bin/telnet. This might make a difference.

> 
> Regards,
> 
> Joe Battle
> battlej@webbed.com
> 
>           sin.sin_port = htons(sin.sin_port);
> 

Regards,

-- 
John Birrell                                CIMlogic Pty Ltd
jb@cimlogic.com.au                          119 Cecil Street
Ph  +61  3 9690 6900                        South Melbourne Vic 3205
Fax +61  3 9690 6650                        Australia
Mob +61 18  353  137