[1781] in Kerberos
Re: Kerberos V4 patches for NeXT?
daemon@ATHENA.MIT.EDU (Alan Crosswell)
Thu Feb 27 00:33:45 1992
Date: Wed, 26 Feb 92 23:58:29 EST
From: Alan Crosswell <alan@curta.cc.columbia.edu>
To: Marc Horowitz <marc@mit.edu>
Cc: kerberos@Athena.MIT.EDU
In-Reply-To: Your message of Wed, 26 Feb 92 23:12:19 EST
> Well, the machine with my NeXT sources on it died, and isn't back up
> yet, so I'll try to quote from memory.
>
> 1) find all instances in all Imakefiles and templates of single quotes
> in make comments, and remove them. This will solve the "unterminated
> string constant" problem.
Unterminated string constant can apparently also be fixed by making
sure imake invokes cpp with -traditional. I did this with a shell
script which I point to with the IMAKECPP environment variable.
#!/bin/sh
exec /lib/cpp -traditional $*
The problem is that without the -traditional switch, cpp doesn't
do the /**/ token concatention trick properly. I guess removing
the single quotes may do the same thing?
A/**/B -> A B rather than AB.
> 2) compile -traditional
You need -bsd too; Some of the standard includes do things like
use 'const' in declarations *unless* _STRICT_BSD_ is #defined.
So, -traditional turns off recognition of 'const' but the lack
of -bsd causes it to get used in standard include files.
> 3) I don't remember any ss problems
make_command gives a variety of bogus arguments to /bin/ld, including
-n and combining -r and -s. [* flame The code in make_command is
conditionalized quite poorly for this very system-dependent thing. *]
Maybe the next programmer who comes along will think to make the
correct ld args a #define or something that can be tweaked in the Makefile.
The one I'm on now is a syntax error in des_internal.h, which I have
a feeling is caused by osconf.h not knowing that a NeXT is an m68k....
So, I'm not out of the woods yet and would greatly appreciate it
if when your file server comes back you could dig up your changes
and send them along (hopefully it's not down for the count!)
Thanks.
/a
