[2396] in Kerberos_V5_Development
Re: Another Kerberos resource
daemon@ATHENA.MIT.EDU (Theodore Y. Ts'o)
Wed Jun 18 21:40:02 1997
Date: Wed, 18 Jun 1997 21:39:38 -0400
From: "Theodore Y. Ts'o" <tytso@MIT.EDU>
To: Marc Horowitz <marc@cygnus.com>
Cc: "Theodore Y. Ts'o" <tytso@MIT.EDU>, krbdev@MIT.EDU
In-Reply-To: Marc Horowitz's message of 18 Jun 1997 20:51:41 -0400,
<t53bu53mn36.fsf@rover.cygnus.com>
From: Marc Horowitz <marc@cygnus.com>
Date: 18 Jun 1997 20:51:41 -0400
>> In fact, I'd just as soon recommend that people use gcc (or some
>> other ANSI- compliant C compiler) instead.
Did you really mean that? You might not recognize kerberos when I got
done ;-)
I simply chose to not spend resources fixing that specific piece of
compiler brain-damage.
On the other hand, I'm not convinced it adds value to spend a lot of
time revamping our tree just so the native SunOS compiler will fail to
build Kerberos (it might be good for Cygnus gcc sales, though :-). On
the other hand, using K&R declarations and PROTOTYPE macros surely isn't
that hard.
In the long run, eventually it'll be fair game to assume ANSI C, and I'm
willing to grant that the day when we can do that is becoming ever
nearer. However, if there is some modern platform which suffers from a
native compiler bug, and it isn't too heinous to work around it, I'll
chose to include the work around rather than to say, "NYAH, NYAH, USE
GCC!". It's a case-by-case, individual judgement sort of thing.
- Ted
