[1540] in Kerberos_V5_Development
Re: lib/rpc breaks on non-ANSI compilers
daemon@ATHENA.MIT.EDU (Barry Jaspan)
Mon Aug 12 20:18:35 1996
Date: Mon, 12 Aug 96 20:18:53 -0400
From: "Barry Jaspan" <bjaspan@MIT.EDU>
To: tlyu@MIT.EDU
Cc: krbdev@MIT.EDU
In-Reply-To: <199608130008.UAA24472@dragons-lair.MIT.EDU> (message from Tom Yu
on Mon, 12 Aug 1996 20:08:24 -0400)
Because of unconditional prototypes, stuff in the lib/rpc headers
doesn't work. I'll be working on this later this week.
Uh oh. I knew this would come up eventually. :-/
lib/rpc isn't the only thing that is going to break on non-ANSI
compilers. All of KADM5 has ANSI function definitions. The code also
uses ANSI constant string concatenation freely.
News flash! It is 1996. All reasonable excuses for using K&R C
compilers expired years ago. gcc is more portable than Kerberos. I'm
confident that Windows compilers are all ANSI compliant. Why should
we expend effort to support obsolete compilers?
I suppose someone will respond "because the foobar platform's native
compiler is non-ANSI." Can someone name a site that uses or wants to
use Kerberos 5, that does not already have gcc or another ANSI
compiler, that would refuse to get gcc in order to run Kerberos 5, and
is not such a backwards and confused site that they are best hand-held
by a commercial vendor rather than an overloaded and understaffed
free-software development team?
Barry
