[342] in Kerberos_V5_Development
Re: Mystifying problem with function prototypes....
daemon@ATHENA.MIT.EDU (John T Kohl)
Tue Oct 16 21:13:23 1990
Date: Tue, 16 Oct 90 21:12:44 -0400
From: John T Kohl <jtkohl@ATHENA.MIT.EDU>
To: Ralph Swick <swick@ATHENA.MIT.EDU>
Cc: John T Kohl <jtkohl@ATHENA.MIT.EDU>, krbdev@ATHENA.MIT.EDU
In-Reply-To: Ralph Swick's message of Tue, 16 Oct 90 17:08:47 EDT,
> Date: Tue, 16 Oct 90 17:08:47 EDT
> From: Ralph Swick <swick@ATHENA.MIT.EDU>
> It may require some gross
> X-style #defines for prototypes, or the disabling of prototypes for such
> a build environment.
> If you mean that a gcc user won't be able to use prototypes right
> now, then I think this needs some attention.
Right now, if you use GCC on the VAX the right thing will happen. If
you use CC on the VAX, you will lose if you call any of the functions
with narrow type arguments from your own code.
The hacks that need to be done would enable the building with ANSI C of a
library that can be called from non-ANSI C programs. The cleanest way
to do that with the current code is to use old-style function
definitions and no prototypes; what I want to be able to do is have the
types in the prototypes automagically widened if requested at
compile-time. This is where the gross hacks would be needed.
John
