lynx-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: lynx-dev Weird #defines for OS/390


From: dickey
Subject: Re: lynx-dev Weird #defines for OS/390
Date: Wed, 2 Dec 1998 22:56:16 -0500 (EST)

> 
> Hello, Lyncei, 
>  
> Heretofore, I've done my OS/390 Lynx compilations with the Legacy C 
> compiler.  Now, I'm trying the Posix compiler (actually the same 
> compiler with different options.)  To make this work I must do: 
>  
>     CC=c89 \ 
>     CFLAGS="-D_OE_SOCKETS -D_ALL_SOURCE" \ 
>         ../configure                 
>  
> I'm wondering how much of this I can reflect back to the 
> source tree.  Since I find the following: 

But doesn't 'cc' work?  (no matter what TOG says, in practice c89 is an
alternate name - most of us work on "legacy" code).

Anyway - my experience with turning on POSIX code is that most vendors
interpret it to mean that ANSI library functions are not recognized, and
vice versa which is why you end up contorting the ifdef's to get things to
compile.  So I gave up on that - there's not very many interesting programs
that are purely ANSI or purely POSIX.

>                  The Single UNIX ® Specification, Version 2 
>                        Copyright © 1997 The Open Group 
>      __________________________________________________________________ 
>   
>      c89 - compile standard C programs 
>      __________________________________________________________________ 
>   
>      cc - a C-language compilation system (LEGACY) 
>      __________________________________________________________________ 
>  
> I might reasonably claim that "configure" should make "c89" its 
> second choice for a compiler name, after "gcc", but before "cc". 
>  
> the "-D_OE_SOCKETS -D_ALL_SOURCE" is pure idiosyncratic nonsense; 
> I've complained about it to IBM tech support; it shouldn't be 
> necessary.  Certainly, I could craft tests in "configure" which 
> would fail if the two definitions were necessary, but I'm reluctant 
> to further burden configure for the sake of one target system. 
> The alternative is to add the defines to CFLAGS if host_os=os390, 
> but this is the very sort of ad-hoc coding that configure is 
> designed to avoid. 
>  
> Any opinions? 
>  
> Thanks, 
> gil 


-- 
Thomas E. Dickey
address@hidden
http://www.clark.net/pub/dickey

reply via email to

[Prev in Thread] Current Thread [Next in Thread]