Re: compiling ncurses on Darwin / Mac OS X

[Thomas Dickey]

On Mon, Jun 18, 2001 at 08:29:48AM -0700, Eugene Lee wrote:
> On Mon, Jun 18, 2001 at 10:58:06AM -0400, Thomas Dickey wrote:
> : 
> : I check for GNU extensions, but don't use (or rely) on them if they're
> : not present.  About all I know about Object C is that it uses a different
> : verb for including files - but iirc, it's supposed to accept #include
> : as well.  Perhaps Apple's compiler happens to have an option by the
> : same name, but this is what gcc's info says:
> 
> [...]
> 
> : (While I've made an attempt to work around broken cpp's such as on AIX 3.x,
> : which implement a mix of K&R/ANSI, it's not certain that ncurses would
> : work properly if you set the cpp to "-traditional-cpp" even if it compiled).
> 
> Hmmm.  Good point.  If you want, I can do a clean build of ncurses with
> the latest patches and capture the output (including errors).  Where
> should I mail that output?  To you directly, the bug-ncurses list, both?

to me (but I'm going out of town til next Wednesday, so I'll look at it
then).
 
> : > I ran into similar problems with compiling Vim pre-5.8, which required a
> : > similar fix.  A quick search turned up tons of similar fixes from people
> : > porting to Darwin.
> : 
> : Vim's not written in ANSI C (it's extended C, which is a bit different).
> 
> <RANT>
> When I asked them about problems compiling on Mac OS X, one of the Vim
> developers responded that it was plain old ANSI C.  I didn't think their
> answer was correct, but I discovered the -traditional-cpp flag from a
> Darwin list and dropped the issue.  Those liars!  :-)
> </RANT>

They probably don't understand the difference (Bram does, and has put some
comments in the code to that effect).

This is ANSI C:

        int foo(int a);
        int foo(int a) { return a; }

This is extended C:

        int foo(int a);
        int foo(a) int a; { return 1; }

-- 
Thomas E. Dickey <address@hidden>
http://dickey.his.com
ftp://dickey.his.com