🚀 go-pugleaf

Hi, GNUrus!

I'm having a whacking annoying problem using gcc/g++ on an RS/6000
running AIX 3.2.5.  The version of the compiler I have installed
(2.6.0) was built on another RS/6000 (370) and installed on this one
(320).  I'm trying to use g++ to compile my source which uses the
socket routines.  Unfortunately, g++ cannot seem to get things
arranged so the socket routines will resolve at link time.  Everything
else works fine, just the socket routines are broken.  

In order to save disk space, I did not install the AIX c compilers on
the 320.  The libraries and include files are there, but the compilers
are not.  Also, the libraries are the c libraries not the c++
libraries.  I suspect this may have something to do with the problem,
however, if it does, why does it only affect the socket routines?  If
not, why don't the socket routines work?  'Tis puzzling.

If I compile in C++ mode, g++ barks that the function socket (int,
int, int) is undefined and my options are socket (), or socket (
struct socket * ).  To my knowledge, the two "available" definitions
are not valid declarations for the socket call.

On the other hand, if I stay in C++ mode but tell g++ that the socket
routines are actually C routines, then g++ works fine.  Unfortunately,
ld starts barking and says it cannot resolve the symbols
.socket__Fiii, .bind__FiP8socketaddrUl, .listen__FiUi,
.accept__FiP-sockaddrPDUl, .recv__FiPvUli, and .send__FiPvUli.  Huh?
I've never heard of 'em.  Obviously, some internal routines, but why
would g++ cause ld not to find them?

So, since g++ barks and growls, I tried gcc and converted the program
to pure ANSI C.  This isn't really an option for me, I just thought I
would try.  Now gcc barks and says it does not know what the variables
sockaddr_in or sockaddr are, both of which are known by g++.  Again,
huh?  I use the same include files.  

I've been writing C for 10 years, C++ for three and I've flown AIX
since the days of 2.2.1 on the RT (Uugh!).  But...this one has me
stumped.  If anyone can give me any suggestions, I would really
appreciate it--even if it is only RTFM (actually, I would be ecstatic
if it were to be RTFM, but I think it is a bit more complicated than
that).  By the bye, these programs compile fine on the 370 under xlC,
but I have the same problem with the socket routines using gcc/g++.
Also, they work fine on an HP 9000/720 using the HP C/C++ compilers
that I am occationally forced, quite against my will, to use.  :-)  I
have not installed the gcc compiler on the HP as I do little
development on the wretched thing, so it would simply be a waste of time.

Please help...I'm really stuck.

Thanks in advance.

<---------------------------------------------------------------------------->
Todd Templeton         | All opinions are solely my own and do not reflect the
Lead Software Engineer | views of Phelps Dodge Corporation or its management.
<---------------------------------------------------------------------------->