🚀 go-pugleaf

In article <9307211404.AA113162@killroy>, Rob.MacKinnon@bsc.NO (Rob MacKinnon) writes:
> I am encountering a very strange bug with gcc when compiling C++ code. Has
> anyone else seen this behaviour below and might be able to tell me how to
> fix it?
> 
> Environment: IBM RISC System/6000, AIX 3.2.3e, gcc 2.4.5, g++ library 2.3.1
> 	     the assembler on these systems are properly fixed.
> Symptom:
>         /usr/local/bin/gcc -D__ANSI_CPP__ -DAIXV3 -DAIX -Dvsyslog=syslog -DSYS_FCHOWN=1 -DSYS_FCHMOD=1 -I. -I../port/aix32 -I../util -I../libtiff -g  -o faxinfo faxinfo.c ../util/libfaxutil.a ../libtiff/libtiff.a ../iv/libivhack.a -lg++ -liberty
> 0706-317 ERROR: Unresolved or undefined symbols detected:
>                  Symbols in error (followed by references) are
>                  dumped to the load map.
>                  The -bloadmap:<filename> option will create a load map.
> .realloc__FPvUl
> .malloc__FUl

etc.

[...]

These are all straight C calls that have been mangled into C++.  The only
fix I've found -- and I haven't looked that hard -- is to:

extern "C" {
  realloc(...);
  malloc(...);
  /* etc. */
}

There's probably an include file somewhere that does this, but I can't find
it and neither, apparently, can gcc.  Might be a mung-up in g++ installation:
I get a couple of non-fatal error messages when I build libg++, but since I
haven't had time to really poke at it, I haven't done an official bug-report.

> 	.
> <output truncated for brevity>
> -- 
> Robert MacKinnon,    UNIX Support | Internet/Email: robmack@bsc.no
> Bergen Environmental Sciences and | 		    robmack at bergen
> Solution Centre,  Bergen,  NORWAY | Phone:          +47-5-544618
> ********************************************************************************
> p'ch'zarni'chiwkov:  the nasty little sound of a sword being unsheathed right
> behind one at just the point when one thought one had disposed of one's enemies.
> ********************************************************************************

-- 
IBM pays absolutely no attention to my
opinions, ergo the foregoing cannot	
possibly represent its position.	
					
					
Christian H. L. Moller