🚀 go-pugleaf

   Return-Path: help-gcc-request@prep.ai.mit.edu
   To: help-gcc@prep.ai.mit.edu
   Date: 6 Mar 92 06:48:59 GMT
   Organization: Solbourne, User Interface Group
   From: imp@solbourne.com (Warner Losh)
   Sender: help-gcc-request@prep.ai.mit.edu
   References: <A33092@AC.maus.de>, <TOMW.92Mar4123856@latchup.ucsb.edu>, <1992Mar6.020554.22495@bond>roy
   Subject: Re: gcc-2.0 name mangling
   
   
   
   While we're on the subject of name mangling, I thought I'd ask the
   following question.  I think I know the answer, but I haven't seen any
   direct statement:
   	Is g++ 2.0 "compatible" with cfront as shipped from AT&T (or
   sun) on a sparc?  By compatible, I mean can I mix and match objects
   from the the two of them and expect things to work?
   
   I gather the answer is no, they are only compatible on the source code
   level, but have different name mangling, virtual function table
   assignment, and other incompatibilities so I must either use cfront or
   g++, but not both.
   
   Some simple experiments seem to indicate that this is the case.  I'd
   like to know if this is a bug in some planned compatibility, or if it
   is beyond the scope of g++ and I shouldn't try to mix the two.
   
   Thanks a bunch for your time,
   
   Warner
   -- 
   Warner Losh		imp@Solbourne.COM	  MMP to DoD #882
   "You better stay away from him, he'll rip your lungs out Jim,
    He's looking for James Taylor" -- Warewolves of Boulder WZ

I just send mail to help-g++ about this.  cfront and g++ do not build
objects the same way, nor should you expect them to.  There is no ANSI
effort whatsoever to try to reconcile this fact.  You must be happy
with *source* compatibility only.  If you want more information about
why this must be, read all about from help-g++.

Michael