Re: gcc-2.0 name mangling

Article View: gnu.gcc.help

Article #53206

#53206

From: tiemann
Date: Wed, 04 Mar 1992 13:55

21 lines
1455 bytes

    In article <A33092@AC.maus.de>, Marc_Van-Woerkom@ac.maus.de (Marc Van-Woerkom) writes:

    >te> My questions are: why does it do this?  Can I do anything to make them
    >te> work correctly together?

    > Name mangling is different on different compilers! (No bug - a fact of life)

    Huh?  This seems idiotic to me.  That means I can't use C++ libs
    compiled with a different compiler?  I'm outraged!  C doesn't have these
    kinds of problems.

Add this to your outrage: G++ has a different policy about where it
puts virtual function table pointers, virtual baseclass pointers, how
it lays out the virtual baseclasses of objects, and the interface to
constructors that initialize objects with virtual baseclasses.

C++ source compatibility is possible.  C++ object compatibility is
much less so---another reason I prefer source availability when I have
the choice.

Michael

Technical Details

Message-ID: <9203050555.AA14227@cygnus.com>

Path: arc1-ca.pugleaf.net!sparky!uunet!cis.ohio-state.edu!cygnus.com!tiemann

References: <TOMW.92Mar4123856@latchup.ucsb.edu>

Show Full Headers

Complete Article Headers

X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 11318c,d442a3a97625629c
X-Google-Attributes: gid11318c,public
X-Google-ArrivalTime: 1992-03-05 07:44:52 PST
Newsgroups: gnu.gcc.help
Path: arc1-ca.pugleaf.net!sparky!uunet!cis.ohio-state.edu!cygnus.com!tiemann
From: tiemann@cygnus.com
Subject: Re: gcc-2.0 name mangling
Message-ID: <9203050555.AA14227@cygnus.com>
Sender: daemon@cis.ohio-state.edu
Organization: GNUs Not Usenet
References: <TOMW.92Mar4123856@latchup.ucsb.edu>
Distribution: gnu
Date: Wed, 4 Mar 1992 13:55:17 GMT
Lines: 21

Reply to Parent Message