From: Andrew Crabtree <andrewc AT typhoon DOT rose DOT hp DOT com>
Message-Id: <199710221613.AA278586807@typhoon.rose.hp.com>
Subject: Re: compiling C++ with -a switch
To: eliz AT is DOT elta DOT co DOT il (Eli Zaretskii)
Date: Wed, 22 Oct 1997 9:13:26 PDT
Cc: wallace AT ami1 DOT md DOT essd DOT northgrum DOT com, djgpp AT delorie DOT com,
        andrewc AT rosemail DOT rose DOT hp DOT com
In-Reply-To: <Pine.SUN.3.91.971022104822.7037J-100000@is>; from "Eli Zaretskii" at Oct 22, 97 10:53 am
Reply-To: andrewc AT rosemail DOT rose DOT hp DOT com
Precedence: bulk

 
> DJ Delorie pointed out that this patch was dangerous, since HAVE_ATEXIT is
> used by other parts of libgcc.  Sorry I didn't catch this myself.
 

I don't know the details of the first patch, but I believe I added the 
following lines to config/go32.h.  Your mention of HAVE_ATEXIT just
makes we worried I may have overlooked something when I did this last 
April-ish.

/* djgpp has atexit ().  */
#undef HAVE_ATEXIT
#define HAVE_ATEXIT

/* djgpp automatically calls it's own version of __main, so don't define one
   in libgcc, nor call one in main().  */
#define HAS_INIT_SECTION



> (Andrew, can you please see if this patch still applies to the latest 
> snapshots?  I will then send it to the GCC maintainers.  Thanks.)

typhoon:gcc$ patch < elipatch.diff
patching file `libgcc2.c'
Hunk #1 succeeded at 1688 (offset 77 lines).
typhoon:gcc$ emacs libgcc2.c
typhoon:gcc$

Looks OK (this was against the pgcc 1008 sources).  
Remember to send to the egcs list too, as well as gcc2
(or I could do this if you want).

Andy