From: Andris Pavenis <pavenis AT latnet DOT lv>
To: djgpp AT delorie DOT com, Rob Kramer <robk AT starhub DOT net DOT sg>
Subject: Re: Building a gcc 3.2.2 cross compiler..
Date: Mon, 14 Apr 2003 08:27:31 +0300
User-Agent: KMail/1.5.1
References: <3e979062 AT news DOT starhub DOT net DOT sg> <3e997711 AT news DOT starhub DOT net DOT sg>
In-Reply-To: <3e997711@news.starhub.net.sg>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200304140827.32028.pavenis@latnet.lv>
Reply-To: djgpp AT delorie DOT com

On Sunday 13 April 2003 17:41, Rob Kramer wrote:
> Hi,
>
> On second thought, I'm trying to build a c/c++ cross compiler, not a native
> DOS executable. I think this means most of the patches in gcc322s2.zip are
> not required for me, since almost all seem to be related to fixing DOS
> things (c++ -> cxx, fix paths, etc). 
>
> One I'm not sure about:
>
> *** sdbout.c~1  Mon Sep 16 17:52:54 2002
> --- sdbout.c    Mon Sep 16 18:17:42 2002
> ***************
> *** 853,859 ****
>         name = IDENTIFIER_POINTER (DECL_ASSEMBLER_NAME (decl));
>
>         if (GET_CODE (value) == MEM
> !         && GET_CODE (XEXP (value, 0)) == SYMBOL_REF)
>         {
>           PUT_SDB_DEF (name);
>           if (TREE_PUBLIC (decl))
> --- 853,861 ----
>         name = IDENTIFIER_POINTER (DECL_ASSEMBLER_NAME (decl));
>
>         if (GET_CODE (value) == MEM
> !         && GET_CODE (XEXP (value, 0)) == SYMBOL_REF
> !           && GET_CODE (XEXP (XEXP (value, 0), 0)) != PLUS
> !           && GET_CODE (XEXP (XEXP (value, 0), 0)) != MINUS)
>         {
>           PUT_SDB_DEF (name);
>           if (TREE_PUBLIC (decl))
>
> I have no idea what this does, but this seems to be the only patch not
> directly DOS compiler related.

There are still patches related to some bug reports I have submitted. As far 
as I remeber this was related assembler warning messages when COFF
debugging info is requested for Fortran program which uses COMMON blocks are 
being used.

A different patch were applied for upcomming gcc-3.3 that fix this problem. 
But it's was not applied for gcc-3.2.2. 

Also the default now is DWARF2 debugging info. COFF debugging info support is 
perhaps rather botroted.

>
> If I don't have to rebuild configures etc, the rest of the process should
> be reasonable standard (ahem)..

Andris