From: pavenis AT lanet DOT lv
Message-ID: <B0000096317@stargate.astr.lu.lv>
To: "Mark E." <snowball3 AT bigfoot DOT com>, djgpp-workers AT delorie DOT com
Date: Wed, 28 Jul 1999 21:12:09 +0300
MIME-Version: 1.0
Content-type: text/enriched; charset=US-ASCII
Content-transfer-encoding: 7BIT
Subject: Re: Another script for building binutils
In-reply-to: <199907281712.RAA31774@out5.ibm.net>
References: <B0000096305 AT stargate DOT astr DOT lu DOT lv>
X-mailer: Pegasus Mail for Win32 (v3.11)
Reply-To: djgpp-workers AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

<color><param>0100,0100,0100</param>On 28 Jul 99, at 13:13, Mark E. wrote:


<color><param>7F00,0000,0000</param>> > Here is another script for configuring and building binutils I'm using.

> > Last test were done with binutils snapshots (990727 appears to be 

> > bad: broken debug info, so I'll perhaps return to 990710 which I 

> > used for some time).

> 

> If you haven't already, please report this problem to binutils AT sourceware DOT cygnus DOT com 

> so the binutils folks can look into it. Since it's recent, they should be able to 

> track it down.

> 


</color>I haven't reported it yet.


One more thing I forgot to mention was that also C++ exceptions 
stopped to work when I tested 990727 snapshot with DJGPP.

Rebuilt 990710 using the same script I sent before and both things

appears to be Ok.


<color><param>7F00,0000,0000</param>> 

> > --- binutils-2.9.1/bfd/coff-i386.c.orig	Fri May  1 15:48:04 1998

> > +++ binutils-2.9.1/bfd/coff-i386.c	Thu Mar  4 20:33:02 1999

> > @@ -48,7 +48,9 @@

> >  #define COFF_DEFAULT_SECTION_ALIGNMENT_POWER (2)

> >  /* The page size is a guess based on ELF.  */

> > 

> > +#ifndef COFF_GO32_EXE

> >  #define COFF_PAGE_SIZE 0x1000

> > +#endif

> > 

> >  /* For some reason when using i386 COFF the value stored in the .text

> >     section for a reference to a common symbol is the value itself plus

> > 

> 

> And this too please. I don't know if it's right or not, but I do notice it prevents 

> .stabs sections from being marked as SEC_DEBUGGING.

> 


</color>This patch were suggested by Laszlo Molnar (his comments were 
also about my binary of BFD assembler which solved problems 
with padding in .eh_frame. Now these problems are solved also 
without using BFD assembler).  I haven't submitted it anywhere

----------------------------------------------------------------------------------------------

Date sent:      	<color><param>0000,0000,8000</param>Mon, 1 Feb 1999 12:20:50 +0100

</color>From:           	<color><param>0000,0000,8000</param>Laszlo Molnar <<laszlo DOT molnar AT eth DOT ericsson DOT se>

</color>To:             	<color><param>0000,0000,8000</param>Andris Pavenis <<pavenis AT lanet DOT lv>

<bold></color>Subject:        	<color><param>0000,0000,8000</param>Re: egcs patch

</bold></color>Send reply to:  	<color><param>0000,0000,8000</param>ml1050 AT freemail DOT c3 DOT hu


</color>On Thu, Jan 28, 1999 at 01:53:24PM +0200, Andris Pavenis wrote:

<color><param>7F00,0000,0000</param>> > The new gas works nicely. Where can I download the snapshot 
from?

> <underline><color><param>0000,8000,0000</param>ftp://ftp.cygnus.com/private/gas/


</underline></color>This weekend I played with it a little bit, and found the alignment

problem in strip (and other binutils). There is a COFF_PAGE_SIZE

define in coff_i386.c which sets this symbol to 0x1000. By adding 
an

#ifndef COFF_GO32_EXE...#endif around it, the aligment problem is

solved.


As I'm ready with the cross-binutils, I'll focus on hacking egcs ;-) I

plan to reduce the size of the eh_frame info for djgpp...


bye, LaSzlo

----------------------------------------------------------------------------------------------


I didn't notice problems with it yet so I kept it.


Andris


<nofill>