Date: Thu, 10 Dec 1998 08:57:31 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: DJ Delorie <dj AT delorie DOT com>
cc: djgpp AT delorie DOT com, Leonid Pauzner <leonid AT pauzner DOT mccme DOT ru>,
        "K.B. Williams" <Kbwms AT aol DOT com>
Subject: Re: DJGPP 2.02 fails immediately on FPU-less machine !
In-Reply-To: <AB3PgRsqX4@pauzner.mccme.ru>
Message-ID: <Pine.SUN.3.91.981210084826.490A-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp AT delorie DOT com


On Wed, 9 Dec 1998, Leonid Pauzner wrote:

> Exiting due to signal SIGNOFP
> Coprocessor not available at eip=00003c67
> eax=00029cd0 ebx=00000000 ecx=00000000 edx=00029cd8 esi=00000000 edi=00000001
> ebp=00073918 esp=000738d4 program=C:\DJGPP\BIN/make.exe

DJ, this is due to reversed order of bytes in emu387.dxe.  I didn't check 
all of the emulator, but at least the header, including the magic 
signature "DXE1", is byte-reversed.  This, of course, causes _dxe_load to 
fail, and the net effect is that the emulator is not installed.

I didn't dig deeper (it was after midnight ;-), and I probably can't 
debug this on a PC anyway, but my first guess would be that the 
byte-reversing function in dxegen.c somehow doesn't work (maybe the 
wiseguy otherwise known as GCC 2.8.1 optimized it way too much?).

And here's a *really* long shot: could it be that the strange error 
message from ld.exe while linking emu387.dxe on a PC, reported by 
K.B. Williams a couple of days ago, is also connected to this somehow?