Date: Thu, 9 Jan 1997 15:25:47 +0100 (MET)
From: Robert Hoehne <robert DOT hoehne AT mathematik DOT tu-chemnitz DOT de>
To: Bill Currie <billc AT blackmagic DOT tait DOT co DOT nz>
Cc: "Salvador Eduardo Tropea (SET)" <salvador AT natacha DOT inti DOT edu DOT ar>,
        DJGPP workers <djgpp-workers AT delorie DOT com>
Subject: Re: Debuggers and mouse hooking (or maybe dpmi memory allocatio
In-Reply-To: <32D375F1.39F7@blackmagic.tait.co.nz>
Message-Id: <Pine.HPP.3.95.970109151805.7026F-100000@newton.mathematik.tu-chemnitz.de>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Wed, 8 Jan 1997, Bill Currie wrote:

> The dpmi spec states that int 31 preserves ALL unspecified registers AND
> flags.
> This means that the caller (the application) shlould only see the carry
> flag 
> change; ALL of the others must be as they were BEFORE the int 31 was
> invoced.
OK. Now I have looked also in the spec and you'r right. But your
originally problem was, that you thought dbgcom.c modifies in some
undefined way the flags. But this is not true, because it returns
the flags, which were returned by the original int 0x31 and if
these are bad, than you would get it also bad when running not
under the debugger.

But if you change the patch (at first clearing the CF bit and then
oring it) I think it's OK.

> Not only that, but for some of the functions, the error code was getting
> lost 
> (or the successfull result in ax, maybe both, can't remember)
OK, I think the best would be to resend a patch for dbgcom.c (not against
your last patch).

Robert