Date: Sun, 24 May 1998 19:09:49 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Mr A Appleyard <MCLSSAA2 AT fs2 DOT mt DOT umist DOT ac DOT uk>
cc: DJGPP AT delorie DOT com
Subject: Re: BUG in Gnu C++ (djgpp v2) mouse access under Windows 95
In-Reply-To: <16844584E2C@fs2.mt.umist.ac.uk>
Message-ID: <Pine.SUN.3.91.980524190926.12971B-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On Fri, 22 May 1998, Mr A Appleyard wrote:

>   AAEMACS reads the keyboard by "AX=0x0700; call interrupt 0x21;

My experience with DOS function 7 is extremely bad: it does indeed
lose extended keys.  I think it's because DOS uses old BIOS functions
(0 and 1 instead of 10h and 11h) to read the keyboard, which don't
support the extended keys.

I have no idea why did that problem start when you changed your mouse
handler.

> Please how does Emacs read the keyboard by BIOS?

It's a very complex code, but in general it uses Int 16 function 11 to
check whether a key is pending in the keyboard buffer, and function
10h to read a key if it is there.

You could also use the DJGPP library function `getxkey' which does
roughly the same (but it waits for the key, so you need to test for a
key being available before you call `getxkey').

> I heard that djgpp Gnu C++ hooks interrupt 9 anyway as part of trapping the
> ctrl-C breakin event. What has come of the code to do this, that I
> sent to one of the djgpp email groups a while ago?

You need to ask DJ Delorie <dj AT delorie DOT com>.  He's the one who decides
what will and what won't be added to the library.  (DJGPP v2.02 is
still in alpha phase.)