From: Michal Mertl <xmerm05 AT manes DOT vse DOT cz>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Problems with ScitechMGL for DJGPP
Date: Wed, 22 Oct 1997 17:27:03 +0200
Organization: Prague University of Economics
Lines: 31
Message-ID: <Pine.ULT.3.95.971022171057.7138D-100000@manes.vse.cz>
References: <Pine DOT GSO DOT 3 DOT 96 DOT 971020004915 DOT 5402A-100000 AT calvin DOT univalle DOT edu DOT co> <344D2118 DOT 37BF AT cam DOT org>
NNTP-Posting-Host: manes.vse.cz
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
cc: Guillermo Porras <guporras AT calvin DOT univalle DOT edu DOT co>
In-Reply-To: <344D2118.37BF@cam.org>
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Precedence: bulk

On Tue, 21 Oct 1997, Vic wrote:
> Guillermo Porras wrote:
> > 
> > I installed the base package and the MGL libraries from
> > SCitech for the DJGPP 2
> <snip>
> >.What
> > could it be .
> the DJGPP support is broken. It does NOT work. You should wait untill
> the DJGPP libs are again available then download them again.
> -- 
Did someone try to locate the problem ? Kendall thinks it's in pmode, and I
agree. I succefully compiled it (with problems - makefile doesn't work for
me, there is problem in makedefs.dj2(?) in section describing how to compile
asm file using tasm, when I correct it and check if it should work with
dmake -n (just show executed commands don't process them) the commands are
right, but without -n it still executes original version of command (problem
with forward slashes which tasm doesn't understand), really strange like dmake
uses different rule with and without -n switch). I do understand most of the
code, but not that with ctrlbreak, int10 and some other handlers where they
install some realmode code, why is this nessesary? In other handlers (like
keyboard) they just hook the PM interrupt from dpmi and according to DJGPP faq
DPMI server should always (if no RM handler installed by the program) pass the
control to PM handler. I also noticed strange thing - e.g. library and mouse
test program compiled with pgcc partially worked and with normal gcc it always
at least SIGSEGVed. I quite gave up already. 

Michal "MiMe" Mertl
  xmerm05 AT vse DOT cz