Message-Id: <9703111850.AA26262@hum.amu.edu.pl>
Comments: Authenticated sender is <grendel AT ananke DOT amu DOT edu DOT pl>
From: "Mark Habersack" <grendel AT ananke DOT amu DOT edu DOT pl>
Organization: KGB
To: chambersb AT juno DOT com (Benjamin D Chambers)
Date: Tue, 11 Mar 1997 19:42:04 +0100
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7BIT
Subject: Re: [opendos] Multitasker API and a demo program 
Reply-To: grendel AT ananke DOT amu DOT edu DOT pl
Cc: "'OpenDOS Mailing List'" <opendos AT mail DOT tacoma DOT net>
In-Reply-To: <19970310.174935.10167.7.chambersb@juno.com>
Sender: owner-opendos AT mail DOT tacoma DOT net
Precedence: bulk

On 10 Mar 97 (at 20:46) Benjamin D Chambers became famous by saying:

> >either (though I could be wrong). Besides, doing it What version of
> TASM?  Also, did you remember the compiler directives?
> ie:
> .286		Enable 286 Instructions
> .386		Enable 386 Instructions
> .486		Enable 486 Instructions
> .586		Enable 586 Instructions
> TASM 5 supported Pentium, I don't know if anything knewer has come along.
So did TASM 4. And even with older ones it's not a problem. Just get any 
opcode reference and create a macro for that!

>  I don't like their other compilers, but Borland's TASM really is a good
> piece of work (IMHO).
Provided you don't use their high-level directives - they can screw things up 
an make your life miserable for two weeks ;-))

> >stand-alone with 
> >TASM 
> >means that I'd have to find out about the C calling conventions. 
They're easy:

  1. The caller sets up the parameters on stack by 'pushing' them in the 
     reverse order: last one goes first
  2. The function is invoked
  3. Caller cleans up the stack from parameters by adding the pushed 
     arguments' size to (e)sp register (assuming 8 bytes were pushed):
         addl  $8, %esp    - GNU AS syntax
         add   esp, 8      - Intel syntax

That's it!

> I don't think so - I vaguely remember something about using the macro
> CALL and it would take care of it for you.  Also, when you set up a
True. There is such a high-level version of the instruction - never used it, 
though.

> procedure, you can call it C or PASCAL (I know for sure Pascal, so I
You can define it to be PROLOG, C, CPP, PASCAL, BASIC (AFAIK):

   foobar PROC FAR C

or just define the default calling convention for entire module by putting 
this at the beginning:

   MODEL LARGE, C
or
   MODEL TPASCAL     - for interfacing with Turbo Pascal only!

> assume you can do the same for C).  I don't remember the actual commands,
> check it's manual (I lost mine years ago :( ).
> 
> >Why Borland? It's small (v3.1), easy to use (*great* IDE) and is 
> >perfectly 
> >suited for the task.
> 
> I was really more interested in whether or not the API could be used by
> djgpp programs - but now that I think about it, isn't there a bug in the
> DPMI provider?  (So the multitasking wouldn't work with djgpp programs
> anyways - would it?)  I haven't actually gotten OpenDOS yet (I plan to wait
You don't have to use the DPMI server that comes with OD - just put DPMI=OFF 
on its command line in (D)CONFIG.SYS and use CWSDPMI. And there is no reason 
why the API wouldn't work with DJGPP. It'd be just a lot of RM calls ;-) And 
it hurts...

=============== OpenDOS - feel the power! http://www.caldera.com ==
From the delequeue to the regiment, a profession in a flash. But
remember Monday signings when from door to door you dash... On the
news the nation mourns you - Unknown Soldier count the cost: for a
second you'll be famous but labeled posthumous... Forgotten Sons...
---
Visit http://ananke.amu.edu.pl/~grendel