Mail Archives: djgpp/1993/06/14/06:35:45
> I can probably get by without interrupt-driven transmission, but it
> would mean some nasty kludging in my code to make sure it behaves
> itself when there is a lot of output to the line going on. Does
> anybody have any suggestions as to how best to implement interrupt
> driven transmission in "async"? Has anybody tried doing this? Does
> anyone care to volunteer hints/tips/warnings on this?
>
> Thanks a lot. Your a great bunch...
>
> ---------------------------------------------------------------------
> Dani Flexer, MENNEN MEDICAL LTD.
I just picked this async stuff up recently, and haven't figured
it out yet (I think there is something else I haven't tracked
down yet: the original DJ version and documentation).
Last year I dealt with my async problems by writing a series
of assembler routines which checked whether a port was ready
to read or write before attempting it, waited or returned without
wait, checked (and buffered and updated) status, etc.
I did still have a problem which seemed only would go away
by handling interrupts, so I rewrote it to work with a FOSSIL
driver (I think it was X00 I got it working with).
I meant to knock it into shape, but have moved on to other
things (and to another country where it is too expensive to
use a modem from home). I've reformatted my harddisk since
then, but I mean to dig through my backups for this sometime.
Working with a FOSSIL driver made everything pretty simple,
and the second library was I think all written directly in
GCC (C or C++, I forget which). (With some FOSSIL drivers
I had problems relating to status signals which my modems
and cables didn't use, but X00 was the one that worked for
me I think.)
dP
--
E-mail: powers AT acm DOT org || powers AT inf DOT enst DOT fr
P-mail: Dr David Powers, Visiting Professor, TEL: +33-1-45.81.80.86
Telecom Paris (ENST), Departement Informatique FAX: +33-1-45.81.31.19
46, rue Barrault - 75634 Paris cedex 13 SEC: +33-1-45.81.78.70
- Raw text -