> PE>myself).  If I manage to obtain DFLAT, I'll make it
 > PE>available on my board.  David Nugent is meant to have
 > PE>ported it to OS/2, but I don't get ACH yet, and I can't
 > PE>FREQ from Michael Butler, so I'm waiting for David Begley
 > PE>to finish his exams to get that.  BFN.

 > Do you mean DFLAT that was described in DDJ? I have DDJ files here up to
 > about July this year, many seem to contain DFLAT, presumably
 > progressively refined versions.

Bear in mind that there are two strains of dflat - dflat (in C, up to v1.5)
and dflat++ (in C++ as at v2.0 in the November 1993 DDJ).

My port was dflat++ - the C++ version - and it's pretty much in the 'hack'
class since it was only used as a basis for quickly testing SL/2, the
library I'd written for MSDOS which emulates the main OS/2 subsystems
providing a better basis for implementing event driven user interface
software than the standard MS-Gross software interrupt kludges.

The port I did of dflat++ (it was v1.0, not 2.0 although I was surprised
that there isn't all that much different) just "works", and
that's all - I make no further guarantees since although I fixed a fair few
bugs in it,  there are certain aspects of the design I am not at all happy
with. Some sections of the code - in particular dispatching menu selections
- is extremely poor programming in that it is unsafe and very un-OOP for
something that is supposed to be written in pure C++. Type casting is used
far too freely, and you can easily get yourself into lots of trouble with
it (which, it seems to me, is something that C++ was specifically designed
to avoid).

I believe that Jon Wright (contactable at 3:633/379, if you're interested
in following this up) has done a more complete port of version 2.0 to OS/2
and has come up with something more solid - he's apparently writing an
article on it. I know he's rewritten most of the device interface
completely, whereas I just hacked it to get it working. However, you could
probably use SL/2 with it anyway and therefore use his port for MSDOS as
well, since the OS/2 subsystem API is a constant and SL/2 follows it fairly
religiously with only a few exceptions. I'll give this a try myself when I
come across his port.

SL/2 has also been further debugged since the last release into AUSTC, with
some minor additions and corrections. If anyone is interested, I'll hatch a
'beta 3'.  I'll see if I can find someone with a Borland compiler locally
to get it working with BC++ and TASM for MSDOS as well.


Cheers,

  david

---