On Wed, Nov 8, 2017 at 1:28 AM, Ralf Quint <freedos...@gmail.com> wrote:
> On 11/7/2017 10:24 PM, Random Liegh via Freedos-user wrote:
>> I'd like to ask if anyone has used the Desmet C compiler
> And yes, I would like to see this included in FreeDOS, as it is the only
> C compiler that both produces 16 bit code and can even run on a
> minimalistic FreeDOS system with only 256KB of RAM and two floppy discs.
360 kb floppies, I presume? I can fit a slimmed, old-but-good DJGPP
(32-bit DPMI, compressed with 7-Zip) on one 1.44 MB floppy. I thought
that was minimal enough! But I'm still sympathetic to 8086, of course
(even if I don't have any working hardware from that era).
Anything that works is welcome, but most people cross-compile these
days (or are willing to natively run a 386+ "cross"-compiler for
16-bit target, e.g. OpenWatcom).
> But when I mentioned this here on the list years back, the idea was
> pretty much shot down by all those proponents of Open Watcom and GCC,
> which neither one runs on FreeDOS itself without jumping through hoops...
Nothing has been shot down, but nobody has cleaned it up. It's not
even 100% obvious what files you need. My misunderstanding was that it
wasn't ANSI because it lacked some of the headers, but maybe those
were in one of the other (dozen) .ZIPs. It might work okay for your
needs, but it's tons inferior to OpenWatcom, no contest.
OpenWatcom, DJGPP, FPC, FBC, NASM, FASM, and many others don't require
"hoops", just 32-bit / 386+ DPMI or similar DOS extender (DOS4GW or
compatible). Most developers long ago gave up on 16-bit hosting and
focused on 16-bit "target". That way they can use more RAM and aren't
hampered by code size or small stack or longint kludges or similar.
There's nothing wrong with supporting 8086, and I still like it. But
it's far from burdensome to require 386 these days. Heck, we're
already a million years beyond that. Most developers don't even want
to support less than SSE2, and a lot want to kill 32-bit kernels
entirely. Don't forget AVX-512 and FMA, BMI, CLMUL, and whatever
gobbledegook they're throwing at us these days. We've long ago jumped
Is 8086 compatibility cool? Yes. Is it gravely important? I would say
no (but optional support never hurt anyone). We have bigger fish to
fry (e.g. being rebuildable with Free tools!) than worry about cpus
that most of us don't even own or use anymore.
Just to be clear, I'm not saying stop 8086 support (what little we
have left) or abandon 16-bit kernels. If someone wants to do that,
more power to them. But we can't afford to rule out every improvement
just because "it requires 386".
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
Freedos-user mailing list