>> IIRC, the FreeDOS kernel is written largely in C, so the ASM source
>> isn't directly usable. It may be useful to go spelunking for the
>> algorithms used and how corner cases were handled.
> It certainly can help to deal with issues that arise out of undocumented
> features/bugs/issues, which in the past had to be re-implemented by
> re-engineering or plain guess work...
Frankly, not so much.
the relevant facts about MSDOS like internal structures, memory layout
aso. have been re-engineered/disassembled, documented and commented
by Andrew Schulman, Mike Podanowsky, and MANY others, and
merged in an almost complete (and almost correct) documented DOS API by Ralph
Brown.
thanks to them, and there is close to nothing to be learned by
studying old MSDOS sources.
there are not many 'algorithms' needed to implement xxDOS, and 'corner
cases' (once you know there is such a case) are easily identified, and
traced by either writing a small test program, or simply stepping
MSDOS execution. Easier at least then trying to understand by reading
MSDOS 2.0 sources.
>> (Speaking personally, I'd love to see *FreeDOS*
>> re-licensed under something other than the GPL.)
> Now THAT is something I would agree with you, even if just to get rid of
> Stallmanitis (thanks Tom! ;-) )
while I agree, this is not going to happen.
a) there is no such thing as FreeDOS with a single license; even in my
*really* minimum setup of
kernel, command, himem
you have 2 different licenses.
b) even for the kernel alone, you have a dozen+ developers; at least
one of them is dead. no way to have them agree to a different
license.
tom
_______________________________________________
Freedos-user mailing list
Freedos-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-user
