Hi again,
On Sat, Oct 12, 2019 at 8:15 AM TK Chia <[email protected]> wrote:
>
> > Okay, I think I've done enough fiddling to be worth sharing.
> >
> > Further testing and contributions are welcome, of course.
>
> Thanks! This looks like ... a rather major revamp of the makefiles. :-)
With the new build(s), I tried rebuilding the FD kernel ("tests /t
ke2041" under MetaDOS, "set NASMENV=-O9v" first). It seems the TCPP101
version (NAS16TCC) ran out of memory on KERNEL.ASM! (Dunno why,
haven't checked closely yet.) Is that a tradeoff for using Huge model?
Or some other problem? Naive question: should it use halloc() or
overlays?? Ugh. (I need to study the manuals closer, too.)
But the OpenWatcom (Large model) build (NAS16WAT) seemed to work fine
(KERNEL.SYS' CRC32 matches). FreeDOS wisely prefers OW anyways, but
still ....
Despite no actual code changes, I have improved my MAKEFILE.BOR again.
I haven't uploaded it yet. I may do some more fiddling (and also
derive a TC201-compatible makefile). It's also simple enough that I
could just write a .BAT to build it. (I also did some local fixes to
avoid dependency on sed. That was the whole reason for my included
SNPRINTF.DIF, to not burden anyone with that requirement.) It's also
reproducible now (avoiding __DATE__), hence easy to verify.
Just FYI.
_______________________________________________
Freedos-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freedos-user
