Mail Archives: djgpp/1994/11/02/23:13:16
> I wrote:
> > > DEVICE=C:\DOS\EMM386.EXE NOEMS X=C800-C9FF
> > >
> > > Commented this out and demacs came up fine! Now... what have I lost
> > > by doing this?
>
> Pieter Kunst and others wrote basically:
> > I would replace 'NOEMS' by 'RAM FRAME=NONE'.
>
> I did this. Demacs 1.2.0 now comes up fine for me, _and_ I
> have high memory available. Thanks!
I had this problem also. I would switch between the two cases. I used (from
memory) these two setups:
DEVICE=C:\DOS\EMM386.EXE 8192 NOEMS M9
because this gave me UMBs to LOADHIGH, and
DEVICE=C:\DOS\EMM386.EXE 8192 RAM M9
to get the VCPI for emacs, but not enough space to load everything high, so
that some software requiring more than 500K of normal memory would not load.
I tried without EMM386 and emacs came up fine, but I had no UBMs of course.
I tried
DEVICE=C:\DOS\EMM386.EXE RAM FRAME=NONE
and emacs came up fine, but very slowly and lots of disc access (paging out I
suppose). I had UMBs so that some stuff was loaded high. I specified 8M for
VCPI as before
DEVICE=C:\DOS\EMM386.EXE 8192 RAM FRAME=NONE
and emacs came up quickly and without the unexpected disc activity.
However, I still did not have all the UMBs that NOEMS gave me in the first
place. But, I tried
DEVICE=C:\DOS\EMM386.EXE 8192 RAM FRAME=NONE M9
which told EMM386 to make use of the memory E000 and up. This recovers most of
the regular memory and eliminates the need to switch between the two EMM386
commands depending on the application.
Any comments?
(I suppose this could be directed somewhere else, except that using djgpp and
emacs got me into these conflicts with MSDOS. Maybe someone else here is in
the same situation).
Peter Horan School of Computing and Mathematics
peter AT deakin DOT edu DOT au Deakin University
Geelong
+61-52-27 1234 (Voice) Victoria 3217
+61-52-27 2028 (FAX) Australia
- Raw text -