From:	Fokkema DOT DBRA AT delorie DOT com, 1043730 <dfokkema AT cs DOT vu DOT nl>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: far pointers again
Date:	Sat, 9 Oct 1999 21:40:07 GMT
Organization:	Fac. Wiskunde & Informatica, VU, Amsterdam
Lines:	21
Message-ID:	<7tocnn$gjj@cs.vu.nl>
References:	<E1FF8239A075D311AF7200A0C9D60AE30A671A AT probe-2 DOT acclaim-euro DOT net> <7tho7c$p4i AT cs DOT vu DOT nl> <37FD0358 DOT 68F04540 AT hmc DOT edu>
NNTP-Posting-Host:	galjas.cs.vu.nl
X-Newsreader:	TIN [version 1.2 PL2]
X-Poster-Key:	sha1:HyNaHtXifNW5k+VsSV2O6mVpYRQ=
Cancel-Lock:	sha1:jaF3y9qcMF+MVsPQRrXcTgSGIdE=
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

Hi All!

I like to thank you for commenting on my post and enlightening me on
several subjects. At least I understand now that far pointers aren't standard
C at all and break down a fundamental concept of C (thanks to Hans-Bernard),
while you don't even need them for a protected os (thanks to Eli, Nate and
Shawn). I decided on leaving my segmented memory model because it just
isn't common. And if I can get the same protection - I want to prevent that
any program could overwrite another - using paging that is fine with me. It
certainly keeps me with djgpp since I don't need far pointers anymore. That is
a relief, actually. And besides, my protection schemes aren't fancy in their
basic forms. Anyway, thank you all for your comments since they helped me
a lot.

Nate Eldredge wrote:
: And yes, there are very few other architectures that use a segmented
: memory model.  In fact, I can't think of any others.

Should say enough.

David

- Raw text -