From: "Albert D. Cahalan" <acahalan@cs.uml.edu>
Date: Mon, 12 Jun 2000 01:49:16 -0400 (EDT)
> Well, you're right as it currently stands. The problem is those UID
> 'compatibility mode' options were meant to hold us over until the
> filesystems could support capabilities,
They are not just to "hold us over". The traditional system must
continue to work, because many people will not want to change.
That's fine. The traditional system will continue to work --- and does,
today.
What's there now is already broken. As I wrote above, there
might be some major changes in the future. By that I mean the
basic inheritance algorithm could go from draft 17 to draft 16.
Actually, D16 to D17 is hardly a major change, and it's not at *all*
clear we're going to make that move. I'm being polite and listening to
folks, but the more people sutdy the problem, the more it becomes clear
that the folks original Posix committee had dealt with many of these
issues already and had come up with a pretty good design.
The major problem here is that a number of people had some fundamentally
deep misunderstandings about how things like the Inheritable set was
supposed to work. They made claims that simply weren't supported by the
actual draft standards (either D15, D16, or D17). This spawned
confusion on various mailing lists, but it was mainly caused by people
who didn't grok the concept, not the concept itself.
Ever wonder why? The system is not compatible with UNIX.
Capabilities have been implemented in many different Unix systems,
actually: Trusted Solaris, Trusted AIX, Trusted Irix, etc. As much as
we can, we should try to be compatible with these other systems, unless
there are some fundamental reasons why that wouldn't work.
- Ted
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Jun 15 2000 - 21:00:25 EST