[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Hurd features (continued)

From: olafBuddenhagen
Subject: Re: Hurd features (continued)
Date: Sat, 10 Feb 2007 20:44:49 +0100
User-agent: Mutt/1.5.13 (2006-08-11)


On Sun, Feb 11, 2007 at 12:37:12AM +1300, Shams wrote:

> Does/will Hurd support metadata capabilities or once again can/do
> translators play a role in this? If translators were the case then how
> could it achieve this (via extended attributes??).

No idea what you mean.

> Could you please eloborate on ACL vs UID part a bit more.
> Here is the scenario 1. UserA creates a file and wants to grant access
> to lets say 5 groups via ACL (each group containing a certain number
> of users) to a certain file.
> Then how does UID solve this problem?

UIDs don't solve this problem.

The standard UNIX solution is very simple: Create a group with all users
allowed to access the file.

The presented scenario is extremely unlikely in practice, and in fact it
demonstrates a common misunderstanding of the UNIX group mechanism:
Groups do not represent any kind of "teams" or whatever; rather, groups
are tied to specific uses. If you want more abstract "teams", you need
an admin tool that manages them in an own database, and maps them to
UNIX groups as needed.

Of course, in UNIX, only the admin can create groups, which can be a
serious limitation in some situations. (Which is only very poorly solved
by POSIX ACLs IMHO.) The hurdish solution is to run an your own access
control mechanism, and do whatever you want. You could run your own
filesystem that uses different rules to decide who is allowed to access
the files. Or, probably more useful in most cases, you still store your
files on the system-provided FS, but you run a proxy FS server that
implements your own access scheme on top of it. This can be POSIX ACLs
if you really want them, but it can also be anything else you can think


reply via email to

[Prev in Thread] Current Thread [Next in Thread]