From: smart@koel.mel.dit.CSIRO.AU (Robert Smart) Subject: Re: [PATCH] CTRL-T status line patch for 0.99.pl11 Date: Sun, 1 Aug 1993 23:17:53 GMT
ORAKEL@rzmain.rz.uni-ulm.de (Framstag) writes:
>Next to do are logicals and ACLs (ok, the last one concerns more the FS).
>About logicals, Olaf Titz and I are preparing an article for discussing
>this topic.
I think ACLs are (or will be) in ext2fs. However introducing them in
different ways in different versions of unix is dubious. We need some
experiment in software clearly labelled "experimental - this will change".
Then we need to get it through posix so that we get a total system. It
is very confusing to introduce ACLs but not also change "ls -l".
Logical names are perfectly well handled in unix with 2 facilities:
environment variables and soft links. If you combine soft links with
IFS you can even get the effect where local definitions supercede
global ones.
We don't need another name space in Linux. Arguably there are too many.
>Last question: will this CTRL-T patch in later offical kernel releases!?
>Linus, are you reading? ;-)
I vote for including a CTRL-T option in Linux. But Linus has to maintain
it and it is important that the kernel doesn't get so big that he can't
handle it all. An alternative is to do it at user level. For example the
screen program could (optionally) intercept CTRL-T and dig around in the
operating system to find out what is going on. The nice thing about CTRL-T
(which is a TOPS-10 feature that DEC only reluctantly added to VMS) is
that the user gets some idea when the system is dead (as opposed to his
program waiting for something). This is handled by job control in unix:
you can CTRL-Z and bg the job and then watch it. CTRL-T would be easier
though.
Bob Smart