Re: nodes have authors

[marco]

On Tuesday 12 February 2002 19:04, you wrote:
> * Thomas Bushnell, BSG writes:
> > ams@kemisten.nu (Alfred M. Szmidt) writes:
> >> - Print the "unknown permission bits" (what is the official name
> >> for this?).
> >
> > Well, I would say it's the "nobody" bits or maybe the "anonymous"
> > user bits or the "no ID" bits.  We should figure out one standard
> > name to use.
>
> Right, I used "unknown" because that was what was in the header file,
> bits/stat.h.  "nobody" would be confusing as there is on most systems
> a user by that name (or do we even care what weird things they do on
> those UNIX systems? :). Maybe we could use "unknown user"?
>
> [snip]
>
> >> And I am working on setting the permission bits with chown, and
> >> changing the author bit with chown (chown owner:group:author?), and
> >> will probably implement the chauth (I think thats a better name
> >> then chauthor) program.
> >
> > Why is chauth better than chauthor?
>
> Same reason why we have chown instead of chowner, I guess.  But maybe
> users would get confused with the auth translator, in that case chauthor
> would be a better name.

I'm sure users will get confused...

>
> >> I was thinking on maybe removing all the "normal" UNIX file modes
> >> and only have ones for active/passive translators, and showing
> >> output similar to when you have a symlink (this will only work for
> >> passive translators):

You could also  add the PID of an active translator, that would be a cool
feature IMHO

> >
> > The normal modes really are important information; the filesystems
> > are setting them because it's a useful hint to users about how the
> > file behaves.
>
> Yes, but I thought that with the `symlink like' (the one suggested bellow)
> info this could be used instead of the normal modes.
>
> >> trw-rw-rw-rw- 1 root root root 0, 0 Dec 27 22:25 /dev/null =>
> >> /hurd/null
> >>
> >> Note, I used => instead of the normal -> to show that it is really
> >> different.
> >
> > Good idea.
>
> Should this also apply to symlinks? Or should the default behaviour be
> used.

I think the translator for symlinks shouldn't be shown or it should
be an option.

>
> Oh, can we define S_IRUNK , S_IWUNK and S_IXUNK in a libc header?
>
> #define S_IRUNK (S_IRUSR << S_IUNKSHIFT)
> #define S_IWUNK (S_IWUSR << S_IUNKSHIFT)
> #define S_IXUNK (S_IXUSR << S_IUNKSHIFT)

I think the right place for those bits is bits/stat.h