Re: Should we implement a /etc/profile.d?

[Raul Miller <rdm@test.legislate.com>]

On May 5, Lars Wirzenius wrote
> Er, no, I think that is an extremely good example of what
> should not go into /etc/profile.d. Packages should not rely on
> a non-standard magic environment variable to work properly.

I was thinking more in terms of administrative decisions than
package maintainer decisions.  Packages should work fine
without /etc/profile.d/.  Having a single point of control is
purely an administrative convenience.

Also, I forgot to mention resource management support (e.g.
ulimit).  Again, this is something that the local
administrator would like to be able to set without having
to worry about the difference between an ssh user who uses
tcsh, an xdm user who uses es, and a tty user who uses zsh.

I don't see how the current /etc/profile.d/ proposal even
begins to cover this kind of ground.

-- 
Raul


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .