Right now the lsb package requires /var/spool/rwho because the lsb test
looks for it.  However, frankly I think this is wrong.  

If you look at the LSB specification it is clear that the
/var/spool/rwho is getting pulled in because of the inclusion of FHS
2.2.  It also says "The FHS allows many components or subsystems to be
optional. An application must check for the existence of an optional
component before using it, and should behave in a reasonable manner if
the optional component is not present." [1]

So I took a look at the FHS 2.2 to see what it said about
/var/spool/rwho.  Which says about a variety of directories under
/var/spool: "The following directories, or symbolic links to
directories, must be in /var/spool, if the corresponding subsystem is
installed" [2]  And specifically mentions that rwho is an optional
component.

As a result the requirement for /var/spool/rwho in the lsb package IMHO
actually violates the LSB specification, which actually allows for those
optional components not to be installed.  If the LSB test is actually
requiring that directory to exist that the test is wrong.

Additionally, the only applications that should be messing with the files
in /var/spool/rwho is rwho and rwhod.  The specification does not
provide a format for those files, so any other application using them
would be straying off the LSB path anyway and into possible conflicting
formats (if someone were to ever reimplement rwho in a slightly
different way).

[1]
http://www.linuxbase.org/spec/refspecs/LSB_1.3.0/gLSB/gLSB/execenvfhs.html

[2]
http://www.pathname.com/fhs/2.2/fhs-5.14.html

-- 
Ben Reser <[EMAIL PROTECTED]>
http://ben.reser.org

"What upsets me is not that you lied to me, but that from now on I can
no longer believe you." -- Nietzsche

Reply via email to