2012/8/4 Petter Reinholdtsen <[email protected]>: > [Martin-Éric Racine] >> Btw, this is an incorrect assumption. The check applies to >> iniscripts, because we're using commands that reside in the /usr >> hierarchy, which cannot be assumed to be available until $remote_fs >> has been mounted. > > I assume that some of the scripts in initscripts are part of the > definition of $remote_fs and only use commands in /usr when it is safe > to do so, and thus got this correct. Some of these scripts are 'half > of $remote_fs', and can not depend on themselves. :) > > And this I believe is a correct assumtion. If I am wrong, please > provide a concrete example. :)
Let's do this the other way around: provide concrete examples that using commands from the /usr hierarchy without first ensuring that $remote_fs is mounted is perfectly safe. Also demonstrate how some of those scripts can be a part of $remote_fs when networking (and thus mounting of remote filesystems) only happens later during rc2, and yet we are talking about scripts shipped as a part of sysvinit, which exist within rcS. _______________________________________________ Pkg-sysvinit-devel mailing list [email protected] http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-sysvinit-devel
