On Tue, 08.02.11 13:47, Alexey Shabalin (a.shaba...@gmail.com) wrote: > > On Tue, Feb 8, 2011 at 12:21 PM, Lennart Poettering wrote: > > On Mon, 07.02.11 14:24, Alexey Shabalin wrote: > >> How chenge (Condition1 || Condition2) to (Condition1 && Condition2) > > I'd very much prefer not having to extend the condition language to more > > than what it currently covers. So, before we think about this more, I'd > > like to hear a very good use case for this. Your example doesn't really > > convince me, and I'd suggest a different fix (see above) instead. > > Ok. Let's look at fedora-storage-init.service and > ExecStart=/lib/systemd/fedora-storage-init from initscripts in Fedora, > for example. > We see in fedora-storage-init
The fedora specific scripts are not particularly great examples how we really should be doing things. They are mostly there as a result of the rc.sysinit split up and because we didn't deem them good enough to be included in the stuff we ship with systemd upstream. Sooner or later the mpath stuff is probably going to be pulled in via udev, too. In fact we already have discussed this a couple of times with its maintainer. So here too my reply would be: a different fix is better ;-)... > if ! strstr "$cmdline" nompath && [ -f /etc/multipath.conf ] && \ > [ -x /sbin/multipath ]; then > modprobe dm-multipath > /dev/null 2>&1 Grmpf. I don't really like this modprobe line, this should be made auto-loadable, so that the kernel APi can just be used and if it isn't there it will be loaded. Explicit invocations of modprobe are usually an indication that something needs to be fixed. Lennart -- Lennart Poettering - Red Hat, Inc. _______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel