Hi Hans, On Wed, 13 Feb 2008 13:07:33 +0100, Hans de Goede wrote: > Jean Delvare wrote: > > On Mon, 11 Feb 2008 13:04:10 +0100, Hans de Goede wrote: > >> I've received the attached patch through Fedora's bugzilla. It fixes the return > >> values of the various error exit cases to match the LSB spec: > >> http://refspecs.linux-foundation.org/LSB_3.2.0/LSB-Core-generic/LSB-Core-generic/iniscrptact.html > >> > >> It looks sane to me, if there are no objections I'll commit it to svn. > > > >> --- lm_sensors.orig 2008-02-07 11:37:22.000000000 -0500 > >> +++ lm_sensors 2008-02-07 11:41:04.000000000 -0500 > >> @@ -40,15 +40,15 @@ > >> > >> # Don't bother if /proc/sensors still doesn't exist, kernel doesn't have > >> # support for sensors. > >> - [ -e /proc/sys/dev/sensors ] || exit 0 > >> + [ -e /proc/sys/dev/sensors ] || exit 6 > > > > Note that this doesn't work with a 2.6 kernel anyway, which makes me > > wonder whether it's worth keeping these init files in the lm-sensors > > tree. Obviously no distribution uses them as-is. > > > > It helps to read the whole script before jumping to conclusions like this, > above this is: > if grep -q sysfs /proc/mounts; then > WITHSYS=1 > else > WITHSYS=0 > fi > > if [ $WITHSYS == "0" ]; then > > So this piece of code doesn't get executed on 2.6 kernels with sysfs enabled. Oops, you're right, I totally missed that, probably because lm_sensors.init.suse doesn't have this conditional. > And actually atleast one distro (Fedora) is using the scrip as is. The script we ship in openSuse is based on lm_sensors.init.suse, but it's heavily patched. In fact the patch is bigger than the script itself. My impression is that lm_sensors.init.suse has not been updated in parallel with lm_sensors.init and is totally lagging behind by now. So I'd rather delete it, and let openSuse base their patch on lm_sensors.init (which is certainly a saner base) or even just maintain their own init script. -- Jean Delvare
