Our unit file doesn't work, as it is setting stdin to NULL and stdout to syslog. A corrected one can be found at https://bugzilla.redhat.com/show_bug.cgi?id=1091566#c8 .
Source RPM: saned => sane-1.0.28-1.mga8.src.rpmAssignee: bugsquad => lists.jjorge
Our unit file works fine on my system. And it matches the example given in the saned man page.
CC: (none) => mageia
Actually, the man page gives two versions of the saned@.service, one for the case where saned has been compiled with systemd support, which matches the one we ship, and one for the case where saned has been compiled without systemd support, which matches the sample given by the link I provided (almost - the salient point being that it sets StandardInput to "socket"). So the question is whether we compile saned with systemd support.
(In reply to Frank Griffin from comment #2) > So the question is whether we compile saned with systemd support. We do: http://svnweb.mageia.org/packages/cauldron/sane/current/SPECS/sane.spec?view=markup#l344
OK, then it comes down to whether we believe the reporter in the link I provided, and/or whether modifying our provided service file(s) to his specifications actually fixes the problems we're seeing here.
(In reply to Frank Griffin from comment #4) > OK, then it comes down to whether we believe the reporter in the link I > provided See the later comment in that report: https://bugzilla.redhat.com/show_bug.cgi?id=1091566#c18 > and/or whether modifying our provided service file(s) to his > specifications actually fixes the problems we're seeing here. As I reported on qa-discuss, it works for me as it is.
OK, since I haven't experienced the problem myself, I'll close this as WORKSFORME. But it still doesn't explain why both the OP in the link and the OP in qa-discuss have the problem, or possibly whether it's in MGA7 and not cauldron.
Status: NEW => RESOLVEDResolution: (none) => WORKSFORME
Sorry, got the link wrong. The comment I meant to refer to is https://bugzilla.redhat.com/show_bug.cgi?id=1091566#c16 which explains why the OP in your link had the problem and why we don't. Having followed my suggestions, the OP in qa-discuss (Herman) no longer has a problem. drakscanner does need to be updated though, as it tries to enable saned.service, whereas it should now enable saned.socket.