[Fingerforce-devel] Bug#683863: libfprint0: udev rules not applied when libfprint0 is installed

[Andrew Pimlott]

Excerpts from Didier Raboud's message of Sun Aug 05 02:33:45 -0700 2012:
> Le samedi, 4 août 2012 23.56:02, Andrew Pimlott a écrit :
> > libfprint0 installs some udev rules to make fingerprint readers
> > accessible to the plugdev group.  However, since many fingerprint
> > readers are built-in to computers, they are never plugged in, and thus
> > the udev rules never fire. 
> 
> In my experience, that's partially wrong: the udev rules get run (at least) at 
> boot time. Are you really experiencing this problem for a device or is it a 
> theoretical problem?

Right, I should have said: the udev rules never fire until a reboot.  So
the fingerprint reader doesn't work until a reboot.

> > libfprint0 should trigger the udev rules when it installs them. 
> 
> I don't think that libfprint should be special-cased here. On my system, there 
> are 32 different packages installing udev rules under /lib/udev/rules.d and 
> libfprint is certainly not the only one that would benefit from "udevadm 
> trigger" runs.

That sound reasonable.  I would definitely be happy with a solution
that covers other packages.  The only thing that might be different is
that most other devices you can simply remove and plug in again.  I
can't do that with my build-in fingerprint reader.

> In my understanding of the situation of the udev rules, there is a requirement 
> to reboot to have things working correctly; and that's nothing libfprint 
> should fix for its own benefit.

That doesn't sound reasonable to me!  I'm not used to rebooting after I
install packages, even if they are desktop-oriented.

Thanks,
Andrew