On Friday 08 April 2005 17:42, Julien BLACHE wrote: > Gerhard Jaeger <[email protected]> wrote: > > Hi, > > > sorry for the late response. > > These are the problems, that I've expected, when introducing this > > feature. But I have no idea how to solve it! > > Well, disabling the locking if your check fails, instead of aborting > the configure script would be a good start :)
Hmmm, not my favourite way, but probably a better one ;) > Correcting the error message would be good too, this one is misleading ;) Oh yes, of course, the current one is more than ugly. I'll fix both issues today. > > Have you any idea how to solve such an issue? > > I'd like to have this feature and therefore I also added the > > possibility to change this default uucp-group: > > --with-group=your_group > > > > Any other comments? > > The programs usually locking the serial port are run setuid root or > uucp or something, so they can write the lock file in /var/lock. > > This is not applicable to SANE, because SANE's only the middleware > (and thus, cannot do setuid()). > > Quite honestly, I have no (portable) idea. Maybe something based on > shared memory could work. If something comes to my mind, I'll share my > thoughts about that :) Okay, I will also spend some time in investigation on the shared memory trail... Thanks Gerhard
