On Wed, 8 May 2019 at 18:16, Bernd Zeimetz <be...@bzed.de> wrote: > Am 8. Mai 2019 22:49:45 MESZ schrieb "Lisandro Damián Nicanor Pérez Meyer" > <perezme...@gmail.com>: > >El mié., 8 may. 2019 17:42, Bernd Zeimetz <be...@bzed.de> escribió: > > > >> systemctl edit gpsd.service > >> > >> Just use common knowledge. > >> > > > >I don't think that's common knowledge, but I would consider this half > >ok if > >a proper Readme.Debian would be provided. > > Well, I think people should know their init system. > > > > >But again, the way you "fixed" the behavior is not right. > > Why? gpsd should never ever be started automatically.
I really don't want to sound aggressive nor anoying here here, but just want to make use of a real-case example. Embedded system running Debian in the middle of nowhere without internet connection and no real time clock. GPS has a backup battery that helps maintain it's clock when not powered, so it will have a valid time as soon as it's being turned on. System starts and gpsd is used as fast as possible to get data into ntp (needs -n). No gpsd clients are present, so nobody will call it's socket. > Actually I think the solution you are looking for is to configure udev to > handle your gps device properly. Unfortunately most gps devices use standard > usb to serial converters, so we can't handle them all or it will break other > things. For those use cases there is gpsd's -n option, like in my example above. > >Would you accept a patch if I prepare a proper fix? There is nothing > >wrong > >in asking me for that. > > Feel free to send patches, but I won't say that I accept them without seeing > them first. Sure thing, you are the maintainer :-) I'll be back with it as soon as time allows me. -- Lisandro Damián Nicanor Pérez Meyer http://perezmeyer.com.ar/ http://perezmeyer.blogspot.com/
