Hi Gustavo, > >>> Letting systemd daemonize connmand, as we are already doing, there's no > >>> need to have an option "-n" since it will work for both debugging and > >>> running it as a daemon. Therefore, just remove the option. > >>> > >>> Other init systems may use whatever method they have to turn connmand > >>> into a > >>> daemon. > >> > >> You are breaking an API by removing an option, now connman will not > >> start if user supplies -n option. > > > > humn? Command line parameters are part of an API? > > > > We provide script to start with systemd, and that is updated as well. > > > >> > >> Anyway, I do not see much point removing functionality that is useful > >> when systemd is not being used. So NACK from me. > > > > The point here is simplification... there's no point in dealing with > > this today - whatever init is starting connmand can turn it into a > > daemon. And if doesn't have native support for this (I really hope all > > init implementations have), it could use helper programs. > > Marcel keeps saying he wants to just support systemd systems, then > this would make sense.
yes, that is the plan. > But really, if you do so you can remove the detaching code, not just > the option... and more, seems systemd is getting (or got?) backtracing > support so it could do the coredump storage and reporting outside of > connman. We can not screw over the developers. So something like debug output going to the command line with -n when testing manually needs to be supported in one way or another. I need to look into this a bit for ELL as well. Regards Marcel _______________________________________________ connman mailing list [email protected] http://lists.connman.net/listinfo/connman
