Preferred way to get signal information

[Aleksander Morgado]

Hey!

> If your using c/c++ consider looking into the library ModemManager
> exposes, this is how mmcli works
> (https://cgit.freedesktop.org/ModemManager/ModemManager/tree/libmm-glib).
> It handles the dbus calls for you and just exposes some nice c code
> wrappers. In your case you should look at how 'mm-modem-signal.c'
> works and consider linking the library and including the libmm-glib.h
> header.
>
> If you're using another language you could also use dbus to access
> information from ModemManager directly, you can read up more on dbus
> here (https://www.freedesktop.org/wiki/Software/dbus/). For all the
> API's exposed by ModemManager via dbus you can checkout the API doc
> page (https://www.freedesktop.org/software/ModemManager/api/latest/).
> Python has a very nice dbus support library and there are lot's of
> examples online. Hope that helps.
>

Definitely this, yes. Running mmcli commands and parsing its output is
not a good idea right now, as the output of mmcli may change at any
point in time so your parsing may break in a future MM update. The
DBus API is stable, and any program written using that API directly
will always work as long as new MM releases are API compatible. For a
simple program that parses the signal output, doing this in python
should be extremely easy.

That said, my OpenWRT integration (that I assume Nick is using, given
he talks about uqmi as well), is based on.... parsing mmcli command
outputs :D yeah, that's very nasty, but in my mind this was because I
did want to have the mmcli output not in human-readable way (hard to
parse) but in some other optional way that would be easier to use via
shell scripts, e.g. returning a list of key=value pairs one per line
directly (and maintaining that API compatible in newer releases). That
idea is still open and I would love to have that implemented, so that
mmcli can then be used safely and easily from within shell scripts.

-- 
Aleksander
https://aleksander.es