org.freesmartphone.GSM.Network Signal Strength Signal

[Thomas Seiler]

On Wed, Jun 25, 2008 at 5:53 PM, Michael 'Mickey' Lauer
<mickey at vanille-media.de> wrote:
> Just added. See
> http://git.freesmartphone.org/?p=specs.git;a=commitdiff;h=bc40b7d8223a9dbb438b4c7e56bd53282722c64f

thanks !

>> The provider name is not something that changes a lot, [...]
>> The signal strength on the other hand is a volatile information, [...]
>
> I agree, however note that we don't send a NetworkStatusChange when the signal
> strength changes, we only send it once a +CREG comes in.

I was under the impression that the modem was actually sending
unsolicited events
for signal changes, when i wrote this. However, as this is not the case,
i stand corrected, sorry for the noise.

>> If introspection is a possible way to learn about listening processes,
>> a similar signal could then be used for CellIDChanged. ophoned would
>> even be in a position to switch on and off the generation of netreg
>> unsolicited events from the modem.
>
> Yeah, that would be nice. Could you make a proposal for that on the dbus list?

The consense on #dbus seems to be that this is best handled by adding
a subscribe method to the API, and then track the unique-names of all subscribed
processes via NameOwnerChanged, then implement a ref-counting mechanism to
understand if somebody is listening or not, and then only emit the
signal, when there
are listeners.

This "tracking until client exits" feature could be implemented in a
general way inside the dbus-bindings,
as it would be useful in many other cases.

I will try to come up with some patches for dbus-python, hopefully soonisch...

Cheers,
Thomas


-- 
Excercise 17:
If the human brain was simple enough for us to understand we'd be so
simple we couldn't understand.
Prove this by induction.