Branch for review: 'riccardo/voice'

[Aleksander Morgado]

On Wed, Jun 3, 2015 at 4:04 PM, Riccardo Vangelisti
<riccardo.vangelisti at sadel.it> wrote:
>>> Anyway, the branch looks good enough to me; there are some things that
>>> it doesn't yet cover, but I believe it's a good start. Dan, what do
>>> you think?
>>> libmm-glib: added support for Modem.Voice and Call interfaces
>>
>> Should mm_call_properties_set_number() check that the number is not
>> zero-length?  I'm not sure how much validation we want to do in
>> mm-call-properties.c but maybe a bit more?
>
> Ok, we've pushed this. All other mm_call_properties_set_* functions are
> already checked by enum type.

That fix you made doesn't really validate the input returning an error
if not valid; it just ignores the string passed if it's not considered
valid.

Anyway, I think we shouldn't do any validation, not even for empty
string, in the setter method in MMCallProperties. This object is a
helper in libmm-glib, which means that it isn't mandatory to use this
helper object to create the call object, you could do it building the
DBus request yourself. The validation of parameters should always be
done in ModemManager when it receives the messages from the bus.

For now I'll ignore the specific commit
(4f6e3a3f47b226f497827837a58f407cd0e4b66e) doing this, won't
cherry-pick it to my branch.

-- 
Aleksander
https://aleksander.es