| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
For easier testing of the ATD>1; command allow setting the voice mailbox
number using a VMBX_NUMBER environment variable.
|
| |
|
|
|
|
|
|
|
|
| |
MSC 4.26 and the text below it in the HFP 1.5 specification conflict with
each other with respect to the order of the callheld and callsetup
indicators. Errata 2459 clarifies the situation by explaining that the
text is correct and the MSC incorrect. Additionally the errata says that
the callheld indicator should be set to 1 just prior to indicting that the
the new call is active and the call setup procedure complete. This patch
updates the Maemo telephony driver logic accordingly.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
otherwise --disable-fortify has no effect and the default is
--enable-fortify anyway
|
| |
|
|
|
| |
This patch got submitted by an openSUSE contributor, I'm passing it
along so that I have one less patch to carry in my package ;)
|
| | |
|
| |
|
|
|
|
| |
The active_calls list can contain calls which are held but we're only
interested in setting the create timer if there's at least one call in
ACTIVE state (since then we know that it will be put on hold).
|
| |
|
|
|
|
| |
HFP 1.5 seems to require a 1->1 value change for the "callheld" indicator
every time a held and an active call are swapped. So the redundant change
check needs to be possible to configure on a per-indicator basis.
|
| |
|
|
|
| |
Now that telephony_update_indicator checks for no-change situations it's
unnecessary to do the same check within the telephony drivers.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The HFP specification requires the AG to enable the "callheld" indicator
with the value 1, i.e. "Call is placed on hold or active/held calls
swapped" when there is an active call and a second outgoing call is
attempted which causes the active call to go on hold. However, with the
current csd and its underlying modem the existing call is set to hold
before any new call creation attempt is indicated upwards in the stack.
This means that we don't know about the second call just with the help of
the CallStatus signal (which maps directly to the modem state) but a new
signal, CreateRequested, is needed.
Since the create request could fail before any call object leaves the IDLE
state or the modem could get reset (in which case a CreateFailed signal
doesn't help) a five second timer is used to clear the call creation
information if the active call doesn't transition to the HOLD state before
that.
|
| |
|
|
|
|
| |
It turns out that the Conference interface has actually some nice signals
for tracking the multiparty status of calls, so use those instead of
sending GetCallInfoAll method calls.
|
| | |
|
| | |
|
| |
|
|
| |
Avoid taking over as default gst sink by registering as GST_RANK_MARGINAL
|
| |
|
|
| |
usually only root is able to invoke l2ping thus move to man section 8
|
| |
|
|
|
|
|
|
| |
There are some state transitions (what these are is still somewhat
unclear) through which a call can enter ACTIVE state multiple times even
though call->on_hold isn't TRUE. The circumstances triggering this still
need investigation, but at least it's an easy one-line change to safeguard
against duplicates in the active_calls list.
|
| | |
|
| |
|
|
|
|
| |
The io channel should be stored in order to cancel connection attempt,
but this also fixes the situation where the headset connects back while
we are trying to connect to it.
|
| | |
|
| |
|
|
|
|
| |
There have been some reports of segfaults which indicate that
bt_cancel_discovery isn't working properly. To clearly catch this in the
future add an assert to the service discovery callback.
|
| | |
|
| |
|
|
|
| |
modify get_adapter_path to use FindAdapter or use DefaultAdapter if none
is given
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
For some reason (bug in csd) csd sometimes gives an incorrect value for
the on_hold property in the GetCallInfoAll reply. This patch fixes (well,
works around actually) it by trusting the call status instead.
|
| |
|
|
|
|
|
|
|
| |
There's no signal from CSD to indicate that the multiparty call status has
changed. The only way to figure out the multiparty status of a call is by
parsing the reply to the GetCallInfoAll method. So, every time a call goes
active or to hold and there is more than one call ongoing use
GetCallInfoAll to update the multiparty bit (and other properties that
come with this method call).
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
introduce optional argument target for remote address, if passed pairing
will be called
|
| |
|
|
|
| |
This patch implements and updates some parts of dbus API missing from
test/agent.c, namely Authorize, RequestPinCode and RequestPasskey.
|
