Change-Id: I6463e8315a84df28fb7a7e8b52f6c8e18899fa44
GitLab: #359

Kateryna Kostiuk authored 4 years ago and Adrien Béraud committed 4 years ago

Change-Id: Icbeffec03c80929f1122a5a02df661124fa683a5

Aline Gondim Santos authored 4 years ago and Adrien Béraud committed 4 years ago

Change-Id: I2b59205b9ad2f75c6c8602a5a273a7d8115e5db6

avoid segfault. Backtrace from the play store

Change-Id: I5a44a5ee190cba22ed31ef4c7f64703b352db7fe

The temporary ice session created by sipcall, might not be
properly released if the call is rejected by the user while
searching/connecting.

Change-Id: I053c2a4c4bd864c4dd4c73488175da30d1bf08db

Change-Id: Iee0b40fd29704f115dbb56f4357de9153f2e2e36

Édouard Denommée authored 4 years ago and Adrien Béraud committed 4 years ago

Change-Id: Ia7fdbfb41b27f3eacdddaa861041e966e455da0c

Sébastien Blin authored 4 years ago and Adrien Béraud committed 4 years ago

getSIPAccount() is returning a reference to the SIPAccount. Account
that can be destroyed and if a mistake is made, the data can be
invalidated, generally even when closing the daemon.

This patch change the behavior of getSIPAccount() to return a
weak_ptr to make sure that the Account is alive when performing
actions on the call

Change-Id: I1dcc05bce0c9826297535b5a5224f9ff22479d80

Since all is async, sometimes, it's possible to call setupLocalSDPFromIce
before that initIceMediaTransport is done, stopping the call in a weird
way.

Change-Id: I5e329a4a44db61e24ac559535a4f6c72f4527ae4

This avoid to wait on Ice initialization for 35 seconds if anything goes
wrong and block the main thread.

Change-Id: I6268802a341b0c366fddcf240f9abf6f1c7b36ad

Kateryna Kostiuk authored 4 years ago and Adrien Béraud committed 4 years ago

Change-Id: Ic034b8d53a7eea29b1230cea0b7dfece06ad40de

Pierre Lespagnol authored 4 years ago and Adrien Béraud committed 4 years ago

- Allows conference host to notify each participant when recording
- Remove unnecessary callback
- Notify host if remaining participant is still recording after conference
- Update host notification if participant leave the conference

Change-Id: Ib17675442f2e175fe8711c695bc737ffa6d89f05

Kateryna Kostiuk authored 4 years ago and Adrien Béraud committed 4 years ago

When call unpaused after microphone was muted, mute state changes to
unmute in call, when in audio input it still muted. After this all attempts
to mute/unmute audio are unsuccessful. This patch removes
mute state update when restarting media, since audio muting managed only
by audio input, and should not be related of rtp session parameters.
This patch left video mute state updates.

Change-Id: I4ac3bea04865fa23b876d5c10fcb5c172c4501a3

Pierre Lespagnol authored 4 years ago and Adrien Béraud committed 4 years ago

Change-Id: I2acae919d2a600e8125e203c169be057ec8fda81

Add user-agent header to most relevant SIP messages (INVITE, RINGING, OK, ...).
The user-agent string is generated from the build version. The user-agent
header is useful to identify build version used by the peer.

Change-Id: Ic99263ca147d3f96c3c092be2ff3d723bdb6fe6f
Gitlab: #287

Pierre Lespagnol authored 4 years ago and Sébastien Blin committed 4 years ago

Change-Id: I711a0b2f446f1620d7dff66945bf9295fd8372ac

Aline Gondim Santos authored 4 years ago and Sébastien Blin committed 4 years ago

Change-Id: I4620a3a2704ea9bca661f00e0f7e6da55f2cb909

Change-Id: I2dd97518218a5a777ddde2ce8d23e2d5dde4dcf0

Sébastien Blin authored 4 years ago and Andreas Traczyk committed 4 years ago

Change-Id: I193328586ea401c16d64393eb0dfb2ebb0d80e9d

Sébastien Blin authored 4 years ago and Adrien Béraud committed 4 years ago

Destroying an ice session take quite some time and must be done in
the io pool

Change-Id: I159d249927a3aecbff640143902c8a6d33a52282

Aline Gondim Santos authored 4 years ago and Adrien Béraud committed 4 years ago

fix preferences set in first plugin loading
small logic changes in libraries loading
related to: https://review.jami.net/c/ring-project/+/14433

Change-Id: I4730f279ee38d337e3ef94dd9562a5dfc1a63c32

Mohamed Amine Younes Bouacida authored 5 years ago and Adrien Béraud committed 4 years ago

Change-Id: If5296e71d4979962f71443fb298891202c8d2afe

Ice destruction can take too long and delay media initialization.
This causes first packets sent by the peer to be missed.

This patch fix multi-device call bug with Android (and MediaCodec)

Change-Id: I4ceb1838e668d5be2ed36d1e0d2f00113833f63c

Change-Id: Ia12da6f9e58ec0cd0100eaba206e1d9e711ca0e1

This patch move checkAudio() out of call mutex. Because it could cause
deadlock if subcallStateChanged() called at the same time.

Change-Id: Iee06874fb79c8e0953e061f7a669b3f61f13d8b9

Sébastien Blin authored 4 years ago and Adrien Béraud committed 4 years ago

The current design describes SIPVoIPLink as a Singleton. This cause
some behaviors where the link is destroyed then immediately created
or vice-versa and its creation/destruction can't really be planned.
This design was made to allow multiple Managers to exist. However,
we never do that, so let's keep it simple, there is only one
Manager and all accounts needs that SIPVoIPLink

This patch change this behavior and the voip link is now owned
by the manager.

Change-Id: I248e41742d342cf452a5503f532fe5ab862166e6

offHold can return false even if the action is pushed in the
requests. This means, that the audio will not be binded to the
call if offHold is waiting for ICE.

Change-Id: I0fbdf4682a613919cee1ae5b2242077d1bd57fd7

During incoming calls to iOS, video input failed to open sometimes.
It happens for two reasons:
1. Stopped by interrupt callback.
2. Keyframe callback was set two late.

This patch:
1. Completes decoder configuration after video loop start running.
2. Sets keyframe callback for video session before starting.

Change-Id: I00e1ace06a78ab2ec2fb3ce6fe6a5991c6593318

This causes the incoming state to come after the current state sometimes and
there is no need to run onPeerRinging on the main thread

Change-Id: I0d98c22b298b1bc5a2e1797ecae4855fa767fec8
Gitlab: #209

This patch avoid to block during a long time in shutdownConnections.

Change-Id: I090e309d4aa5cffcb9bb69b2f78a7769bb204f68

Sébastien Blin authored 5 years ago and Adrien Béraud committed 5 years ago

Change-Id: Iada9a972ece86ea2536b9baa0b9e07b9bb2c35fc

Change-Id: I28f193813ad72ee5936ef0be7f252c40c3b748ee

Sébastien Blin authored 5 years ago and Adrien Béraud committed 5 years ago

pjsip uses a counter to delete objects when the ref counter is equals to 0.
This means that our unique_ptr on the invite will be invalid if resources are
already freed by pjproject. To avoid this, we need to increment and decrement
the counter when we respectively create and destroy our unique_ptr on the
invite session

Change-Id: Ida5c687004b91100f1c10f83e32c1a40264c775c

Change-Id: I1ad6f00a44ceeac14929ca573dd130a678a23a31

Adrien Béraud authored 5 years ago and Sébastien Blin committed 5 years ago

Change-Id: Ie1d1120957fe3e8bb38621b5ffa38f8d1e407b62

Change-Id: Ie1ff7315e0d5dca1a63cb16aac4a3a4bff09415f

Change-Id: Iac06d0897b37992daa959f4adf8bd1fbde33215c

Sébastien Blin authored 5 years ago and Adrien Béraud committed 5 years ago

im::sendSipMessage is locking dialog when callMutex_ is locked, but
handleEvents can lock dialog then callMutex_ in startAllMedias. This
will give a deadlock, so startAllMedias should be rescheduled

Change-Id: I3d1b0b9a981419840f19bfe1832b70503713fe72
Gitlab: #175

Sébastien Blin authored 5 years ago and Andreas Traczyk committed 5 years ago

The UPnP stack now opens and closes ports with an
asynchronous mechanic and with callbacks.

Every time a mapping request (open or close) is sent,
a callback is registered and managed by the context.
When the corresponding protocol executes the request and
gets an answer, the right callback is found and dispatched
to its controller. A timeout of 1 second is implemented
on each request.

Controllers can now be tracked with a unique Id (a string
that represents its location in memory).

The IGD class now tracks its current active ports and
manages the number of users for each port.

The PMPIGD class now operates using a queue of mapping
requests. Whenever a request is made to the natpmp protocol,
it saves the request to its corresponding list (mapToAdd,
mapToRemove or mapToRenew in the PMPIGD class) and notifies
the main natpmp thread. The thread then checks if there
is any pending map requests in the PMPIGD queue and treats
them accordingly.

Add Nat-Pmp support for windows.

Fixed bug where nat-pmp wouldn't find an IGD on a new interface
after a connectivity change. The problem stems from the fact
that, for one, the nat-pmp handle didn't get cleared internally
once a connectivity change occured. This would cause the nat-
pmp library to try and send searches on a socket that was
bound on the old gateway. Secondly, if you don't wait for a
little period of time, the natpmp library would discover the
old IGD on the old interface before the connectivity change
modifications were completed. Now we restart the nat-pmp
process one second after the connectivity change triggers.

Change-Id: I39a9b57e237deaa65eef4464a1838eaa3c5bbb03

Change-Id: I727a0040b4a80afa810db7b6042a3e7ec093f4b4