Gitlab: #226
Change-Id: I058c851795f912f261f371afe3edc2f9f09a9ff5

Change-Id: Ic61c5acaff60b015d1df92b5f1633a63f81bccd9

Change-Id: I4c6ea07a9d6734235e488a914bbaeeab01cee357

Change-Id: I72dcdecda2aab18c251f339d979bf72bda1af3e2

Change-Id: I1256fcb763bd4982714e9205078919d0d05a1e6b

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

Change-Id: Iceb7bb4ec8e710928b9399a75122db2f837d5491

Vladimir Stoiakin authored 4 years ago and Adrien Béraud committed 4 years ago

Change-Id: I010069c97af03b8128c22765eb5814b2fb0a59c2
Gitlab: #254

Change-Id: I6b97da0a32855dc8dc583a30883b31a647d20d15

Change-Id: Iae41a073cfa5a1a842d8025477539db833f506b8

Change-Id: Ibd8a93dceb1dc3fb959cdc6c874aae7da5514647

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

Change-Id: I26660bccd3ab150687c118e22edc27d9e9f9e333

Change-Id: I0af605825cd6cc80bf6bb5cbc05fdb8e81185a39

Change-Id: I0895410c9a3cb602e10ed8911545c3ad9e77b53f

TCP_KEEPALIVE are not enough to detect that TCP connections are dead or not.
Some routers doesn't seems to support TCP_KEEPALIVE and moreover, sometimes
the settings is just not enough as described in this article:
https://blog.cloudflare.com/when-tcp-sockets-refuse-to-die/

RFC 5482 describe a new option for sockets: TCP_USER_TIMEOUT which must be used
with keep alives to detect closed connection in a better way.

In this patch we set TCP_USER_TIMEOUT to 30 seconds (like the keep alive). This is
a TCP level socket option used to specify the max time in ms that transmitted data
can be unacknowledged before a force close.

Change-Id: I217616b9b9e08adfa5a044a451fbbc97faf1939e
Gitlab: #243

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

Change-Id: I64783d990cf463a5a873314c978e42a7b19bb55b
Gitlab: #228

Vladimir Stoiakin authored 4 years ago and Adrien Béraud committed 4 years ago

Change-Id: I9d1d1a53070228b93728655021cdf1764b5eb6b2
Gitlab: #238

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

Because multiple state changes can occurs while the call is in connecting,
only start the fallback check once. Also, only fallback on parent calls, we
do not care about subcall here.

Change-Id: I1ae758d8a6f9aacc2057f736c7b4173f21add11e
Gitlab: #247

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

Because we removed the eventloop to check pending calls, timeout are not checked
and only done on successfully negotiated ICE. We should re-add timeouts to avoid
any lock during the Searching or the Connecting state.

Change-Id: Id4f9a4869c171a5c3f316b76e9ecede1bc76ed97
Gitlab: #244

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

Change-Id: I81aaa48382725cc95d08ebf84ee00eb92747d3ff

Change-Id: I32788f8b817c4e38bdf1c689c751d0f0e09c4e74

Change-Id: I720705ed96b8754f039170eca132e952c8ca92a3

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

With this patch, Jami will be able to use a current opened socket with a peer
to transmit files. This means that no ICE negotiation will be necessary
and if multiple files are transferred via the same socket

Change-Id: I8eaf7c38595bbf8e86098d6c8ad21afc9210fe6b
Gitlab: #228

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

Check for endpoint existance to avoid nullptr exception.
Avoid to write multiple packets at the same time on the socket. This avoid
to mix packet.
Avoid to be able to shut the channel multiple times

Change-Id: If5158b51f55f368091616062ced4d641130c8468

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

Change-Id: I058f2230952b8dc92c85e247aa83fd4304df46bb

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

Change-Id: If6e2bbc65a4e0029d55675717cb853e0a3dcf3c4

Vladimir Stoiakin authored 4 years ago and Sébastien Blin committed 4 years ago

Change-Id: I767d7baa288edeeb93ceb16d6f2bbfff2b48d5c9
Reviewed-by: Sébastien Blin <sebastien.blin@savoirfairelinux.com>

Change-Id: Ief11a7a484973c15ff5fd03fb64240f267527b2b

other updates
* fix config test
* remove bad import (allows dringctrl.py to run, where
  previously it would result in an import error)

Change-Id: I401629956b64d2c8d51c499113f1e03b96ae3338

Change-Id: Iac4f7395a26f4d3ca42df1c65838dc04ed801114

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

Change-Id: Id3ae8c82ed11e1c6408b2733621fb5ddb58401ff

Change-Id: If1d318acf90ce34f3de8372bd9cf48c1ab0bde7a

Because setContacts() just change the contact list but never notify
the client, the client can't know the contacts from the new account.

Change-Id: I02e27f3b768dd7def661d3568a796a91eaa78751
Gitlab: #230

Send outgoing packets to the default remote address (the address
present in the 'c=' or 'a=rtcp' lines in case of RTP connections)
instead of the ICE nominated address.

Gitlab: #223
Change-Id: I4b17ee55ebd92292f994baf7a972514d95ae728b

Change-Id: Ieb7ee800f2b6f50f7ed2605a50b8dbeb32881645

Kateryna Kostiuk authored 4 years ago and Sébastien Blin committed 4 years ago

When call state is RINGING manager will enumerate through all calls
to check current call. It will cause lock when subcall and parent call
update state at the same time.

Change-Id: I0850ea685046cf3c981700e524ca184becdb96e5

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

Change-Id: I5b3200ac03212c0c082331b3ead563d4cbfa2396

Change-Id: I5693a690856cc45c9cc1a9833f544ff2f54ef963

Change-Id: Ibd1c8649483cb9a43d65a41a580968e3ca25561f

Change-Id: I43d881096bc375b9c33984c3078a04036de07737

Change-Id: I5e4e54f64152c6fb6c81dfc5b1c8890878d9fdd4