Change-Id: I5a30a4f5b7c7f2b74e5457973fdf42849fbc604c
GitLab: https://git.jami.net/savoirfairelinux/ring-project/-/issues/1282

SipTransportBroker::shutdown triggers pjsip_transport_shutdown
that can call transportStateChanged synchronously. In this case,
transportMapMutex_ is locked, causing a deadlock on a single thread.
We don't care about the transport's map at this point as we are
destroying the transportbroker, so we can just ignore the state
changed here.

Change-Id: I180ff1807fe87d1c48d72851f17e94ae2f892823

The handleEvent should finish. If after 3 secs nothing it's still
running, it should be stopped anyway. Add a log to detect this case.

Change-Id: Icc64f9450d5a8aecfcb6a5aa2e6a13882f4d0e8b

With this patch, contact left is removed in one to one. So, if
a contact left, they still receive message notification and they
will generate conversation requests and retrieve the old history.
Moreover, if they remove and re-add a contact, they will receives
a confirmation for the previous conversation and retrieve the old
history.
This avoid duplicates and read-only conversation.

GitLab: https://git.jami.net/savoirfairelinux/ring-project/-/issues/1282
Change-Id: Id820482e82e23334632e76bf4920255efb45f4cb

Olivier Dion authored 3 years ago and Sébastien Blin committed 3 years ago

Change-Id: I8d5eaf4c90cb425e8f96eaac25e727b6a6aae24f

Do not emit CallSignal::StateChange while locked. Because if the
client uses any method that lock callMutex_ it will be a deadlock.

Change-Id: I49aa7cb4ee60b23ef069751a7c1abbf4c9fcec95

This makes the log sending messages in the order the client will
show it and avoid to calculate linearized parent for each commit.
However, this linearized parent can change over time. In fact, if
a merge happens, the parents of the messages shown can be
modified.

Change-Id: I26372aa52de0e5f89140680a4beaf5fc7ec52163
GitLab: https://git.jami.net/savoirfairelinux/ring-project/-/issues/1282

Change-Id: I295f444373aaf00ca11e56dd594e61548237364e

This patches introduces two new concepts in order to reduce the code
of JamiAccount.

ChannelHandlers to manages protocols logic. The idea of this class
is to handle channels per protocol, accept and reject it.
AccountModule, to be able to separate logic between call managements
datatransfer, config and conversation but give an interface to
detect when some events should occurs.

Change-Id: I34ff07852c06d7266411f1ffb32b71a1834aba4f
GitLab: #603

A deadlock is caused by a lock-order-inversion of the private
mutex from pulse-audio thread loop, and the callMutex_ mutex
from the Call class. It occurs when concurrently calling
writeToSpeaker() and onNegoDone() callbacks.
The deadlock is prevented by avoid to a access Call data from
the pulse-audio thread.

Gitlab: #623

Change-Id: I91d936b37528db3de24e93b0d49d686f5ca11813

Change-Id: Ib4377b5a109dc9b1554a35f2341820307ec6a500
GitLab: #624

There is an issue with the nomination of triggered checks in ICE
that causes ICE negotiation to conclude on sub-optimal pairs.
The issue was reported, and acknolegded by PJSIP maintainers.
This patch will fix the issue.

Gitlab: #622

Change-Id: I4bed7191692051d6215fbde8ee2ca9f94b76e0f9

Use a higher Ta value (50ms instead of 20ms) as
recommended by newer ICE spec (RFC-8445).
This will reduce bandwith usage (peaks) needed for the connectivity
checks as described in RFC-8445.

Gitlab: #605

Change-Id: I6a228a47d5f6cf574003d1354c4410294aedf95a

Aline Gondim Santos authored 3 years ago and Sébastien Blin committed 3 years ago

Gitlab: #562

Change-Id: I577c86b3e6a30c171f0233c4354ccfb79e494d40

fixes commit b954f3b5

Gitlab: #617

Change-Id: I82b6f9f229b173279bfcd3c200b73690d752b670

Change-Id: I7e46255256ebce1a1e133c53d170539f6c8e1f8c

Sébastien Blin authored 3 years ago and Mohamed Chibani committed 3 years ago

Change-Id: Icee8afc872c9b25fce2d049e66673459de254b4d
GitLab: https://git.jami.net/savoirfairelinux/ring-project/-/issues/1282

Change-Id: I505b63153f49aa9733d278c613c9813c67febb62

Change-Id: I384bca2ca4e97588b21041ebca72083c57abb915

Mohamed Chibani authored 3 years ago and Sébastien Blin committed 3 years ago

Resort the check list  when new prflx candidates are discovered,
to avoid the nomination of candidates pairs with less priority.

Gitlab: #617

Change-Id: I93559facd8b6ce99708ccf1311925837d9cd119d

ICE "aggressive nomination" was enabled to reduce connection time.
However, this feature may cause ICE to nominate a sub-optimal pair
during the negotiation process. For example, nominating a
relay-prflx pair, while a srflx-prflx is more optimal.
Furthermore, aggressive nomination was dropped in the latest ICE
specification (RFC-8445)
Also fix logging of nominated pairs

Gitlab: #601

Change-Id: Ib08103b830fc0c5d158cf776a1082a0655317270

When receiving a 'connectivity change', UPNP search must be restarted
if the host address changes and also if there is no valid IGD.
Also reduce the number of retries on failure.

Gitlab: #615

Change-Id: I80e6d61ab79ef91f480bb0073b9cec8e7e67e559

This brings regression and unit test are failing

Change-Id: I0bd7373bd75f502b2f08f55932a6b6c4210d31d2

Change-Id: I5924025c4c9e37f7ebe554151b39a8da5eb5154e

Change-Id: I692df9c8e315e703fb079e136962905b9e83dce8

Change-Id: Ief1fd8fc1058f77784260d5be7a116d13ce06480

Change-Id: I7de1fb513238707930d500472bb6d8b2139aba2a

Change-Id: I08dc1c58529872c031ce9e5c36706173857f3f83

Change-Id: Ie6c21d1505f218a0de3bf79aec4565fc3713a922

Change-Id: I748bcb66be07c70021c422a6d04627bc62591e6a

Ming Rui Zhang authored 3 years ago and Ming Rui Zhang committed 3 years ago

https://datatracker.ietf.org/doc/html/rfc8599

Follow rfc8599, when the sip account registers, it will add
 1. pn-provider
 2. pn-param
 3. pn-prid
parameters into the Contact header.

pn-provider will either be fcm or apns
pn-param is not used currently
pn-prid will be set from setPushNotificationToken api

Once pushNotificationReceived is called, all sip acc that are enabled
will re-register.

Change-Id: If39a8183e7f45b76ec71641f45805e48690943bc

Gitlab: #612
Change-Id: I923c6cf4dcbe94cdf3667717f977e21f8d1b7ce1

Change-Id: Ia16f429032fd940b52393d946400300fe269048e

Ming Rui Zhang authored 3 years ago and Ming Rui Zhang committed 3 years ago

Change-Id: I0b05cf38d6a01711c994b124d002795bcd9d6a96

Olivier Dion authored 3 years ago and Sébastien Blin committed 3 years ago

Change-Id: I3840037c7710a6a4874ea03af49b139e314dcd95

Ming Rui Zhang authored 3 years ago and Sébastien Blin committed 3 years ago

Change-Id: Ie40d86f8ceac139d6d55fcbf81fc26e252d95a04

Olivier Dion authored 3 years ago and Andreas Traczyk committed 3 years ago

Variables `stun_sock->*_nb` are initialized to -1 and incremented by 1 for each
new socket.  Thus, the `<=` comparison operator should be used for iteration
instead of `<`.

Gitlab: #613
Change-Id: I3c6971f533ead466b25381f3d4edbc4dc1827950

Change-Id: Ia8ea4818d7966c80be32e44a26d7d84fc4ba008e

Olivier Dion authored 3 years ago and Sébastien Blin committed 3 years ago

Change-Id: I790b47971c6c202f7efeeb8c10db00ef430ca39c

Maxim Cournoyer authored 3 years ago and Amin Bandali committed 3 years ago

This is to prevent resource contention on shared file systems such as
NFS.

* contrib/src/main.mak (download) <wget>: Guard with flock.
<curl>: Likewise.
<fetch> Likewise, and properly quote shell arguments.
(download_git): Likewise.

Change-Id: I8e2139d780a34ad054cfa8e188c6cbaf94df0609