Before merging a remote branch for the conversation, each
commit is validated. This avoid to get unwanted commits.
Change-Id: I8bfe559df3ceec406293da0a586b35cf1b67ef99
GitLab: #345
Add a 'mediaNegotiationStatus' signal to report media negotiation (SD)
status (SUCCES/FAIL).
Gitlab: #380
Change-Id: Id3ecb14484466ce80be716e05071a1eb31d50c2a
They were being added by Autoconf following our own defaults, leading
to the erroneous -g -O3 -g -O2 options being passed to GCC. This is
documented in the C Compiler and C++ Compiler sections of the Autoconf
manual.
* configure.ac: Clear the default CFLAGS and CXXFLAGS variables.
Reported-by: Adrien Beraud <adrien.beraud@savoirfairelinux.com>
Change-Id: Id5e24cb2544ac4492b229746ad7ea1dee7cbb963
Link removeConversation. For this feature there is two possibilities:
+ If we are the last member, remove all the files.
+ If not, wait for another user to sync.
Change-Id: I387a20fe93a72524ae701e7e98d66633d91a9ede
GitLab: #305
Add APIs to set a list of desired media to included when making
outgoing calls.
Also add APIs to update the media of an ongoing call. The update
can be used for example to "mute/un-mute" a media, or to add a new
media (add video to an audio-only call).
Gitlab: #380
Change-Id: Iab32a72547f3a0a2328408aadc6a7656c971a7b0
When 2 conferences are merged, a host can transmit
moderator commands to another host.
Change remote layout is not supported, the layout
management will be moved client side.
Gitlab: #449
Change-Id: I1550677ef7f8babf19a729dad12a224f12226fbd
Possible workaround for issues with "no response from DHT". It can be a
blocked mutex, but I can't reproduce, so this is a tentative to improve
the behavior and to be sure it's not a locked mutex there.
Change-Id: I4c8f152be08f4243ef1654cd36d661989783be0b
GitLab: #421
Some interfaces (typically from docker or virtual machines) were added
in the candidates even if down. For example:
9: br-b8251442502e: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue
state DOWN group default
Even if state=DOWN if_enum_by_af consider the ip as valid and adds it in the
candidates because IFF_UP is true. However, if_enum_by_af should use IFF_RUNNING
to detect if the interface is up and running before adding it. Like
iproute2/bridge/link.c does for print_link_flags.
Resource: https://man7.org/linux/man-pages/man7/netdevice.7.html
Upstream ticket: https://github.com/pjsip/pjproject/issues/2686
Change-Id: I168954509ec935bb45e39700e57a0c12ac837931
GitLab: #495
If the peer sends too many remote candidates, the negotiation
will just fail, but should not, as we have enough infos (even too
much) to connects.
So, instead of failing, just construct a better list.
Change-Id: Ic8ad86b2b97e04aca9a5660ab2f65068b01a928f
GitLab: #495
This patch introduces three major changes:
1- When a "Connectivity Change" notification is received from the
configuration manager, the current IGDs and the provisioned mappings
were reset and a new search and provisioning was started anew.
Now, the reset/restart is performed only if the host address actually
changes.
2- Many code simplification and clean-up. Obsolete code used to
handle asynchronous operations on libupnp was also removed.
Most of the calls to libupnp are now synchronous. Blocking calls
are dispatched on the internal execution queue (worker thread).
3- Add synchronization to avoid thread race at shutdown
Gitlab: #416
Change-Id: I26970d78db494795f23379c7a9af5a20a665bb06
This reverts commit 17f9bda885.
Reason for revert: major design issue. When plugin are disabled, every plugin will be disabled, then every plugin are unloaded, so no handler can remain after disabling plugin.
Change-Id: Icd393d179b45866e46dd4cc3d902174699ac2d98
This prevents downgrading the audio experience to ALSA on system
running a vanilla pulseaudio configuration that relies on pulseaudio
being able to manage the pulseaudio server itself.
* src/media/audio/pulseaudio/pulselayer.cpp (PulseLayer::PulseLayer)
[PA_CONTEXT_NOAUTOSPAWN]: Replace flags argument value with
PA_CONTEXT_NOFLAGS.
* src/media/audio/pulseaudio/pulselayer.h: Fix typo.
GitLab: #473
Change-Id: I219223dd8a1237a51b3e5bf60d6535518d748b63
On recent gcc, daemon fails (static assertion failed: result type
must be destructible) cause futures must access the destructor
since commit/71ed3c0c9a3458998bded8e2443c0a680c2eb8cd in gcc
Change-Id: I1e7d35da6aeaffffcb6aa9362d335625ea370052
