- Oct 23, 2021
-
-
Change-Id: Ic5862a3e7ce879d8015a50dd348489c81b893603
-
ASAN does not like this: -------------------------------------------------------------------------------- runtime error: reference binding to null pointer of type 'const value_type' -------------------------------------------------------------------------------- Change-Id: Idca4720dda401c756bd6ea5dc054dd9e63779153
-
- Oct 22, 2021
-
-
Sébastien Blin authored
If a conversation is not cloned, when detecting a new presence, needsSyncingWith MUST return true if one conversation needs to be cloned. Else, it will result in an infinite syncing. Change-Id: I3a9bcab9f8a845161ceac7d75d8842f8b338ed57
-
- Oct 21, 2021
-
-
Changing the media source of a media stream (video in particular) must trigger a reinvite. Since currently we do not support changing video properties (resolution in particular) in the same media session, any changes in the video properties must trigger a re-invite to restart the media locally and on the remote peer. Gitlab: #653 Change-Id: Ibd6192135d272908963cfbc585bec95c66881a28
-
Sébastien Blin authored
acceptTrustRequest will set the conversation as ready. removeConversation should be done after Change-Id: I966cc2142d00bda880d63f11c52c36faceb11aa7
-
Sébastien Blin authored
A ChannelSocket can starts to receive datas while waiting acceptance for the channel. In this case, received datas will be injected when setOnRecv will be called on the channel. However, in channeled_transport, setOnRecv injects data to pjsip and is handled by the SIPVoIPLink where transaction_request_cb will retrieve the infos via the transports_. This means, that, in order to successfully handle data, we MUST inject data when transports_ is correct. In this patch, SipTransportBroker::getChanneledTransport is modified to connect the callbacks after adding the transport to the map and inject data in the correct methods. This fix sporadic failures for testInviteFromMessageAfterRemoved Change-Id: I2767801a9dad77439fb2f2adedbc9b900add8cea
-
- Oct 20, 2021
-
-
Change-Id: I724883bf523ce7670f87979c884d97a86688cc38
-
- Oct 19, 2021
-
-
Adrien Béraud authored
Change-Id: If306734d2c29c017115e829b3fb4811333cea1f4
-
Adrien Béraud authored
* use fmt for better performances * fix dtls * remove double space after display name Change-Id: I9daeb69c65838b130432dfee0b4a5f92d0e53ad9
-
Change-Id: I53308fca61c46551129320babc99915d927a81dd
-
- Oct 14, 2021
-
-
Mohamed Chibani authored
The contact header field was stored in a member varialbe of Jami/SIP accounts. There was a condition in which the contact header could be concurrently accessed by multiple threads. This typically happens when terminating calls in batch (when calling hangupCalls() for instance). Managment of SIP contact header in Jami and SIP accounts was reworked to prevent such race. Gitlab: #633 Change-Id: Ib9295070a5295969bf114ec29e66e36b1c5c5e03
-
GitLab: #644 Change-Id: Ia2b8e05d6ffbb22336bb284ea9ac5fc292eacb85
-
- Oct 13, 2021
-
-
GitLab: jami-client-ios#145 Change-Id: Ic4ae63ce2f1df1176c4a74339d1c3d82a9570ad5
-
Change-Id: I2bcbbf872b54262f8e4406fb01921a1ec07be3b3
-
Change-Id: I09876059293d18e25c34e755deafba8d224dc740
-
Change-Id: Ib5194ef6ca475df5be621c923d7660266cf7256e
-
Change-Id: I0c44b1a5fdd9727a8acf3d9deb88f9c2b5f71dd8
-
Change-Id: I1b21d80b7ea24bfc9334f2f63f3a57f0d4a57522
-
Change-Id: I59d9d812b6b5a2c66e2832328fbfdf663d9019af
-
Change-Id: I42fe3a4682bb9468ecf5417d6d9823e151e10bec
-
The agent will be written in Guile Scheme instead of C++ in order to give maximum flexibility to developers of scenarios. Thus, bindings will be added for the public Jami interface. In other words, the agent will be a client of yet another Jami's bindings. Change-Id: Ic2cd333007d0b1aad56c02b116ea708b56e96cc5
-
- Oct 12, 2021
-
-
Adrien Béraud authored
Change-Id: Iab2caaa3fb8377b7a42301c3635d70233f0cc623
-
Adrien Béraud authored
Change-Id: I2c04e3831f68a55b7682c42eb93464e1fab78b1b
-
Adrien Béraud authored
Change-Id: I4bdef901c2a2bc07dc1a9263eb0a9796655c051a
-
Change-Id: I55568eac3c95105131751d245224fbc3fd44a782
-
Adrien Béraud authored
RingBuffer has its own mutex. This avoids potential deadlocks. Change-Id: I93175aafe9fbb3f927da28124a7d42e61585523d
-
Adrien Béraud authored
Change-Id: Idde24be13daf0fca55ce5a2607e6b9835c18f3c5
-
Change-Id: Iaa356f723a34c67cfd66190c04a53aea88bbea07
-
Adrien Béraud authored
Change-Id: I30b7d9db2cbac84c6ae15d15b0fe8a0f93e3a271
-
- Oct 11, 2021
-
-
Adrien Béraud authored
Change-Id: I3b0fd2a34289545dbf06b85c2e825fbf87bb42ad
-
Adrien Béraud authored
Change-Id: Idbb8720de936661b01fcea97faeab92d0526b6c2
-
Hardware decoding will still be used even if the codec level is unknown or higher than the hardware's reported supported level. From FFmpeg: "It's generally a good idea to pass this flag unless you have a specificreason not to, as hardware tends to under-report supported levels." Change-Id: I83614b0eab6166a42973f062e666783c2a8e7627
-
- Oct 10, 2021
-
-
Change-Id: I19bc2fe4edef3c8dc7992473235f6ffd002c6e0b
-
- Oct 07, 2021
-
-
Adrien Béraud authored
Change-Id: I12148f354905cd888e284ab5583a80f9fe721248
-
Adrien Béraud authored
Change-Id: Ieb61181e437278e1975fadfa9908c0bc413b1add
-
Adrien Béraud authored
Change-Id: If641609b25cf6dad1a2c91fe534c2fa3a7119ad1
-
Adrien Béraud authored
Change-Id: I5e328cc2d0e38aa26bf62cb3a5bd083d499b9a02
-
Adrien Béraud authored
Change-Id: If509355da31d71e9301d463fbbd64c202530cdfc
-
- Oct 06, 2021
-
-
If UPNP is enabled by the user, the UPNP controller must registered only if the account is enabled. Gitlab: #639 Change-Id: I993b50ea37fafa847460650d901ac3c3d5039718
-
- Oct 04, 2021
-
-
Add logs to help debug ICE intialization issues Gitlab: #619 Change-Id: If81067e2261cf7fbb054b4c9abe641e2ebc84a0a
-