summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Version 0.2.40.2.4Olivier Crête2014-05-052-2/+9
|
* msnconnection: Fix typoOlivier Crête2014-05-051-2/+2
|
* msnconnection: Double check return value of recv()Olivier Crête2014-05-041-3/+10
| | | | Even though it has already been peeked at!
* rtpsession: Check that there is either a blueprint or a profileOlivier Crête2014-05-041-0/+6
| | | | Having neither is always invalid!
* msnconnection: Make sure token is correctly readOlivier Crête2014-05-041-1/+5
|
* rtp: Check both variants, not only one!Olivier Crête2014-05-041-1/+1
|
* multicast: Assert that udpsock is valid if there are ttls leftOlivier Crête2014-05-041-0/+2
| | | | | If the sock is not valid, that we should be the only user and the ttl should have been flushed.
* rtptfrc: Fix off by one errorOlivier Crête2014-05-041-1/+1
| | | | 128 is dynamic and needs checking
* rawudp: udpsock is never NULL thereOlivier Crête2014-05-041-2/+1
|
* multicast: udpsock is never NULL thereOlivier Crête2014-05-041-2/+1
|
* rtpkeyunitmanager: Correctly check for local ssrcOlivier Crête2014-05-031-2/+3
| | | | Found by coverity
* rtpsession: Since GStreamer 1.2, the real internal SSRC is on the incoming capsOlivier Crête2014-05-031-3/+18
|
* rtpsession: Also notify of SSRC change on caps changeOlivier Crête2014-05-031-0/+11
|
* tests: Disable upnp tests by defaultOlivier Crête2014-05-031-1/+1
|
* Fix documentationOlivier Crête2014-05-024-3/+36
|
* docs: Use Farstream API version when installing docsOlivier Crête2014-05-013-2/+233
| | | | This also forces us to make a private copy of the gtk-doc.mak
* include <sys/socket.h> for setsockopt(2)Jasper Lievisse Adriaanse2014-03-301-0/+1
|
* Include <netinet/in.h> for struct sin_addr.Jasper Lievisse Adriaanse2014-03-301-0/+1
|
* Include <sys/uio.h> for struct iovec.Jasper Lievisse Adriaanse2014-03-301-0/+1
|
* raw: Fix crash where the stream would try to contact its session before its ↵Olivier Crête2014-02-271-1/+2
| | | | been set
* transmitters: include <netinet/in.h> for IPPROTO_*Ryan Lortie2014-02-142-0/+2
| | | | | | | | POSIX says that we need <netinet/in.h> for IPPROTO_* to be defined, so make sure we include it. It also ensures that we get a definition of 'struct sockaddr' which appears in the rawudp header as an argument type.
* rtpsession: Need to read the method on stop tooOlivier Crête2013-12-051-5/+4
| | | | Otherwise it used an initialized variable. Thank you clang-analyzer
* examples: Remove unused variableOlivier Crête2013-12-051-2/+2
|
* rtp-codec-nego: Actually test that the codec id is validOlivier Crête2013-11-071-1/+1
| | | | Bug found by David Binderman <dcb314@hotmail.com>
* tests/rtp/sendcodecs: pass a GError to parse_launch.Mathieu Duponchelle2013-10-161-1/+3
| | | | | Otherwise for some reason launch_full returns a pipeline even when an element is missing, despite the FATAL_ERRORS flag.
* Add files from newer autotools to .gitignoreOlivier Crête2013-09-051-2/+8
|
* multicast: Remove not required non-standard headerOlivier Crête2013-06-251-1/+0
|
* tests: Only use matching host candidates for nice force_remote_candidates testOlivier Crête2013-06-201-10/+18
| | | | The other candidates may or not may not work
* Prefer dynamic PT 101 for telephone-event at clock rate 8000Simon McVittie2013-06-031-0/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | The WebRTC implementation in Google Chrome <= 26 would reject calls if there was "a telephone-event payload type less than 101"[1] and as of 2013-06-03, the Google Mail web UI with the VoIP extension seems to have a similar signalling bug. Experimenting with the web UI indicates that telephone-events with clock rate != 8000 are irrelevant, and only clock rate 8000 matters. Hopefully the same was true in WebRTC (I can't find a libjingle commit that looks likely to have fixed this). Meanwhile, many SIP implementations and at least one Jingle implementation (Freeswitch's mod_dingaling) either hard-code payload type 101 to be telephone-event, or make the payload type for telephone-event a configuration option. I can't help thinking this was not how dynamic payload types were meant to work, but interoperability is interoperability... This fixes interop when Empathy 3.8 + telepathy-gabble 0.17.4, on a system with not many codecs installed) calls the Google Mail web UI. When the same setup is called by a peer that specifies a different PT for telephone-event:8000 (the Google Mail web UI uses 126 in its outgoing calls), the peer's choice of PT takes precedence. [1] https://code.google.com/p/webrtc/issues/detail?id=1783 https://bugs.freedesktop.org/show_bug.cgi?id=65311
* Version 0.2.3.1Olivier Crête2013-04-151-1/+1
|
* Version 0.2.30.2.3Olivier Crête2013-04-152-2/+13
|
* tests: Use G_GSIZE_FORMAT where appropriateOlivier Crête2013-04-153-6/+6
|
* fs-rtp-discover-codecs: plug memoryleakHavard Graff2013-04-081-1/+1
| | | | use g_list_delete_link to free the list as well
* rawudp: Use GSocket abstraction for portabilityOlivier Crête2013-04-042-112/+65
|
* multicast: Use gio instead of getaddrinfo for resolvingOlivier Crête2013-04-041-14/+23
|
* tests: Use GSocket instead of getaddrinfo to parse IP addressesOlivier Crête2013-04-044-33/+38
|
* Misc win32 portability fixesOlivier Crête2013-04-049-16/+18
| | | | Based on a patch by Conrad Poelman
* codec-discovery: Intersect different parts of the same caps to reduce themOlivier Crête2013-04-021-6/+34
| | | | | We do this because a caps may have the static payload in a separate structure from the encoding-name We just want both in the same structure
* rtpsession: Set error in all error casesOlivier Crête2013-04-021-2/+12
|
* rtpsubstream: Don't free codec after setting it inside substreamOlivier Crête2013-03-291-8/+6
| | | | Bug discovered by Havard Graff
* session: Add API to set the transmitter parameters as a GHashTableOlivier Crête2013-03-263-14/+114
|
* candidate: Add helper function for pygiOlivier Crête2013-03-262-0/+19
| | | | This way, it can set a FsCandidateList to a GValue
* candidate: Allow various elements to be NULLOlivier Crête2013-03-261-5/+5
|
* fs-codec: plug memory leakHavard Graff2013-03-261-2/+0
| | | | encoding_name is already g_strdup'ed in codec_new
* rtp-stream: plug session leakHavard Graff2013-03-261-0/+3
|
* fs-rtp-substream: Don't leak caps on errorOlivier Crête2013-03-261-0/+2
| | | | Based on patch from Havard Graff
* rtp-codec-discovery: Intersect instead of mergeOlivier Crête2013-03-261-6/+4
| | | | | We want the semantics of intersection, not merging, as this will produce a caps with two separate structures in some cases.
* Use == instead of = test for portabilityOlivier Crête2013-03-221-1/+1
| | | | Bug reported by OBATA Akio
* rtp: Tune pulsesink/pulsesrc latency values furtherArun Raghavan2013-03-221-3/+2
| | | | | | This makes for lower overall values without forcing a bunch of underruns at the start which we got by having pulsesink's buffer-time as 2*latency-time.
* Remove deprecated GStaticMutex usageOlivier Crête2013-03-225-21/+21
|
View Lorry Controller Status