Version 0.0.17
tests: Add test for telephone-event events parameter nego
rtpspecificnego: Add handling of telephone-event event ranges
tests: Skip tests if no local candidates are produced
rtcpfilter: Reduce the packet size when reducing the packet
tests: Skip libnice tests if it finds no local candidates
rtpdtmfsoundsource: Respect the ptime/maxptime too
tests: Add test ptime/maxptime passing
rtpsession: Set the ptime/maxptime on the send codec bin caps
rtpcodecnego: Negotiate the ptime/maxptime
rtpconference: Add function to make gst caps while keeping the ptime
rtpcodecnego: Add function to copy the list of codecs with the send-side ptime
tests; Add test for fscodec ptime/maxptime handling
codec: Add ptime
codec: Add maxptime
tests: Take rtpsession lock during message emissions
This ensures that it is not held across message emissions.
tests: Add debug-blocks
rtpsubstream: Keep ref on substream while callbacks are invoked
rtpsubstream: Put codec/codecbin inside loop
rtpsubstream: Use rw-lock to make sure the substream really stops
rtp: Move locking into callback
rtpsubstream: Don't hold session lock too much while setting new codecbin
rtpsubstream: Move modification locking to blocked function
Also allow only one thread to be in substream blocked function at once.
rtp: Move substream blocking logic into substream
rtp: Don't include marshaller headers in headers
rtp: Depend on the correct var for marshaller list generation
rtcpfilter: Add gst-p-base paths to Makefile.am
Patch contributed by Armijn Hemel <armijn@loohuis-consulting.nl>
rawudp: Remove upnp-request-timeout, it was a terrible idea
Substitute deprecated Glib symbol: g_mapped_file_free
Use g_mapped_file_unref if Glib >= 2.22 is available
http://bugs.freedesktop.org/show_bug.cgi?id=21422
rtpsession: Only add stream to list if its creation worked
README: Require gst-p-bad 0.10.17 for dtmfsrc
dtmfsrc can do do more than 8000 Hz, that has only been fixed in
gst-plugins-bad 0.10.17
rtpdtmfsound: Try hardwired PCMx only if the clock-rate is 8000
rtp: Lookup codec with config is always for sending, so make it explicit
Also, the dtmf sound will always get a valid codec now.
rtpconference: Make message about gst_bin_add failure more accurate
rtpdtmfsoundsource: Ignore codecs that don't have a blueprint
tests: Test dtmf as sound
tests: Make recv-pipeline per test
rtpdtmfsoundsource: Use main codec if PCMA/U are not available
rtpspecialsource: Make local class_get_codec function static
rtp: Regroup CodecBlueprint related functions in one place
rtpspecialsource: Rename negotiated_codecs to negotiated_codec_associations
This way, the list contents can be guessed
rtpsession: Don't need to set queue-delay anymore
rtpsession: Split codecbin generation from factory from profile
tests: Make it build against GUPnP 0.13
msnsession: Check if dispose has already been called
fstransmitter: uint can't be < 0
rawudp: Bring upnp discovery timeout down to 2 seconds
tests: Verify that it is not possible to disable all codecs
Add a reserve-pt to guarantee that it is not possible to disable all codecs
rtpcodecnego: Verify if there are any valid local codecs left after applying preferences
rtpsession: Make error message less cryptic
Version 0.0.16.1
Version 0.0.16
rtpspecialsource: Remove want_source() method
get_codec() function does the same thing
rtpdtmfsoundsource: Implement get_codec method
rtpdtmfeventsource: Implement get_codec method
rtpspecialsource: Add new get_codec method
rtp: Check if the codec changed when removing special sources
rtp: Allow checking if a codec is valid for sending even if it has no way to build a codecbin
rtpcodecnego: Fix doc string
rtpspecialsource: Move static function closer to its use place
rtpspecialsource: Fix over-80 line
rtpsession: Check/update secondary sources even if the primary one doesn't change
tests: Tests changing the dtmf PT mid-call
tests: Make sure dtmf events are really received
test: Test changing the dtmf_id
tests: dtmf method is not always auto
rtpsession: Only emit send-codec-changed message after the special codecs have been changed
rtpsession: Don't leak iterator on linking failure
rtpsession: Cleanup send codecbin on failure
rtpsession: Print error on session dispose problems
rtpdtmfsoundsource: Correctly check the presence of elements
rawudp: Use %d for ints, not %s
configure: quiet automake portability bs
msnstream: Make send sink async=false for now
msnstream: Don't keep lock into set_remote_candidates
tests: Test invalid property name in fs_element_added_notifier_from_keyfile
element-added-notifier: Don't crash on invalid property
rtpconference: Don't assert on non-existing sdes parts
rtpspecialsource: Dispose is not always called twice, cleanup in finalize
rtpsession: Remove useless ref
Version 0.0.15.1
Version 0.0.15
Require gst-p-bad 0.10.14 for mimic
tests: Unlock src before setting it to playing
tests: Refrain from using the thread unsafe version of failure in the nice test
rtpsession: Keep ref on stream while associating substreams to it
rtpsubstream: Remove another double-unlock in error case
rtpsession: Don't double-unlock
rtpsession: Fix leaking caps on signals after dispose
rtpsession: Fix potential leak if already disposed
rtpsubstrea: Remove unused variable
elementaddednotifier: Use g_connect_signal_object
Otherwise each element had a ref on the notifier and relied on the not thread
safe weak references.
rawudp: Emit local candidates if there are no local interfaces suitable for UPnP
rawudp: Add some UPnP debug messages
glib-gen: Use single = instead of == for portability
msnconnection: Check return values from recv()
msnsession: Conference must always set before get_property
msnsession: Only try to lock conference if it has been set
rtpsession: Initialise variable to NULL
Makes coverity happy
msnconnection: Remove unused variables
rtpstream: Correct documentation
rtpsession: Unref transmitter src/sink in dispose
Unref element from g_object_get(), fixes leak
elementaddednotifier: Unref element in iterator loop
Fixes leak
elementadded: Use gst_value_deserialize to read properties
Use the existing function instead of having our own less-capable re-implementation
Version 0.0.14.1
The Farsight project is an effort to create a framework to deal
with all known audio/video conferencing protocols. On one side it
offers a generic API that makes it possible to write plugins for
different streaming protocols, on the other side it offers an API
for clients to use those plugins.
The main target clients for Farsight are Instant Messaging
applications. These applications should be able to use Farsight
for all their Audio/Video conferencing needs without having to
worry about any of the lower level streaming and NAT traversal
issues.
Farsight forms an integral part of the Telepathy framework. It is
used by Empathy through the Telepathy-Farsight library. It can also
be easily used on embedded platforms by using Stream-Engine. The
Telepathy-Farsight library binds it to the Connection Managers via
D-Bus and the Telepathy Media Stream Spec and is used for all their
streaming requirements.
