summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2013-12-04record: Queue buffers and send out multiple frames at onceaudio-recordaudioBenjamin Franzke2-26/+152
Also restart sequence on discontiuity. FIXME: Research whether we can always just ask for the GST_BUFFER_DISCONT flag. Sending audio works quite reasonable with a server that has 72kbit/s now. :)
2013-12-04Timestamp playback buffers with appropriate flagsBenjamin Franzke1-7/+8
* GST_BUFFER_FLAG_DISCONT helps to play immediately after a pause (i.e. no talking) TODO: Not sure GST_BUFFER_FLAG_RESYNC does really help, but from the decoumentation it does what we want.
2013-12-04Port to gstreamer 1.0 APIBenjamin Franzke3-46/+88
Includes fixes: - use a callback, not a signal for appsink's new_buffer
2013-12-04Fix buffer timestampsBenjamin Franzke4-12/+72
* fix to start sequence at 0, if needed * do own buffer timestamping TODO: Sound is still choppy
2013-12-04audio: Improve pulseaudio config message a bitBenjamin Franzke1-3/+4
2013-12-04Add a FIXMEBenjamin Franzke1-0/+2
2013-12-04Set has_<key> for optional integer protobuf elementsBenjamin Franzke2-0/+4
2013-12-04Get OS name from /etc/os-releaseBenjamin Franzke3-1/+51
2013-12-04cmumble: Default to unix username as usernameBenjamin Franzke2-5/+5
2013-12-04Make use of CELT_GET_BITSTREAM_VERSIONBenjamin Franzke3-7/+12
Instead of hardcoding this.
2013-11-22message: Return true after receiving udptunnelBenjamin Franzke1-1/+1
2013-11-22message: Use message typedef for udptunnelBenjamin Franzke1-2/+2
2013-11-22message.c: Remove the send mutexBenjamin Franzke1-4/+0
Since we're now sending from the main thread only, this is no longer needed.
2013-11-22audio: Pull and send the recorded buffer from the main threadBenjamin Franzke3-1/+30
Tests showed this lets the appsink pull more buffers, since the streaming thread is not blocked due to network write.
2013-11-22connection: Do not unref the input source after creationBenjamin Franzke1-1/+0
It is done later in cmumble_connection_fini.
2013-11-22cmumble.c: Use message typedefs in callbacksBenjamin Franzke1-13/+11
2013-11-22Coding style fixesBenjamin Franzke5-9/+7
* remove trailing whitespace * inline where fit for 80 columns
2013-11-22audio.c: Move variable declaration before codeBenjamin Franzke1-3/+4
2013-11-22cmumble.c: Fix function name typoBenjamin Franzke1-1/+1
2013-11-22Check for udp type in tunnel messageBenjamin Franzke1-2/+14
Since we're currently support celt only, do nothing when retrieving other codecs.
2013-11-22Pass message type implicity in cmumble_send_*Benjamin Franzke5-54/+36
This also removes the unneeded "struct mumble_*" which is ugly in use.
2013-11-21message: Use g_printerr instead of printfBenjamin Franzke1-1/+1
2013-11-21io: Add another missing headerBenjamin Franzke1-0/+1
2013-11-21Create playback pipeline not for ourselfsBenjamin Franzke1-1/+5
FIXME: This should rather be created ondemand, and the problem is gone at all. To be done later.
2013-11-21recv_user_state: Check for availability of the session idBenjamin Franzke1-0/+6
2013-11-21ping: Use new cmumble message abstractionBenjamin Franzke1-5/+5
2013-11-21message.h: Add wrapper around protobufs ugly func namesBenjamin Franzke2-8/+46
They are of cause autogenerated. Only the one wrapper for the reject enum is added as one makro by hand.
2013-11-21message: Quit on G_TLS_ERROR_MISC errorBenjamin Franzke1-1/+2
Dont know why this happens sometimes, but quit at least. The szenario seems to be: - mute mic - wait around 10 - unmute mic -> G_TLS_ERROR_MSG on next write to socket
2013-11-21cmumble.c: Inline a makroBenjamin Franzke1-2/+1
Does not take 80 columns at all, no linebreak needed.
2013-11-21Add missing headersBenjamin Franzke2-0/+3
There were not warnings yet, since gstreamer 0.10 includes them implicitly. This will change with gstreamer 1.0.
2013-11-21mumble.proto: Rephrase the copy warningBenjamin Franzke1-1/+1
2013-11-19gen_message_list.sed: Remove trailing whitespace in outputBenjamin Franzke1-2/+2
2013-11-18Revert "recv_user_state: Verify needed, but optional values"Benjamin Franzke1-9/+0
This reverts commit f3b388b80daebb13398f2b6b269eda0aa947e469.
2013-11-18configure: Fix protoc-c detectionBenjamin Franzke1-3/+9
* Use --with-protoc=command instead of env var * Error out when protoc is not found.
2013-11-18recv_user_state: Verify needed, but optional valuesBenjamin Franzke1-0/+9
We need at least always the session id. Didn't looked through the mumble source code, when this event is useful without. Store a user only, if we've been given a name, id and channel.
2013-11-18Implement channel joinBenjamin Franzke2-1/+35
2013-11-18Add verbose optionBenjamin Franzke3-7/+22
Disable all debug output using that flag for now.
2013-11-18Implement reject messageBenjamin Franzke1-1/+21
2013-11-18Fix typo in cmumble struct nameBenjamin Franzke13-59/+59
2012-02-10Derive message type from position in .proto fileBenjamin Franzke4-48/+29
This involves changing the .proto file, to actually reflect the type in the protocol.
2012-02-10Use sizeof in encode_varint call for buffer sizeBenjamin Franzke1-1/+1
2012-02-10gen_message_list.sed: Add an exmapleBenjamin Franzke1-0/+2
This should help understanding that code easily.
2012-02-10s/ctx/cm/g; s/struct cmumble_context/struct cmumlbe/gBenjamin Franzke13-168/+168
2011-10-08audio: Another type fixBenjamin Franzke2-2/+2
2011-10-08audio.h: Fix typeBenjamin Franzke1-1/+1
2011-09-29Support delimiter in cmdline argumentsBenjamin Franzke1-4/+32
Delimiters are " and '.
2011-09-28Support sending TextMessagesBenjamin Franzke1-0/+31
2011-09-28Support receiving TextMessagesBenjamin Franzke1-1/+11
2011-09-28Store own user in cmumble_contextBenjamin Franzke2-0/+6
2011-09-28Store channel in cmumble_userBenjamin Franzke2-1/+2