[pulseaudio-discuss] [ANNOUNCE] PulseAudio 0.9.15-test6

Lennart Poettering lennart at poettering.net
Tue Mar 31 17:38:04 PDT 2009


A new test release for you:

http://0pointer.de/public/pulseaudio-0.9.15-test6.tar.gz

Without further ado here's the shortlog:

A S Alam (1):
      Adding Punjabi Language

Colin Guthrie (1):
      raop: Add call to pa_sink_set_max_request()

Finn Thain (1):
      revive solaris module

Kyle Cronan (1):
      Specifying ALSA mixer control

Lennart Poettering (89):
      Ignore HAL NoSuchProperty errors when looking for capabilities field. Closes rhbz #489394
      deduce a proper icon for TV sets
      get rid of setsockopt() calls since they have never been implemented upstream
      document todo items discussed at bt meeting
      fix prototypes of sbc functions a bit
      unfuck i18n
      document things that need to be fixed with FIXME
      fix a comment
      introduce pa_assert_fp() for fast path assertions
      introduce pa_sink_input_get_relative_volume()
      always remap relative volume properly
      simply bluetooth nrec handling a bit
      fix channel mapping for a52 devices. Closes #508
      don't try to unref reserve stuff if we don't use it
      fix prototypes of remaining sbc.[ch] API regarding size_t/const
      rework bluetooth IO loops
      fix a misplaced assert
      make sure we dispatch messages in order
      add pa_hook_is_firing
      rework device discovery to share a single device list among all modules
      at a couple of #ifdef NOKIAs for now
      add functions that modules can call whenever they now the volume changed
      readd volume control logic
      enable bluetooth default by support
      only store card profile if flagged for that
      add pa_assert_cc() for compile time assertions
      don't fail when no session bus is available
      only decrease timer slack, never increase
      simplify latency range by not allowing stored 'wildcard' ranges anymore
      set request/rewind sizes only via accessor functions
      minimal reordering
      Allow calling pa_{sink|source}_update_proplist() without an actual proprlist
      fix bad memory access when destroying m-b-d
      get rid of 'default' min/max latencies, simplify things by just having absolute boundaries
      make pa_source_set_max_rewind() work similar to pa_sink_set_max_rewind()
      introduce new flag that marks sinks/sources which can adjust the latency dynamically
      simplify things and make sure timing setters can be called in most contexts
      link jack modules's max_request to the jack buffer size
      set latency range only in tsched mode
      don't fiddle with latency range because we cannot adjust it dynamically
      Don't initialize userdata twice
      again, don't fiddle with latency range in sinks with static latency
      initialize max_request
      don't fiddle with latency range in sources with static latency
      initialize max_request to SO_SNDBUF
      beef up esd sink properties a bit
      trivial simplification
      add simple ref counting debugging framework
      don't show full so path in backtrace
      optionally skip initial frames in backtrace
      fix misplace _ref() calls that should have been _assert_ref()
      fix value of DYNAMIC_LATENCY
      use u->use_tsched everywhere
      copy latency flags from sink to monitor source
      add missing whitespace
      on monitor source be fine with any latency range set by the sink
      make sure the discovery module is only loaded once
      add missing initialization
      fix typo
      Small fix for MacOSX compat
      don't enabled tsched on software ALSA devices
      Trigger move callback a little bit earlier so that no IO thread is running
      simplify latency config functions a bit and make them callable in more contexts
      document more often the context certain functions are called in
      add pa_memblockq_apply_attr()/pa_memblockq_get_attr()
      typo fix
      notify clients about tlength changes
      rework tunneling code
      Use pa_source_set_max_rewind_within_thread() for updating the monitor source's max_rewind
      introduce typedef for pa_bt_audio_state and use it everywhere
      get rid of old 'Connected' property parsing and make sure we don't execute two case branches
      make sure we always read in all properties
      make sure we never access an invalid pa_bluetooth_device object
      fix compiler warning
      take org.pulseaudio.Server instead of org.pulseaudio to match the interface name already used in the HAL module
      minor cleanups
      allow nofail mode only when no server string was specified
      Revert "make sure we always read in all properties"
      explain ff7033c11d9248fe837204b03c8397231dc511fe
      fix an error where a signal was accidently freed when it is tried to register it twice
      update command name table
      handle buffer_attr changed messages properly
      Log underruns
      be a bit more verbose about max_request changes
      fix buffer defaults
      add buffer_attr callback stuff to exported symbol list
      in verbose mode log buffer attr changes
      increase timing update interval exponentially
      prepare test6

Luiz Augusto von Dentz (5):
      Maintain the original code style for sbc.
      Fix misuse of 'frame.joint' when estimating the frame length.
      Update ipc to match new message headers introduced on BlueZ 4.34.
      Query and make use of the current configuration.
      Do not reconfigure capabilities.

Maarten Bosmans (2):
      handle failure to parse proplist in cli-command gracefully
      fix some typos in doxygen comments

Marc-André Lureau (21):
      bluetooth: reset read/write index when starting to stream
      bluetooth: s/handled/not_yet_handled for signals
       bluetooth: accept temporarily unavailable error
      bluetooth: restart timer when write begin
      bluetooth: stream also when source is suspended
      bluetooth: load bluetooth device with connected profile
      bluetooth: update SCO over PCM with latest changes
      bluetooth: mark info_valid when receive Connected
      bluetooth: rework the info_valid logic to be more tolerant
      pulse: check context (do not user pstream when NULL)
      bluetooth: fix #NOKIA, correctly unlink sink/src
      bluetooth: connected can be -1, check > 0
      bluetooth: fail when switching on non-connected profile
      bluetooth: don't access outside array range
      bluetooth: remove racy GetProperties to check profile
      bluetooth: GetProperties after profile UUID show up
      bluetooth: use new audio State properties
      dbus: split dbus-util into dbus-shared
      daemon: take org.pulseaudio
      pulse: client connect to dbus
      gtk-test: updated to use PA_CONTEXT_NOAUTOSPAWN|PA_CONTEXT_NOFAIL

Piotr Drąg (1):
      Updated Polish translation

Xavier Conde (1):
      Update catalan po

beckerde (3):
      Sending translation for Spanish
      Sending translation for Spanish
      Sending translation for Spanish

kmilos (2):
      Sending translation for po/sr.po
      Sending translation for po/sr at latin.po

Lennart

-- 
Lennart Poettering                        Red Hat, Inc.
lennart [at] poettering [dot] net         ICQ# 11060553
http://0pointer.net/lennart/           GnuPG 0x1A015CC4



More information about the pulseaudio-discuss mailing list