Heya! http://0pointer.de/public/pulseaudio-0.9.15-test3.tar.gz We are approaching the final 0.9.15. Threre's probably going to be another iteration of test release before the final release though, given that my list of items to fix before the final release is, uh, .... oscillating. Hmm, in contrast to what I promised earlier I had to change a few strings. Also I added new strings because I earlier forgot to make some things translatable that need to be translatable. Hence, if you care about i18n, please use Transifex to update the translations. Colin Guthrie (2): Remove references to trademarked terms. Add the module dir to the libpulse pkgconfig file (needed for paprefs) Fabian Affolter (1): Updated German translation Francesco Tombolini (2): Italian language by ubuntu team Updated linguas file for it lang Iain Hibbert (1): Optionally disable IPv6 Kelemen G?bor (2): Make pulseaudio.desktop and org.pulseaudio.policy translatable. Make .desktop and .policy file in src/daemon translatable Lennart Poettering (86): the service fd is a stream socket, so handle things accordingly minor service IO fixes use the same service fd shutdown logic when destructing module and changing profile don't claim that profile changes are always successful don't try to recycle rtpoll objects add full set of argument description add a module that forwards cork/uncork requests to X11 as fake pause/resume key events fix aiff channel mapping for 6 channels we reinit proplist since the server will copy from client proplist anyway drop check for PA_PROP_APPLICATION_NAME since often enough we can deduce this better from g_get_application_name() make PA_GCC_PACKED and PA_GCC_MALLOC actually work add definition for GCC style weak references make PULSE_PROP env vars non-overriding but introduce PULSE_PROP_OVERRIDE for allowing overriding properly read icon/application name/display from gtk/glib/gdk ignore gtk-test show whether gtk+ support is enabled after configure document all currently known properties add doxygen comment for PA_GCC_WEAKREF unify ALSA mixer initialization don't open the alsa devices in hw:xxx mode bluetooth: fix message queue/rtpoll bump required alsa version allow pa to be run in a chroot() environment tht lacks /proc simplify pa_alsa_init_proplist_pcm() a bit and include resolution bits in alsa device props minor optimizations refer folks to the ALSA devs, not us allow importing of more memory blocks than exporting if we fail to import a memblock fill in silence to guarantee stability of timing export pa_match() introduce pa_realpath() export card information for sinks/sources and number of sinks/sources a profile would create to clients handle both positive and negative errno's additional validity check additional validity check make interpol-test useful for recording as well add pa_timespec_load modernize pa_msleep() a bit split out mixer setup into seperate functions to make things more readable don't try to use weakref stuff on older compilers that don't really support it make profile names translatable make it easier to debug timing related problems Make sure we actually call _() for translating profile names add test tool for debugging broken timing in sound drivers fix comment big alsa module rework to make things more robust to broken sound drivers and make it easier to debug them run make update-po simplify some code tell gcc to ignore invalid gtk header files split off foreign code into a convenience library to make gcc warnings go away check for ENABLE_LEGACY_RUNTIME_DIR with #ifdef, not #if add new property PA_PROP_APPLICATION_PROCESS_SESSION_ID and initialize it by default introduce default channel map in addition to the default sample spec print warnings about driver bugs at most once prefer profiles that match the default channel map fix wording, closes #484 make string translatable (fixes #483) rework logging to make it more modular make sure we don't choke on PULSE_PROP_OVERRIDE make sure we don't choke on overly long lines in .desktop files initialize selem index rework suspending/resuming prepare test for tracking down ens1371 issue Allow passing a NULL proplist to pa_xxxx_update_proplist() to just fire a notification a couple of boring updates add new function pa_alsa_get_driver_name_by_pcm() add new wrapper pa_alsa_safe_delay() around snd_pcm_delay() print driver name when we encounter driver bugs use pa_alsa_safe_delay() where appropriate allow scache entries to have arbitrary names set esound.byte_order property implement esd sample panning. closes #428 complete esd suspend/resume implementation it's probably more appropriate to return the configured latency instead of the actual latency fix handling of stereo update gitignore introduce new well-known role 'a11y' if ALSA gives us nonsensical data at least try to fix it up a little in case alsa lies to use don't spin forever implement device reservation scheme update reserve.c from upstream git don't put both the card and the pcm name in the description of a device if one contains the other set reserve interface application device name run make update-po ignore tags file make sure we check the sink status for PA_SINK_INPUT_FAIL_ON_SUSPEND only after module-suspend-on-idle had the chance to resume the device hide shave in gitignore Marc-Andr? Lureau (35): bluetooth: update SBC from upstream bluetooth: don't crash on pa_thread_mq_done() if pa_init() fail bluetooth: print SBC encoder implementation info pulse/context: add --enable-legacy-runtime-dir dbus-util: avoid double free hal-detect: make sure r is initialized, so we don't take random path alsa-util: make sure we check an initialized cn variable log: don't leak bt card-restore: it's not useful to check an array, let's check the length protocol-native: fix get_info() for cards paplay: check if pa_context_connect() succeed pactl: check if pa_context_connect succeed tests/sync-playback: check if pa_context_connect succeed tests/interpol-test: check if mainloop_start() succeed tests/thread-mainloop-test: check if threaded_mainloop_start() succeed pactl: return in case of error reading file (avoid using freed d) main: remove unused lf variable pulsecore: remove unused variable from cli_command_load() rtp-recv: remove unused variable assignment pacat: remove unused variable pulsecore: fix check for cb (m is already checked before) padsp: don't use si if it's NULL tests/ipacl-test: check inet_pton() pulsecore: unused variable e in hashmap_put() pulsecore: use r returned from fgets() daemon-conf: make sure c->log_level < LEVEL_MAX protocol-esound: don't accept a request of PROTOCOL_MAX stream-resotre: don't leak a name protocol-native: don't leak a proplist rtp: remove unused variable a pulsecore: don't leak d in case of error pulsecore: don't leak p when make_random_dir_and_link() alsa-util: check if mixer_poll_descriptors_count() < 0 bluetooth: fix suspend on a2dp (to do on HSP non SCO over PCM) build: shave it! Piotr Dr?g (1): Updated Polish translation Timo Jyrinki (1): Updated Finnish translation against current trunk. Xavier Conde (1): Updated catalan po Lennart -- Lennart Poettering Red Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4
