Original commit message from CVS:
* gst/ffmpegcolorspace/gstffmpegcodecmap.c:
(gst_ffmpegcsp_avpicture_fill):
* gst/ffmpegcolorspace/imgconvert.c: (avpicture_get_size),
(avpicture_alloc):
* gst/ffmpegcolorspace/imgconvert_template.h:
Use correct _fill function to get correct strides.
Original commit message from CVS:
* gst/playback/gstdecodebin.c: (gst_decode_bin_class_init),
(try_to_link_1), (get_our_ghost_pad), (remove_element_chain),
(unlinked), (no_more_pads), (close_link):
* gst/playback/gstplaybasebin.c: (gst_play_base_bin_init),
(unknown_type), (add_element_stream), (new_decoded_pad),
(removed_decoded_pad), (setup_source):
* gst/playback/gststreaminfo.c: (gst_stream_info_get_type),
(gst_stream_info_class_init), (gst_stream_info_init),
(gst_stream_info_new), (gst_stream_info_dispose),
(stream_info_mute_pad), (gst_stream_info_set_property),
(gst_stream_info_get_property):
* gst/playback/gststreaminfo.h:
Fix playback of multiple files.
a slightly different approach to handling dynamic pad removals.
This one only looks at pads that we have linked.
Original commit message from CVS:
* gst/playback/gstdecodebin.c: (gst_decode_bin_class_init),
(get_unconnected_element), (remove_starting_from), (pad_removed),
(close_link):
Implement support for dynamic pad changing. We listen to "live"
pad removals (i.e. while playing) and re-setup autoplugging
after that. Playbasebin/playbin need some more work for this
to finally work, but decodebin supports (and replugs) chained
ogg now.
Original commit message from CVS:
2004-10-21 Zaheer Abbas Merali <zaheerabbas at merali dot org>
* gst/tcp/gsttcpserversink.c:
(gst_tcpserversink_handle_server_read),
(gst_tcpserversink_init_send):
Zero some variables first (need for accept not to return EINVAL)
Original commit message from CVS:
* configure.ac: update for swfdec-0.3 and liboil-0.2
* ext/swfdec/gstswfdec.c: update for swfdec-0.3
* ext/swfdec/gstswfdec.h: same
* gst/videofilter/gstvideobalance.c: update for liboil-0.2
* gst/videotestsrc/videotestsrc.c: same
Original commit message from CVS:
* gst/tcp/gstmultifdsink.c: (gst_multifdsink_add),
(gst_multifdsink_remove), (gst_multifdsink_remove_client_link),
(is_sync_frame), (gst_multifdsink_new_client),
(gst_multifdsink_handle_client_write),
(gst_multifdsink_recover_client), (gst_multifdsink_queue_buffer),
(gst_multifdsink_handle_clients), (gst_multifdsink_change_state):
Turn warnings into info.
Don't allow a state change in the streaming thread.
Original commit message from CVS:
* gst/playback/gstplaybin.c: (gen_video_element),
(gen_vis_element), (remove_sinks), (setup_sinks):
Added vis plugin support, need to configure the vis
element to activate it.
Original commit message from CVS:
* gst/playback/gstplaybasebin.c: (setup_source):
Cleanup the previous pipeline a little earlier for the
case that a source element provides raw data.
Original commit message from CVS:
* gst/playback/gstplaybasebin.c: (setup_source),
(gst_play_base_bin_change_state):
Actually clean up streaminfo if output fails. This would trigger
if, for example, there was no CD in the drive. No preroll, so
a streaminfo structure is created, but the subsequent state change
of the thread fails.
* gst/playback/gstplaybin.c: (gst_play_bin_change_state):
Don't change state if parent failed.
Original commit message from CVS:
* gst/playback/gstplaybin.c: (gst_play_bin_class_init),
(gst_play_bin_init), (gst_play_bin_get_property), (handoff),
(gen_video_element), (remove_sinks):
Add small bits of code for screenshot handling.
Original commit message from CVS:
* gst/playback/gstplaybin.c: (gst_play_bin_set_property),
(gen_video_element), (gen_audio_element), (setup_sinks):
Don't assume the user provided sinks are named "sink"...
Original commit message from CVS:
* gst/playback/gstplaybasebin.c: (gen_preroll_element),
(unknown_type), (setup_source), (gst_play_base_bin_remove_element),
(gst_play_base_bin_link_stream):
Do not try to autoplug sources that generate raw streams like
cdparanoia.
disconnect the preroll overrun signal when we don't need it anymore.
Original commit message from CVS:
* gst/playback/gstplaybasebin.c: (play_base_bin_mute_pad),
(gst_play_base_bin_mute_stream), (gst_play_base_bin_link_stream):
* gst/playback/gstplaybin.c: (setup_sinks):
Implement muting/unmuting of streams, mute streams that are not
used.
Original commit message from CVS:
* gst/playback/gstdecodebin.c: (gst_decode_bin_init),
(find_compatibles), (close_pad_link), (try_to_link_1), (new_pad),
(no_more_pads), (close_link), (type_found):
* gst/playback/gstplaybasebin.c: (new_decoded_pad):
* gst/playback/gstplaybin.c: (gen_video_element):
Do not signal the no_more_pads after the first pad when
we are plugging a non dynamic element with multiple
output pads (like swfdec, dvdec, ...).
Original commit message from CVS:
* gst/playback/gstdecodebin.c: (gst_decode_bin_init),
(find_compatibles), (close_pad_link), (try_to_link_1),
(no_more_pads), (close_link), (type_found):
* gst/playback/gstplaybasebin.c: (new_decoded_pad):
* gst/playback/gstplaybin.c: (gen_video_element),
(gen_audio_element):
Set state on newly added element to READY so that negotiation
can happen ASAP.
Addes some more debug info.
Do not try to plug pads with multiple caps structures or ANY
because it is too dangerous since we do not do dynamic
replugging.
Original commit message from CVS:
* gst/playback/README:
* gst/playback/gstdecodebin.c: (gst_decode_bin_factory_filter),
(gst_decode_bin_init), (find_compatibles), (close_pad_link),
(try_to_link_1), (no_more_pads), (close_link), (type_found):
Add some debug info to decodebin, update README
Original commit message from CVS:
* ext/flac/gstflacdec.c: (gst_flacdec_src_query):
Only return true if we actually filled something in. Prevents
player applications from showing a random length for flac files.
* gst-libs/gst/riff/riff-read.c: (gst_riff_read_class_init),
(gst_riff_read_use_event), (gst_riff_read_handle_event),
(gst_riff_read_seek), (gst_riff_read_skip), (gst_riff_read_strh),
(gst_riff_read_strf_vids_with_data),
(gst_riff_read_strf_auds_with_data), (gst_riff_read_strf_iavs):
OK, ok, so I implemented event handling. Apparently it's normal
that we receive random events at random points without asking
for it.
* gst/avi/gstavidemux.c: (gst_avi_demux_reset),
(gst_avi_demux_src_convert), (gst_avi_demux_handle_src_query),
(gst_avi_demux_handle_src_event), (gst_avi_demux_stream_index),
(gst_avi_demux_sync), (gst_avi_demux_stream_scan),
(gst_avi_demux_massage_index), (gst_avi_demux_stream_header),
(gst_avi_demux_handle_seek), (gst_avi_demux_process_next_entry),
(gst_avi_demux_stream_data), (gst_avi_demux_loop):
* gst/avi/gstavidemux.h:
Implement non-lineair chunk handling and subchunk processing.
The first solves playback of AVI files where the audio and video
data of individual buffers that we read are not synchronized.
This should not happen according to the wonderful AVI specs, but
of course it does happen in reality. It is also a prerequisite for
the second. Subchunk processing allows us to cut chunks in small
pieces and process each of these pieces separately. This is
required because I've seen several AVI files with incredibly large
audio chunks, even some files with only one audio chunk for the
whole file. This allows for proper playback including seeking.
This patch is supposed to fix all AVI A/V sync issues.
* gst/flx/gstflxdec.c: (gst_flxdec_class_init),
(flx_decode_chunks), (flx_decode_color), (gst_flxdec_loop):
Work.
* gst/modplug/gstmodplug.cc:
Proper return value setting for the query() function.
* gst/playback/gstplaybasebin.c: (setup_source):
Being in non-playing state (after, e.g., EOS) is not necessarily
a bad thing. Allow for that. This fixes playback of short files.
They don't actually playback fully now, because the clock already
runs. This means that small files (<500kB) with a small length
(<2sec) will still not or barely play. Other files, such as mod
or flx, will work correctly, however.
Original commit message from CVS:
* gst/tcp/gstmultifdsink.c: (gst_multifdsink_add),
(gst_multifdsink_remove), (gst_multifdsink_remove_client_link),
(is_sync_frame), (gst_multifdsink_client_queue_buffer),
(gst_multifdsink_new_client),
(gst_multifdsink_handle_client_write),
(gst_multifdsink_recover_client), (gst_multifdsink_queue_buffer),
(gst_multifdsink_handle_clients):
* gst/tcp/gstmultifdsink.h:
Make syncing to keyframes actually work for new clients and lagging
clients.
Original commit message from CVS:
* gst/playback/gstplaybasebin.c: (new_decoded_pad):
Only signal the no_more_pads signal when we have
added the stream to our list.
Original commit message from CVS:
* gst/playback/gstplaybasebin.c: (remove_prerolls),
(new_decoded_pad):
* gst/playback/gstplaybasebin.h:
* gst/playback/gstplaybin.c: (setup_sinks):
Don't try to preroll or decode more than one audio/video
track.
Original commit message from CVS:
* gst/playback/gstplaybasebin.c: (gst_play_base_bin_change_state):
Throw error if we failed to find a suitable output. This should
throw an error if we successfully set up a pipeline (e.g. because
we recognized a media file) but found no decodable streams in it
(e.g. because it contains only media stream types for which we
have no decoders, or because it's not a media type).
Original commit message from CVS:
* ext/dirac/Makefile.am:
* ext/dirac/gstdirac.cc:
* ext/dirac/gstdiracdec.cc:
* ext/dirac/gstdiracdec.h:
Do something. Don't actually know if this works because I don't
have a demuxer yet.
* ext/gsm/gstgsmdec.c: (gst_gsmdec_getcaps):
Add channels=1 to caps returned from _getcaps().
* ext/ogg/gstogmparse.c: (gst_ogm_audio_parse_get_type),
(gst_ogm_video_parse_get_type), (gst_ogm_audio_parse_base_init),
(gst_ogm_video_parse_base_init), (gst_ogm_parse_init),
(gst_ogm_audio_parse_init), (gst_ogm_video_parse_init),
(gst_ogm_parse_sink_convert), (gst_ogm_parse_chain),
(gst_ogm_parse_change_state):
Separate between audio/video so ogmaudioparse actually uses the
audio pad templates. Both audio and video work now, including
autoplugging. Also use sometimes-srcpad hack.
* gst-libs/gst/riff/riff-read.c: (gst_riff_read_seek):
Handle events better. Don't hang on infinite loops.
* gst/avi/gstavidemux.c: (gst_avi_demux_class_init),
(gst_avi_demux_init), (gst_avi_demux_reset),
(gst_avi_demux_src_convert), (gst_avi_demux_handle_src_query),
(gst_avi_demux_stream_header), (gst_avi_demux_stream_data),
(gst_avi_demux_change_state):
* gst/avi/gstavidemux.h:
Improve A/V sync. Still not perfect.
* gst/matroska/ebml-read.c: (gst_ebml_read_seek),
(gst_ebml_read_skip):
Handle events better.
* gst/qtdemux/qtdemux.c: (gst_qtdemux_handle_sink_event),
(gst_qtdemux_loop_header), (qtdemux_parse_trak),
(qtdemux_audio_caps):
Add IMA4. Improve event handling. Save offset after a seek when
the headers are at the end of the file so that we don't end up in
an infinite loop.
* gst/typefind/gsttypefindfunctions.c: (qt_type_find):
Add low-priority typefind support for files with no length.
Original commit message from CVS:
* ext/gdk_pixbuf/pixbufscale.c: (gst_pixbufscale_getcaps):
Correct caps negotiation
* gst/volume/gstvolume.c: (volume_chain_float),
(volume_chain_int16):
Modify debug output to be little more informative
* sys/ximage/ximagesink.c: (gst_ximagesink_check_xshm_calls):
* sys/xvimage/xvimagesink.c: (gst_xvimagesink_check_xshm_calls),
(gst_xvimagesink_xvimage_destroy):
Add XSync calls after detaching from the shared memory segment to
avoid a crash.
Original commit message from CVS:
* gst/asfdemux/gstasfdemux.c: (_read_var_length), (_read_guid),
(gst_asf_demux_process_segment), (gst_asf_demux_handle_data),
(gst_asf_demux_process_chunk), (gst_asf_demux_handle_sink_event):
Prevent infinite loops. More correct error reporting.
* gst/auparse/gstauparse.c: (gst_auparse_chain):
Error out if negotiation fails.
* gst/playback/gstplaybasebin.c: (setup_source),
(gst_play_base_bin_change_state), (gst_play_base_bin_error),
(gst_play_base_bin_found_tag):
Error/tag forwarding. Pre-roll fixes for source errors on state
changes (e.g. "file does not exist") to prevent hangs.
Original commit message from CVS:
* ext/mad/gstmad.c: (gst_mad_check_caps_reset),
(gst_mad_change_state):
Allow for mp3 rate/channels changes. However, only very
conservatively. Reason that we *have* to enable this is smiply
because the mad find_sync() function is not good enough, it will
regularly sync on random data as valid frames and therefore make
us provide random caps as *final* caps of the stream. The best fix
I could think of is to simply require several of the same stream
changes in a row before we change caps.
The actual testcase that works now is #
* ext/ogg/Makefile.am:
* ext/ogg/gstogg.c: (plugin_init):
* ext/ogg/gstogmparse.c:
OGM support (video only for now; I need an audio sample file).
* gst/asfdemux/gstasfdemux.c: (gst_asf_demux_base_init),
(gst_asf_demux_process_stream), (gst_asf_demux_video_caps),
(gst_asf_demux_add_video_stream):
WMV extradata.
* gst/playback/gstplaybasebin.c: (unknown_type):
Don't error out on single unknown-types after all. It's wrong.
If we found type of video and audio but not of a subtitle stream,
it will still error out (which is unwanted). Will find a better fix
later on.
* gst/typefind/gsttypefindfunctions.c: (ogmvideo_type_find),
(ogmaudio_type_find), (plugin_init):
OGM support.
Original commit message from CVS:
* gst/tcp/gstfdset.c: (gst_fdset_fd_has_closed),
(gst_fdset_fd_has_error), (gst_fdset_fd_can_read),
(gst_fdset_fd_can_write), (gst_fdset_wait):
* gst/tcp/gstmultifdsink.c: (gst_client_status_get_type),
(gst_multifdsink_init), (gst_multifdsink_add),
(gst_multifdsink_remove), (gst_multifdsink_get_stats),
(gst_multifdsink_remove_client_link),
(gst_multifdsink_client_queue_buffer),
(gst_multifdsink_handle_client_write),
(gst_multifdsink_recover_client), (gst_multifdsink_handle_clients),
(gst_multifdsink_close), (gst_multifdsink_change_state):
* gst/tcp/gstmultifdsink.h:
* gst/tcp/gsttcpserversink.c: (gst_tcpserversink_class_init),
(gst_tcpserversink_removed):
Small cleanups in fdset.c
Use a hastable to map fd to the client structure for faster
lookup in _remove and get_stats.
Added virtual function to close the fds.
Handle clients even when the select/poll call was unblocked because
of a command.
Implement syncing to keyframe in the recovery procedure.
Original commit message from CVS:
* configure.ac: remove NASM check, since we don't use it. Update
dirac check to 0.4
* ext/dirac/gstdiracdec.cc: update to current 0.4 API
* gst/audioconvert/gstaudioconvert.c: (gst_audio_convert_link):
Initialized variables.
* gst/qtdemux/qtdemux.c: (gst_qtdemux_change_state),
(gst_qtdemux_loop_header), (qtdemux_parse), (qtdemux_parse_trak),
(gst_qtdemux_handle_esds), (qtdemux_audio_caps): Fix seeking, add
SVQ3 format
Original commit message from CVS:
* gst/tcp/gstmultifdsink.c: (gst_multifdsink_remove_client_link),
(gst_multifdsink_handle_clients), (gst_multifdsink_change_state):
Don't close the fd in multifdsink as we didn't open it in the
first place. Some cleanups.
Original commit message from CVS:
* gst/playback/gstplaybasebin.c: (state_change), (setup_source),
(gst_play_base_bin_change_state):
Handle the case where we failed to setup a clear pipeline. This
will throw an error (or EOS, another nice case) and if you don't
catch that, the app will wait for the signal forever (and thus
hang).
Original commit message from CVS:
* ext/gnomevfs/gstgnomevfssink.c:
(gst_gnomevfssink_uri_get_protocols):
* ext/gnomevfs/gstgnomevfssrc.c:
(gst_gnomevfssrc_uri_get_protocols):
* ext/gnomevfs/gstgnomevfsuri.c: (gst_gnomevfs_get_supported_uris):
* ext/gnomevfs/gstgnomevfsuri.h:
Use _uri_new() instead of _open(), so it doesn't take as long and
Christophe's computer won't hang.
* gst/playback/gstplaybasebin.c: (unknown_type):
Throw error on unknown media type, so apps actually display it.
Original commit message from CVS:
* gst/playback/gstplaybasebin.c: (queue_overrun), (no_more_pads),
(setup_source), (gst_play_base_bin_set_property),
(gst_play_base_bin_add_element):
* gst/playback/gstplaybin.c: (gst_play_bin_send_event):
Some more work on making sure seeking pauses the pipeline and
that changing the uri actually does something.
Original commit message from CVS:
* gst/tcp/gstfdset.c: (gst_fdset_wait):
* gst/tcp/gstmultifdsink.c: (gst_multifdsink_close):
* gst/tcp/gsttcpserversink.c: (gst_tcpserversink_init_send),
(gst_tcpserversink_close):
Be a bit more paranoid when freeing memory.
Original commit message from CVS:
* gst/playback/gstplaybasebin.c:
(gst_play_base_bin_dispose), (gst_play_base_bin_set_property):
Handle double disposals, and proper change of URIs.
Original commit message from CVS:
* ext/alsa/gstalsamixer.c: (gst_alsa_mixer_update),
(gst_alsa_mixer_get_volume), (gst_alsa_mixer_set_volume),
(gst_alsa_mixer_set_mute), (gst_alsa_mixer_set_record),
(gst_alsa_mixer_set_option), (gst_alsa_mixer_get_option):
Update mixer (to sync with other sessions) if we try to obtain
a new value. This makes alsamixer work accross applications.
* ext/alsa/gstalsasink.c: (gst_alsa_sink_get_time):
Only call sync functions if we're running, else alsalib asserts.
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_src_query):
Sometimes fails to compile. Possibly a gcc bug.
* gst/playback/gstplaybin.c: (gen_video_element),
(gen_audio_element):
Add a reference to an application-provided object, because we lose
this same reference if we add it to the bin. If we don't do this,
we can only use this object once and thus crash if we go from
ready to playing, back to ready and back to playing again.
Also add an audioscale element because several cheap soundcards -
like mine - don't support all samplerates.
* sys/ximage/ximagesink.c: (gst_ximagesink_xcontext_get),
(gst_ximagesink_xcontext_clear), (gst_ximagesink_change_state):
Fix wrong order or PAR calls. Makes automatically obtained PAR
from the X server atually being used.
Original commit message from CVS:
* gst/tcp/gstfdset.c: (gst_fdset_free), (gst_fdset_set_mode),
(gst_fdset_get_mode), (gst_fdset_add_fd), (gst_fdset_remove_fd),
(gst_fdset_fd_ctl_write), (gst_fdset_fd_ctl_read),
(gst_fdset_fd_has_closed), (gst_fdset_fd_has_error),
(gst_fdset_fd_can_read), (gst_fdset_fd_can_write),
(gst_fdset_wait):
* gst/tcp/gstfdset.h:
* gst/tcp/gstmultifdsink.c: (gst_multifdsink_add),
(gst_multifdsink_client_queue_buffer),
(gst_multifdsink_handle_client_write):
* gst/tcp/gstmultifdsink.h:
Some extra checks in gstfdset.
Only use send() when the fd is a socket. Don't try to
read from write only fds.
Original commit message from CVS:
* gst/tcp/gstfdset.c: (ensure_size), (gst_fdset_wait):
Realloc test fdset in the lock and right before starting
the poll call. Bump the limit to 4096.
Original commit message from CVS:
2004-08-17 Zaheer Abbas Merali <zaheerabbas at merali dot org>
* gst/audioscale/gstaudioscale.c:
* gst/audioscale/gstaudioscale.h:
made audioscale resample from any sample rate to any sample rate
Original commit message from CVS:
* gst/tcp/gstmultifdsink.c: (gst_multifdsink_class_init),
(gst_multifdsink_init), (gst_multifdsink_add),
(gst_multifdsink_client_queue_buffer),
(gst_multifdsink_set_property), (gst_multifdsink_get_property):
* gst/tcp/gstmultifdsink.h:
Added option to send a keyframe to clients as the first buffer.
Make timeout property writable.
Original commit message from CVS:
* gst/tcp/gstfdset.c: (ensure_size), (gst_fdset_new),
(gst_fdset_add_fd), (gst_fdset_remove_fd),
(gst_fdset_fd_has_closed), (gst_fdset_fd_has_error),
(gst_fdset_fd_can_read), (gst_fdset_fd_can_write),
(gst_fdset_wait):
Make sure the pollfds are not changed when the poll call is
running. Protect against array out of bounds.
Original commit message from CVS:
* gst/tcp/gstmultifdsink.c: (gst_unit_type_get_type),
(gst_client_status_get_type), (gst_multifdsink_class_init),
(gst_multifdsink_init), (gst_multifdsink_remove_client_link),
(gst_multifdsink_handle_client_read),
(gst_multifdsink_handle_client_write),
(gst_multifdsink_recover_client), (gst_multifdsink_queue_buffer),
(gst_multifdsink_handle_clients), (gst_multifdsink_set_property),
(gst_multifdsink_get_property):
* gst/tcp/gstmultifdsink.h:
* gst/tcp/gsttcp-marshal.list:
Starting to prepare for specifying buffer time in other units
than buffers. Expose remove reason in signal.
Original commit message from CVS:
* gst/tcp/gstmultifdsink.c: (gst_multifdsink_add),
(gst_multifdsink_remove), (gst_multifdsink_clear),
(gst_multifdsink_remove_client_link),
(gst_multifdsink_handle_client_read),
(gst_multifdsink_client_queue_data),
(gst_multifdsink_client_queue_buffer),
(gst_multifdsink_handle_client_write),
(gst_multifdsink_queue_buffer), (gst_multifdsink_handle_clients),
(gst_multifdsink_chain), (gst_multifdsink_close):
* gst/tcp/gstmultifdsink.h:
Added more debugging info. Changed the way clients are
removed from the lists. Fixed a bug where a bad file descriptor
could cause many clients to be removed.
Original commit message from CVS:
* gst/tcp/gstmultifdsink.c: (gst_multifdsink_class_init),
(gst_multifdsink_add), (gst_multifdsink_get_stats),
(gst_multifdsink_client_remove),
(gst_multifdsink_handle_client_read),
(gst_multifdsink_handle_client_write),
(gst_multifdsink_queue_buffer), (gst_multifdsink_handle_clients):
Do a bit more logging, make the client_read code more robust.
Original commit message from CVS:
* gst/tcp/gstmultifdsink.c: (gst_multifdsink_class_init),
(gst_multifdsink_add), (gst_multifdsink_get_stats),
(gst_multifdsink_client_remove),
(gst_multifdsink_handle_client_read),
(gst_multifdsink_handle_client_write),
(gst_multifdsink_queue_buffer), (gst_multifdsink_handle_clients):
Make sure we don't try to read more from a client that what
ioctl says us or we deadlock.
Original commit message from CVS:
* gst/videotestsrc/gstvideotestsrc.c:
(gst_videotestsrc_get_capslist), (generate_capslist),
(plugin_init):
generate the list of supported caps at startup and reuse it instead
of always generating it
Original commit message from CVS:
* gst/audioscale/gstaudioscale.c:
- fix templates to only support S16, it's the only format that works
- make caps nego code use try_set_caps_nonfixed and fixation instead
of try_set_caps twice, which is not nice for autopluggers
- change rank to secondary, so autopluggers can pick it up after
audioconvert
Original commit message from CVS:
* gst/tcp/gstmultifdsink.c: (gst_multifdsink_class_init),
(gst_multifdsink_add), (gst_multifdsink_remove),
(gst_multifdsink_clear), (gst_multifdsink_get_stats),
(gst_multifdsink_client_remove),
(gst_multifdsink_handle_client_write),
(gst_multifdsink_queue_buffer), (gst_multifdsink_handle_clients):
* gst/tcp/gstmultifdsink.h:
Recover from a select with a bad file descriptor by removing
the client.
