Implement keyframe seeking in oggdemux by doing the double seek trick. First
seek to the required position, then read pages for all streams to grab the
granulepos (to know the timing of the keyframe) of each stream, then seek back
to the first keyframe.
After a seek, discard all packets before the packet with the granulepos on it so
that the output buffers contain valid timestamps.
Reorder some code so that we check the timestamps before allocating and pushing
an output buffer.
Do more checks on valid packets in ogm mode.
OGM demuxing no longer requires helper elements. It's done internally
in oggdemux. Vorbis comments are still not handled because I don't
have anything to test with.
gstoggstream.c:419: error: format ‘%lld’ expects type ‘long long int’, but argument 8 has type ‘gint64’
gstoggdemux.c:2253: error: format ‘%lld’ expects type ‘long long int’, but argument 8 has type ‘GstClockTime’
gstoggdemux.c:2333: error: format ‘%lld’ expects type ‘long long int’, but argument 8 has type ‘GstClockTime’
Timestamp generation was broken by the last commit for formats
with a non-zero granule shift. Also keep track of the last keyframe
so that we can regenerate granulepos for theora.
Add a granule to granulepos conversion function. Fix the duration
function for vorbis. Handle timestamps on header packets differently
and be more careful about calculating OFFSET and OFFSET_END. After
this change, timestamps for vorbis don't exactly match up with the
timestamps that vorbisparse outputs, but it's unclear if vorbisparse
is actually correct and it would add a lot more code to make oggdemux
match vorbisparse. Fixes#602790.
Adds code that parses headers of various formats encapsulated in
Ogg in order to calculate timestamps and durations of each buffer.
Removes the creation of helper decoder elements to do this calculation
via conversion queries.
Fixes: #344013, #568014.
Revert previous 'fix' for bug #588717 and fix it properly, whilst
maintaining the streamheader field on the output caps. Also make
sure we don't leak header buffers we couldn't push when downstream
is unlinked. Add unit test for the presence of the streamheader
field on the output caps and for the issue from bug #588717.
Fix caps warning when there's no element linked downstream, and pass
not-linked flow return value correctly up the chain, so we error out
correctly. Fixes#588717.
For this to work properly, theoradec and vorbisdec need to put
tag events received from upstream into the pending_events list
so they get pushed out after any newsegment event, not before.
When we feed the ogg sync layer, we need to feed it contiguous data even if the
sync layer did not consume all of it yet. This makes sure that it always finds
the next page even for more corrupted files. Use a different read_offset for
this purpose. since we now keep track of the sync layer, we don't have to reset
after finding a start of a page.
Add some more debug info for the error paths.
Only reset the sync layer when we perform a seek operation.
Avoid failure when the next chain has no bos pages but instead simply ignore it.
when we receive unknown page serial numbers mid stream, don't fail but post a
warning and hope that we get back on track later.
Fixes#579642
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_class_init),
(gst_ogg_pad_dispose), (gst_ogg_pad_finalize):
Use G_DEFINE_TYPE for the OggPad to get some threadsafe type
init from glib.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_sink_event):
If no stream was found before receiving EOS, post an error message.
Fixes#561924.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_submit_packet),
(gst_ogg_demux_deactivate_current_chain),
(gst_ogg_demux_perform_seek), (gst_ogg_demux_handle_page),
(gst_ogg_demux_loop):
* ext/ogg/gstoggdemux.h:
Copy seqnums around to track playback segments and messages.
Original commit message from CVS:
Patch by: Daniel Drake <dsd at laptop dot org>
* ext/ogg/gstoggmux.c: (gst_ogg_mux_clear_collectpads):
Unref all buffers when clearing collectpads. Fixes bug #546955.
Original commit message from CVS:
Patch by: Alessandro Dessina <alessandro nnva org>
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_deactivate_current_chain),
(gst_ogg_demux_activate_chain):
Don't add pads and activate them for skeleton streams. These are already
handled inside oggdemux. Fixes bug #537599.
Original commit message from CVS:
* ext/ogg/gstoggmux.c: (gst_ogg_mux_sink_event),
(gst_ogg_mux_request_new_pad):
* ext/ogg/gstoggmux.h:
Don't pretend to support NEWSEGMENT events, instead override the
GstCollectPads event function to return FALSE on NEWSEGMENT events
and do the normal work for other events.
This prevents elements like flacenc to seek to the start and rewrite
some data which then results in a broken Ogg packet.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_chain_peer),
(gst_ogg_demux_perform_seek), (gst_ogg_demux_handle_page):
Don't use GST_CLOCK_TIME_NONE as start of NEWSEGMENT events.
This fixes a critical warning.
Original commit message from CVS:
Patch by: Jan Gerber <j at oil21 dot org>
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_parse_skeleton_fisbone):
Fix calculation of the start time from skeleton streams.
Fixes bug #530068.
Original commit message from CVS:
Patch by: j^ <j at oil21 dot org>
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_parse_skeleton_fishead),
(gst_ogg_pad_parse_skeleton_fisbone):
* ext/ogg/gstoggdemux.h:
Parse presentation time from skeleton streams and use it as offset
for the timestamps. Fixes bug #530068.
Original commit message from CVS:
* ext/ogg/gstoggmux.c:
Update Ogg/Dirac muxing. Removes the weird "KW-DIRAC" bos
packet. Should conform to what we currently think is the
final Ogg/Dirac muxing spec.
Original commit message from CVS:
* ext/ogg/gstogmparse.c: (gst_ogm_parse_stream_header):
GST_TYPE_FRACTION contains gints so correctly cast gint64 arguments to
vaargs functions to gint. Otherwise the fractions will get 0 set
instead of the correct value on big endian systems. Fixes bug #529018.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_reset),
(gst_ogg_demux_chain_elem_pad):
Add small debug line.
Pass return code from the internal decoder instead of the too generic
GST_FLOW_ERROR.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_activate_chain),
(gst_ogg_demux_read_chain):
Refix oggdemux, we only have a problem if we failed to find a chain and
we are not EOF.
Original commit message from CVS:
Patch by: Victor STINNER <victor dot stinner at haypocalc dot com>
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_activate_chain),
(gst_ogg_demux_read_chain):
When we fail to find a BOS page and we and up with no chain, error out
properly instead of segfaulting. Fixes#525665.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_activate_chain),
(gst_ogg_demux_read_chain), (gst_ogg_demux_handle_page):
The new-pad-group sequence is add-pads, no-more-pads, add-pads,
no-more-pads...
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_bisect_forward_serialno),
(gst_ogg_demux_read_chain):
If we find a new serial number but it does not contain a BOS page, make
sure we initialize the chain to NULL because else we will try to scan it
and crash. Fixes#500763
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_read_chain),
(gst_ogg_demux_collect_chain_info):
Also submit the eos page when trying to find the first timestamp.
See #466717.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_find_chains):
The chain should be freed if we error out here, else it will leak.
* gst/playback/gstdecodebin.c: (disconnect_unlinked_signals),
(cleanup_decodebin):
Don't forget to *properly* remove the signals, else it will leak.
Original commit message from CVS:
* tests/check/pipelines/oggmux.c: (validate_ogg_page), (is_video),
(eos_buffer_probe):
Add a test that ensures we set DELTA_UNIT on all non-header,
non-video buffers, if we have a video stream.
* ext/ogg/gstoggmux.c: (gst_ogg_mux_queue_pads),
(gst_ogg_mux_process_best_pad):
Move setting delta_pad to earlier, where we inspect all pads, so
that leading audio pages don't get DELTA_UNIT unset if they come
before the first DELTA_UNIT from video pages. Fixes the newly-added
test. Fixes#385527.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_submit_packet),
(gst_ogg_demux_perform_seek), (gst_ogg_demux_read_chain):
Fix compilation on mingw. Fixes#446972.
Original commit message from CVS:
* ext/ogg/gstoggmux.c: (gst_ogg_mux_send_headers):
The BOS page of the first Dirac video stream needs to come before
the BOS page of any Vorbis streams or other audio streams, just like
it is with Theora.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_parse_skeleton_fisbone),
(gst_ogg_pad_submit_packet), (gst_ogg_demux_perform_seek),
(gst_ogg_demux_read_chain), (gst_ogg_demux_collect_chain_info):
consideratly speedup ogg chain detection by not trying to find a base
timestamp for skeleton streams.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_submit_packet),
(gst_ogg_demux_perform_seek), (gst_ogg_demux_handle_page):
Some more chained streaming ogg timestamp fixes.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_submit_packet),
(gst_ogg_demux_activate_chain), (gst_ogg_demux_perform_seek),
(gst_ogg_demux_handle_page):
Add some FIXMEs.
Fix chain start/stop segment handling based on patch by
<ahalda at cs dot mcgill dot ca> see #320984.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_submit_packet),
(gst_ogg_demux_collect_chain_info):
If there is a stream in a chain without any data packets, ignore the
stream in the total length calculations. Might be related to #436820.
Original commit message from CVS:
reviewed by: Wim Taymans <wim@fluendo.com>
* ext/ogg/gstoggmux.c: (gst_ogg_mux_ogg_pad_destroy_notify),
(gst_ogg_mux_request_new_pad), (gst_ogg_mux_release_pad):
Use newly added GstCollectPads API to free the allocated resources in
the GstOggPad structures (#402393).
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_submit_buffer),
(gst_ogg_demux_get_data), (gst_ogg_demux_get_next_page),
(gst_ogg_demux_get_prev_page), (gst_ogg_demux_do_seek),
(gst_ogg_demux_perform_seek),
(gst_ogg_demux_bisect_forward_serialno),
(gst_ogg_demux_read_chain), (gst_ogg_demux_read_end_chain),
(gst_ogg_demux_find_chains), (gst_ogg_demux_handle_page),
(gst_ogg_demux_chain), (gst_ogg_demux_combine_flows),
(gst_ogg_demux_loop_reverse), (gst_ogg_demux_loop):
* ext/ogg/gstoggdemux.h:
Properly propagate streaming errors when we are scanning the file for
chains so that we don't crash when shut down. Might fix some crashers
when quickly switching oggs in RB such as #332503 and #378436.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_read_chain):
Error out properly if we get an error from libogg while reading the
BOS page(s). Fixes crash parsing 'fuzzed' ogg file (#399340).
Original commit message from CVS:
* ext/ogg/gstoggmux.c: (all_pads_eos), (gst_ogg_mux_collected):
Fix the cases where oggmux doesn't properly figure out that all
sinkpads have gone EOS, and therefore doesn't push out the remaining
buffers and the final EOS event.
Fixes#363379
Original commit message from CVS:
* ext/ogg/Makefile.am:
Dist gstoggdemux.h to fix 'make distcheck'.
* sys/v4l/Makefile.am:
Fix 'make distcheck' even more.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_stream_out):
* ext/vorbis/vorbisdec.c: (vorbis_handle_data_packet):
* gst/playback/gstdecodebin2.c:
(gst_decode_group_check_if_blocked):
Printf format and missing argument fixes.
Original commit message from CVS:
* ext/ogg/gstogmparse.c: (gst_ogm_parse_stream_header),
(gst_ogm_parse_change_state):
Activate pads before adding them to the element.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_page_copy), (gst_ogg_page_free),
(gst_ogg_pad_init), (gst_ogg_pad_dispose), (gst_ogg_pad_reset),
(gst_ogg_pad_stream_out), (gst_ogg_pad_submit_page),
(gst_ogg_chain_reset), (gst_ogg_demux_perform_seek):
Some cleanups.
Handle continued pages in reverse mode.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_submit_page),
(gst_ogg_demux_chain):
Don't just ignore return values from _pad_push().
Small debug improvements.
Original commit message from CVS:
* ext/ogg/gstoggmux.c: (gst_ogg_mux_process_best_pad):
If our incoming buffer is marked as DISCONT, then increment the page
number (so that the discontinuity is marked in the final ogg
bitstream) and flush the previous page.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_submit_page):
Improve a debug line slightly.
* ext/ogg/gstogmparse.c: (gst_ogm_parse_plugin_init):
Call gst_riff_init() in plugin_init, to avoid getting errors from
the debug system (unrelated changes to another plugin made this turn
up; not sure why).
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet dot be>
* ext/ogg/gstoggmux.c: (gst_ogg_mux_release_pad):
Remove the pad from the element in release_pad.
Original commit message from CVS:
* ext/ogg/gstoggparse.c: (gst_ogg_parse_dispose),
(gst_ogg_parse_append_header), (gst_ogg_parse_chain):
Set caps on pushed buffers; fix up refcounting of caps objects.
Original commit message from CVS:
* ext/ogg/gstoggmux.c: (gst_ogg_mux_request_new_pad),
(gst_ogg_mux_release_pad), (gst_ogg_mux_push_buffer),
(gst_ogg_mux_compare_pads), (gst_ogg_mux_queue_pads),
(gst_ogg_mux_send_headers), (gst_ogg_mux_process_best_pad),
(gst_ogg_mux_collected):
Commit patch from James "Doc" Livingston, adds proper EOS handling
in oggmux. GStreamer can, for the first time ever, create a valid
Ogg file! Yay!
* tests/check/pipelines/oggmux.c: (check_chain_final_state),
(oggmux_suite):
Reenable tests now that they pass.
Original commit message from CVS:
* ext/ogg/gstoggmux.c: (gst_ogg_mux_push_buffer):
Timestamps are unsigned; comparision against GST_CLOCK_TIME_NONE was
always true, leading to dropping all timestamps.
Original commit message from CVS:
* ext/ogg/gstoggmux.c: (gst_ogg_mux_clear),
(gst_ogg_mux_push_buffer):
Don't generate out-of-order timestamps from oggmux, instead clamp
output timestamps to be >= the previously output ts.
Fixes#355595
Original commit message from CVS:
* ext/gnomevfs/gstgnomevfssrc.c:
Add docs about icydemux usage in connection with gnomevfssrc
* ext/libvisual/visual.c:
* ext/ogg/gstoggaviparse.c:
* ext/ogg/gstoggdemux.c:
* ext/ogg/gstoggmux.c:
* ext/ogg/gstoggparse.c:
* gst-libs/gst/audio/gstaudiofiltertemplate.c:
* gst-libs/gst/audio/gstaudiosink.c:
* gst-libs/gst/audio/gstaudiosrc.c:
* gst/audiorate/gstaudiorate.c:
More G_OBJECT macro fixing.
* gst/audiotestsrc/gstaudiotestsrc.h:
Fix wrong info in header due to copy & paste
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_init),
(gst_ogg_demux_finalize), (gst_ogg_demux_perform_seek),
(gst_ogg_demux_loop):
Send the GST_EVENT_NEW_SEGMENT from the streaming thread.
Original commit message from CVS:
* ext/ogg/gstogmparse.c: (gst_ogm_audio_parse_base_init),
(gst_ogm_video_parse_base_init), (gst_ogm_text_parse_base_init),
(gst_ogm_parse_class_init), (gst_ogm_parse_dispose),
(gst_ogm_parse_init), (gst_ogm_audio_parse_init),
(gst_ogm_video_parse_init), (gst_ogm_text_parse_init),
(gst_ogm_parse_stream_header), (gst_ogm_parse_comment_packet),
(gst_ogm_text_parse_strip_trailing_zeroes),
(gst_ogm_parse_data_packet), (gst_ogm_parse_chain),
(gst_ogm_parse_sink_event), (gst_ogm_parse_change_state):
Refactor ogm parse, do better input checking, misc. clean-ups.
Cache incoming events and push them once the source pad has
been created. Don't pass unterminated strings to sscanf().
Strip trailing zeroes from subtitle text output, since they
are not valid UTF-8. Don't push vorbiscomment packets on
the subtitle text pad. Output perfect streams if possible.
Original commit message from CVS:
* ext/ogg/gstogmparse.c: (gst_ogm_parse_chain):
Fix crash due to broken bitstream parsing on x86-64: can't make
any assumptions about sizeof(struct) due to alignment/packing
differences on different architectures. Fixes#351790.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_submit_buffer),
(gst_ogg_demux_get_next_page), (gst_ogg_demux_perform_seek),
(gst_ogg_demux_read_chain), (gst_ogg_demux_loop):
Add some more debug info.
Don't crash when a seek failed.
Actually return the result of the seek instead of TRUE.
Ignore multiple BOS pages with the same serial so that we don't create
the same stream multiple times.
Post an error when we fail to do the initial seek.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_src_query):
Implement SEEKING query in its most basic form, so that we can
at least check if we're seekable or not (#350655).
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_internal_chain),
(gst_ogg_demux_do_seek), (gst_ogg_demux_read_chain):
Make seeking in ogg more accurate again by doing the more correct
granuletime to stream time conversion.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_internal_chain),
(gst_ogg_pad_submit_packet), (gst_ogg_demux_read_chain):
*sigh*, when is the compiler going to warn when the comments
are out-of-sync with the code.. Refix case of busted theora
headers with 0 granule pos.
Original commit message from CVS:
2006-07-14 Andy Wingo <wingo@pobox.com>
* ext/vorbis/vorbisparse.h:
* ext/vorbis/vorbisparse.c (vorbis_parse_drain_event_queue)
(vorbis_parse_push_headers, vorbis_parse_clear_queue)
(vorbis_parse_drain_queue_prematurely, )
(vorbis_parse_sink_event, vorbis_parse_change_state): Queue events
until we have initialized our state. Fixes seeking after an
initial pad block.
2006-07-14 Andy Wingo <wingo@pobox.com>
Patch by: Iain * <iaingnome@gmail.com>
* ext/ogg/gstoggdemux.c (gst_ogg_demux_finalize): Fix memleak.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_chain_peer),
(gst_ogg_demux_activate_chain), (gst_ogg_demux_combine_flows),
(gst_ogg_demux_loop):
Combine GstFlowReturn from the source pads to give a
meaningfull result to the upstream peer or to stop the
processing task in case of errors.
Original commit message from CVS:
* ext/ogg/Makefile.am:
* ext/ogg/gstogmparse.c: (gst_ogm_parse_chain):
Extract language tags from ogm subtitle streams, so that
the subtitle menu choices are labelled correctly in
Totem (fixes#344708).
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_perform_seek):
Don't accidently send GST_CLOCK_TIME_NONE as a new segment start
value. Fixes g-critical on trying to play back ogg containing
unknown codec.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_typefind):
Use gst_type_find_helper_for_buffer() to find the type
of stream from the first packet.
* configure.ac:
Bump requirements to core CVS (needed for vorbis
typefinding to work).
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_submit_packet):
Fix seeking performance in the case where a non-header
packet has a 0 granulepos (busted theora case).
Fixes#341719
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_demux_chain_peer),
(gst_ogg_chain_mark_discont), (gst_ogg_chain_new_stream),
(gst_ogg_demux_activate_chain), (gst_ogg_demux_perform_seek):
Mark buffers with DISCONT after seek and after activating new
chains.
* ext/theora/gsttheoradec.h:
* ext/theora/theoradec.c: (gst_theora_dec_reset),
(theora_get_query_types), (theora_dec_sink_event),
(theora_dec_push), (theora_handle_data_packet), (theora_dec_chain),
(theora_dec_change_state):
Fix frame counter.
Detect and mark DISCONT buffers.
* ext/vorbis/vorbisdec.c: (vorbis_dec_src_query),
(vorbis_dec_sink_event), (vorbis_dec_push), (vorbis_dec_chain),
(vorbis_dec_change_state):
* ext/vorbis/vorbisdec.h:
Use GstSegment.
Detect and mark DISCONT buffers.
Don't crash on 0 sized buffers.
Original commit message from CVS:
2006-04-26 Thomas Vander Stichele <thomas at apestaart dot org>
patch by: Wim Taymans
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_submit_packet),
(gst_ogg_demux_perform_seek):
make sure correct newsegments are sent, so that the decoder
and the demuxer agree on timestamps. Fixes playback of a lot
of Ogg files that do not start from 0. Fixes#339833.
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_src_query),
(gst_ogg_demux_chain_peer), (gst_ogg_pad_submit_packet),
(gst_ogg_chain_free), (gst_ogg_demux_sink_event),
(gst_ogg_demux_loop):
More cleanups.
Respect segment stop when emiting EOS or SEGMENT_DONE.
Fixes (#337945).
Original commit message from CVS:
* ext/ogg/gstoggdemux.c: (gst_ogg_pad_src_query),
(gst_ogg_demux_receive_event), (gst_ogg_pad_event),
(gst_ogg_demux_init), (gst_ogg_demux_finalize),
(gst_ogg_demux_sink_event), (gst_ogg_demux_get_data),
(gst_ogg_demux_loop):
Don't leak events.
Remember what error we got when finding chains, if we
were shutdown, that would not be an error.
Original commit message from CVS:
2006-04-06 Andy Wingo <wingo@pobox.com>
* ext/ogg/gstoggmux.c (gst_ogg_mux_push_buffer): Set caps on
buffers being pushed out. Fixes oggmux ! multifdsink.
Original commit message from CVS:
* ext/ogg/gstoggmux.c: (gst_ogg_mux_queue_pads):
Oggmux sucks.
Make it suck slightly less by writing out the final page.
Still can't encode a vorbis-in-ogg file correctly, though.
Original commit message from CVS:
* ext/ogg/gstoggmux.c: (gst_ogg_mux_init):
* gst/adder/gstadder.c: (gst_adder_init):
use DEBUG_FUNCPTR for collectpads
Original commit message from CVS:
* ext/ogg/gstoggparse.c: (gst_ogg_parse_find_stream),
(gst_ogg_parse_chain):
Fix very inefficient usage of linked lists (#335365).
