If it looks like we would be allocating a silly size for our sample
index, just bail out instead of trying to allocate it. Helps with
broken or fuzzed files where we might end up trying to malloc a
couple of hundred MBs otherwise.
Make sure we don't read beyond the atom boundary. Note that the code
behaves slightly differently in the corner case where there is not
enough atom data for the specified number of samples (n_samples_time)
in the atom, but still enough data to fill the pre-allocated index of
n_samples entries: before we would just stop parsing the stts data
and continue, whereas now we will likely error out. This should not
be a problem in practice though. We could maintain the old behaviour
by doing reads with a size check inside the loop if needed.
Use GstByteReader to parse stsz and stsc chunks, and check size of
available data before parsing it, instead of blindly assuming there
will be enough data. Fixes crashes with some fuzzed/broken files.
3GPP specs define a number of tags along with precise layout. While these
are normally expected to be found in a container whose major brand is a
3GPP brand, this may also happen when a 3GPP brand is only mentioned as a
compatible brand. Apply some checks, heuristic and fallbacks to extract
such tags as well.
Once buffering has started (with an mdat atom), continue buffering
until moov atom is reached, which handles cases with multiple
mdat atoms. Also keep adapter/offset better in sync with upstream
and fix some debug statements. Fixes#587426.
This reverts commit 5503a59a57.
Reverting this since it causes regressions with a lot of sample files
I have, all of which worked fine with the last -good release (#586891).
Whenever we alloc something based on a user-supplied size, we should
really use g_try_new(), otherwise we can easily be made to abort by
passing a ridiculously large number to us for allocing. Fixes
problems with some fuzzed files.
Check the possibly 64-bit atom size more carefully before casting it
to an int and passing it to gst_pad_pull_range(), otherwise we might
end up pulling 0 bytes, getting an empty buffer as requested and
dereferencing not available data whilst thinking we actually asked
for and got 0x1000000000000 bytes. Similar fix for push mode operation
where neededbytes ends up being 0 bytes, which makes us assert. Fixes
crash with broken or fuzzed file (NB #122378).
When there are less timestamps that there are samples, fill up the sample table
with the last know timestamp. This situation can happen when the last sample
does not decode and doesn't need a timestamp. We however calculate the total
track length using the last sample timestamp so we need to have something
sensible in there.
Fixes#585056
in24 samples are normally big-endian but an enda box can change this to
little-endian. Recurse into the in24 box and find the enda box so that we get
the endianness right.
Fixes#582515
If the codec is actually something else (e.g. mjpeg) change the caps to
match when parsing the ESDS atom.
Also, for AAC, override rate and channels with correct values read from
ESDS, since the rate/channels values elsewhere are often wrong.
Some clips (trailers) may have (length-wise) unbalanced streams,
which stalls the pipeline if seeking into that region.
Additional stream synchronization can handle this, as well as
sparse (subtitle) streams (at some later time ?)
Cater for DELTA_UNIT flag on buffers, keep track of current
position, remove and warn about edit lists if any (as those
as are de facto discarded anyway), add some debug statements
and indent fixes.
stps atoms contain "partial sync" information, which means that it's
a sync point where pts != dts. This is needed to properly handle
MPEG2, H.264, Dirac, etc., in quicktime.
Link to properties. Correct titles for examples. Document a few trivial cases. Keep lists in section file and docs/plugins/Makefile.am alphabetically ordered.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (qtdemux_parse_segments):
Catch invalid and commonly wrong playback rates in the elst atoms.
Fixes#567800.
Original commit message from CVS:
Patch by: Thiago Sousa Santos <thiagoss@embedded.ufcg.edu.br>
* gst/qtdemux/qtdemux.c:
Fix format string for guint64.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c:
In push mode, error out if we get EOS before we've created any srcpads.
Handle (in pull mode) some files that have a truncated moov atom where
the final sub-atom is a 'free' atom and the contents of that are not
present in the file.
Original commit message from CVS:
Patch by: j^ <j at oil21.org>
* gst/qtdemux/qtdemux.c: (qtdemux_video_caps),
(qtdemux_audio_caps):
Add codec mapping for xvid, fmp4 and ac3 tracks.
Fixes#565850
Original commit message from CVS:
2008-11-25 Julien Moutte <julien@fluendo.com>
* gst/qtdemux/qtdemux.c: (qtdemux_video_caps): Add MPG1 and MPG2
fourcc
to supported qtdemux video codecs as I found some video clips
using
those.
Original commit message from CVS:
Patch by: Tal Shalif <tshalif at nargila dot org>
* gst/qtdemux/qtdemux.c: (qtdemux_audio_caps):
Use G_{BIG,LITTLE}_ENDIAN instead of the non-GLib variants as
the latter don't exist on some systems (mingw). Fixes bug #561992.
Original commit message from CVS:
* gst/qtdemux/qtdemux.h (struct _GstQTDemux):
* gst/qtdemux/qtdemux.c (gst_qtdemux_do_seek): Queue up new
segment events instead of sending them from the seeking thread.
Fixes#559288.
(gst_qtdemux_push_pending_newsegment): New helper, sends out
queued newsegment events.
(gst_qtdemux_loop_state_movie): Voilà, call it here. Only need to
call it here, as we only seek when looping, and only push in the
movie state.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (qtdemux_tag_add_tmpo),
(qtdemux_tag_add_covr), (qtdemux_parse_udta):
* gst/qtdemux/qtdemux_fourcc.h:
* gst/qtdemux/qtdemux_types.c:
Add cover and alternative copyright tag, and enhance some existing
ones by marking them as container atoms.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (qtdemux_parse_trak),
(qtdemux_video_caps), (qtdemux_audio_caps):
Refactor some raw audio caps building, and handle >16-bit cases.
Fix/replace building caps from a string description.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_loop_state_header),
(gst_qtdemux_chain):
Some 'broken' files out there have atom lengths of zero...
which basically results in qtdemux consuming that atom again and again
until the *end of night* !
Detect that and emits an adequate element error message.
Original commit message from CVS:
Patch by: Jonathan Matthew <notverysmart@gmail.com>
* gst/qtdemux/qtdemux.c: (qtdemux_video_caps):
Add mapping for 'tiff' => image/tiff
Fixes#552213
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_change_state),
(gst_qtdemux_loop_state_header), (qtdemux_parse_node),
(qtdemux_parse_trak), (qtdemux_video_caps):
* gst/qtdemux/qtdemux.h:
* gst/qtdemux/qtdemux_fourcc.h:
* gst/qtdemux/qtdemux_types.c:
Add support for video/mj2 mime-type and its additional atoms/boxes.
Fixes#550646.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (qtdemux_audio_caps):
Add mapping for IMA Loki SDL MJPEG ADPCM codec.
Add some alternative byteswapped mappings that seem to pop up sometimes.
Fixes#550288.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c:
Use audio/x-qdm for caps. Collect some info - mplayer has a decoder
for it but ffmpeg does not.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_add_stream),
(qtdemux_parse_trak):
Revert ISO base media spec based pixel-aspect-ratio calculation.
Fixes#543300.
Original commit message from CVS:
Patch by: Thiago Sousa Santos <thiagoss at lcc dot ufcg dot edu dot br>
* gst/qtdemux/qtdemux.c: (qtdemux_audio_caps):
Correctly distinguish 8bit vs 16bit raw audio. Fixes#542410.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_add_stream),
(qtdemux_parse_trak):
Set pixel-aspect-ratio in caps using display width and height
provided in track.
Original commit message from CVS:
2008-05-31 Julien Moutte <julien@fluendo.com>
* gst/qtdemux/qtdemux.c: (gst_qtdemux_find_keyframe),
(gst_qtdemux_find_segment), (gst_qtdemux_perform_seek),
(gst_qtdemux_seek_to_previous_keyframe),
(gst_qtdemux_activate_segment), (gst_qtdemux_loop): Make sure we
we don't clip the segment's stop using the main segment duration
as
that could crop quite some video frames. Make reverse playback
support
more robust and support edit lists. Support seeking to the last
frame,
and fix reverse looping playback. Add some debugging.
* win32/common/config.h: Updated.
Original commit message from CVS:
Patch by: j^ <j at oil21 dot org>
* gst/qtdemux/qtdemux.c: (qtdemux_video_caps):
Add caps for DVCPRO50 and DVCPRO HD PAL/NTSC. See #526481.
Original commit message from CVS:
* configure.ac:
* gst/qtdemux/qtdemux.c: (qtdemux_parse_trak):
Use new error code for encrypted streams (which requires core CVS).
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_videosrc_template),
(gst_qtdemux_audiosrc_template):
Fix swapped pad template names, spotted by Thiago Sousa Santos.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_perform_seek),
(gst_qtdemux_activate_segment),
(gst_qtdemux_prepare_current_sample),
(gst_qtdemux_loop_state_movie), (qtdemux_parse_trak):
Make sure we always send a DISCONT after a seek by setting the sample
index to an undefined value after a seek.
Original commit message from CVS:
2007-11-20 Julien MOUTTE <julien@moutte.net>
* ext/taglib/gsttaglibmux.c: (gst_tag_lib_mux_render_tag),
(gst_tag_lib_mux_adjust_event_offsets):
* gst/qtdemux/qtdemux.c: (qtdemux_parse_theora_extension):
* sys/osxaudio/Makefile.am:
* sys/osxvideo/cocoawindow.h:
* sys/osxvideo/cocoawindow.m: Fix build on Mac OS X 10.5
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_change_state),
(gst_qtdemux_prepare_current_sample),
(gst_qtdemux_loop_state_movie), (qtdemux_parse_theora_extension),
(qtdemux_parse_node), (qtdemux_parse_trak), (qtdemux_video_caps):
* gst/qtdemux/qtdemux_fourcc.h:
* gst/qtdemux/qtdemux_types.c:
Add suppport for theora in quicktime according to XiphQT.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_change_state),
(qtdemux_parse_samples):
Properly free QTDemuxSamples array.
Protect table write with a sensible check, some files apparently DO contain
stts values starting with 0 :(
Original commit message from CVS:
* gst/avi/gstavidemux.c:
* gst/qtdemux/qtdemux.c:
Drop EOS in _handle_src_event(). Fix the refcount in qtdemux that
previous commit messed up.
Original commit message from CVS:
* gst/avi/gstavidemux.c:
* gst/qtdemux/qtdemux.c:
Sync _handle_src_event() with oggdemux. In avidemux also ref the
element when we should, but not when we its not needed.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_activate_segment),
(gst_qtdemux_combine_flows), (gst_qtdemux_loop_state_movie):
Smarter combine_flow code that also deals with downstream elements
returning UNEXPECTED when they receive data out of the segment
boundaries. Fixes#491305.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_loop_state_movie):
If the buffer was entirely clipped ... don't try sending it :)
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_activate_segment),
(gst_qtdemux_clip_buffer), (gst_qtdemux_loop_state_movie),
(qtdemux_parse_trak), (qtdemux_video_caps), (qtdemux_audio_caps):
Clip raw audio and video when we can, keep track of current output
segment.
Don't leak buffers and events when there is no output pad.
Improve debugging here and there.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (qtdemux_parse_node):
Fix parsing of esds atoms inside mp4a atoms so that we can set correct
codec_info for AAC audio. Fixes#457097 along with a whole other bunch
of qt/aac files.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (qtdemux_parse_samples),
(qtdemux_video_caps):
* gst/qtdemux/qtdemux_fourcc.h:
Add MJPG to the variants of motion jpeg.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (qtdemux_parse_trak):
For AMR-NB streams, export the AMRSpecificBox as codec_data on the
caps.
Fixes#447458
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_prepare_current_sample),
(gst_qtdemux_chain), (qtdemux_parse_samples):
* gst/qtdemux/qtdemux_dump.c: (qtdemux_dump_ctts):
* gst/qtdemux/qtdemux_dump.h:
* gst/qtdemux/qtdemux_fourcc.h:
* gst/qtdemux/qtdemux_types.c:
Process 'ctts' atoms, which are present in AVC ISO files (.mov files
with h264 video).
Use the offset present in 'ctts' to calculate the PTS for each packet
and set the PTS on outgoing buffers.
Fixes#423283
Original commit message from CVS:
* gst/qtdemux/qtdemux.c:
* gst/qtdemux/qtdemux.h:
Share qtdemux debug category across all files, otherwise all debugging
in files other than qtdemux.c would end up in the default category.
Original commit message from CVS:
* configure.ac:
* gst/qtdemux/Makefile.am:
* gst/qtdemux/qtdemux.c: (qtdemux_parse_moov):
Check for zlib and if available pass it explicitly to the linker
when linking qtdemux. If not available (or --disable-external has
been specified!), disable the bits in qtdemux that use it. Fixes
build on MingW (#392856).
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_post_progress),
(gst_qtdemux_chain):
Don't post BUFFERING messages in streaming mode if the stream
headers are behind the movie data; instead, post "progress" element
messages as a temporary solution. Apps might get confused and do
silly things to the pipeline state if they see buffering messages
from different sources and don't realize they come from different
sources (#387160).
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_chain),
(gst_qtdemux_add_stream):
Don't output g_warning for an unsupported format, just send a
GST_ELEMENT_WARNING and don't add the pad.
Fix the case where it doesn't check for a NULL pad in streaming mode.
Fixes#387137
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (qtdemux_parse_trak):
Fix crash dereferencing NULL pointer if there's no stco atom.
Fixes#387122.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_get_src_query_types),
(gst_qtdemux_handle_src_query), (gst_qtdemux_handle_src_event):
We don't support seeking in streaming mode, so don't even try.
Implement seeking query so apps can query seekability properly
(see #365414). Fix duration query.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_get_duration),
(gst_qtdemux_handle_src_query), (qtdemux_parse_trak),
(qtdemux_audio_caps):
Add AMR-WB to the list of supported formats.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (qtdemux_parse_tree):
Fix non-working redirects from inetfilm.com (handle 'alis' reference
data type as well). Fixes#378613.
Original commit message from CVS:
Patch by: René Stadler <mail at renestadler de>
* gst/qtdemux/qtdemux.c: (gst_qtdemux_get_duration),
(gst_qtdemux_handle_src_query), (qtdemux_parse_trak),
(qtdemux_audio_caps):
Fix caps for 24 bit raw PCM audio (2).
Fixes#383471.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_get_duration),
(gst_qtdemux_handle_src_query), (qtdemux_parse_trak),
(qtdemux_video_caps):
Handle more H263 variants.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_get_duration),
(gst_qtdemux_handle_src_query), (qtdemux_parse_trak):
Remove some asserts and replace them with a proper error
message. Fixes#379261.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_get_duration),
(gst_qtdemux_handle_src_query), (qtdemux_parse_trak):
Don't parse extra sample params for raw pcm. Fixes#374914.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_get_duration),
(gst_qtdemux_handle_src_query), (qtdemux_parse_tree),
(qtdemux_parse_trak):
Handle unbounded length streams a bit better. Fixes#367696.
Original commit message from CVS:
Patch by: Josep Torra Valles <josep at fluendo com>
* gst/qtdemux/qtdemux.c: (gst_qtdemux_handle_sink_event),
(next_entry_size), (qtdemux_inflate), (qtdemux_parse_moov),
(qtdemux_parse_tree), (qtdemux_parse_trak), (qtdemux_tag_add_str),
(qtdemux_tag_add_num), (qtdemux_tag_add_date),
(qtdemux_tag_add_gnre):
Make compile with Forte compiler, mostly don't do pointer arithmetic
with void pointers (#362626).
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (extract_initial_length_and_fourcc),
(gst_qtdemux_loop_state_header):
Printf format fixes.
* sys/dvb/gstdvbsrc.c:
Use "_stdint.h".
Original commit message from CVS:
* ext/faad/gstfaad.c: (gst_faad_setcaps), (gst_faad_chain),
(gst_faad_close_decoder):
Some cleanups.
Added some more debugging.
Don't ever ignore unlinked, we're not a demuxer.
* gst/qtdemux/qtdemux.c: (gst_qtdemux_add_stream):
Activate pad before adding it to the element.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c:
(gst_qtdemux_do_seek):
Reset each streams last_flow to GST_FLOW_OK.
(gst_qtdemux_activate_segment):
Removing mystic modifications for good.
Original commit message from CVS:
* gst/qtdemux/qtdemux.c: (gst_qtdemux_activate_segment),
(qtdemux_parse_tree):
put back 'segment start<=stop' change that was mystically reverted by
the last commit
