Original commit message from CVS:
2007-07-25 Julien MOUTTE <julien@moutte.net>
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_video_frame):
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_loop): Handle
buffers with invalid duration correctly even in the case where
we have a demuxer AND a decoder from ffmpeg linked.
Fixes: #460274.
Original commit message from CVS:
Patch by: Jan Arne Petersen <jpetersen at jpetersen dot org>
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_src_query):
Use total duration when the individual stream duration is not known.
Fixes#375534.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_type_find):
Reduce the probability of the MPEG-TS typefinder, because there's a
better one in recent base releases.
Original commit message from CVS:
* ext/ffmpeg/gstffmpeg.c: (gst_ffmpeg_av_find_stream_info):
* ext/ffmpeg/gstffmpeg.h:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_open):
Protect calls to ffmpeg functions that will implicitly call
avcodec_open/close with the avcodec mutex.
Fixes#433245
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_type_find):
Allow ffmpeg typefinders to try and typefind very short (<4kB)
files.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_base_init),
(gst_ffmpegdec_register):
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_base_init),
(gst_ffmpegdemux_register):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_base_init),
(gst_ffmpegenc_register):
* ext/ffmpeg/gstffmpegmux.c: (gst_ffmpegmux_base_init),
(gst_ffmpegmux_register):
Don't use GINT_TO_POINTER and GPOINTER_TO_INT with GTypes, this
can break horribly if sizeof(GType) happens to be bigger than
sizeof(int), because GPOINTER_TO_INT might then chop off some bits
of our GType (the reason this seems to works nevertheless is the
put-current-type-also-as-value-0-into-the-hash-table-as-fallback
hack used in the current code).
In any case, instead of just fixing this, let's not use a hash table
with GTypes here at all. g_type_{set|get}_qdata() seems to do the
job just as well.
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw@skynet.be>
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_register):
register yuv4mpegpipe demuxer with MARGINAL rank to
make it also typically available for auto-plugging. Fixes#399108.
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw@skynet.be>
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_perform_seek),
(gst_ffmpegdemux_aggregated_flow), (gst_ffmpegdemux_get_stream),
(gst_ffmpegdemux_loop):
Check the aggregated return value of pushes downstream to figure out if
we really need to stop the task or not.
Original commit message from CVS:
* Makefile.am:
* configure.ac:
* ext/ffmpeg/Makefile.am:
* ext/ffmpeg/gstffmpeg.c:
* ext/ffmpeg/gstffmpegdemux.c:
* ext/libpostproc/Makefile.am:
* ext/libpostproc/gstpostproc.c:
Allow building against an external FFMpeg install. Fixes: #363363
Patch by: Josselin Mouette <joss at debian dot org>
When built against an external install, the plugin description will
say "system install" instead of "local snapshot"
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_register):
Set the rank of all demuxers for which we:
_ Have no alternate demuxer,
_ And are know not to crash like madmans,
to GST_RANK_MARGINAL
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_register):
Disable the amr ffmpeg typefind function, it conflicts with
gsttypefindfunctions' implementation.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_get_stream):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_setcaps):
Printf format fixes.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_averror),
(gst_ffmpegdemux_init), (gst_ffmpegdemux_close),
(gst_ffmpegdemux_push_event), (gst_ffmpegdemux_send_event),
(gst_ffmpegdemux_src_query), (gst_ffmpegdemux_open),
(gst_ffmpegdemux_loop), (gst_ffmpegdemux_sink_activate),
(gst_ffmpegdemux_sink_activate_push),
(gst_ffmpegdemux_sink_activate_pull):
Don't try to use or remove non-existing pads.
Use _scale some more.
Don't try to do maths with invalid timestamps. Fixes#359545.
Prepare for push based scheduling.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps):
Clarify some CODEC_ID <=> mime type conversions for image formats.
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_register):
Cleanup on demuxer registration.
Bumped flv demuxer to RANK_MARGINAL since it proved solid enough.
Will have to go over more demuxers, for which we don't have any
gst-plugins-* alternative, to figure out which ones can move to
RANK_MARGINAL too.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_register):
One of the ffmpeg 'plugins' changed its name with the latest update
of the snapshot. Change the blacklisted typefind list to match.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_register):
Give ffmpeg demuxers a rank of NONE for the time being, so that
they are not auto-plugged; they are not stable enough for that
yet (e.g. #337785). Don't register AAC demuxer, it's not really
useful and can't be used as faad replacement (#338928).
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_perform_seek),
(gst_ffmpegdemux_loop):
Don't try to convert -1 values when seeking.
Adjust timestamps with start_time of the stream.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_register):
Register id3 demuxer, avi demuxer and ogg demuxer with
a rank of GST_RANK_NONE so that they don't get
autoplugged accidentally. We want people to install
the demuxers that are known to work well rather than
mostly untested stuff like ffdemux_*.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps),
(gst_ffmpeg_formatid_to_caps), (gst_ffmpeg_get_codecid_longname):
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_register):
Add new codec ids and fix up some of the codec->mime type mappings
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_sink_activate):
Post nice error message when we cannot activate in pull mode.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_open):
If the duration returned by ffmpeg is 0, it doesn't mean the file has
no duration, but that it doesn't know the end.
Fixed accordingly for segment_stop.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_init),
(gst_ffmpegdemux_close), (gst_ffmpegdemux_handle_seek),
(gst_ffmpegdemux_add), (my_safe_copy), (gst_ffmpegdemux_read_tags),
(gst_ffmpegdemux_open), (gst_ffmpegdemux_loop):
Add Tag support and keyframe seeking (for those formats where ffmpeg
actually fills in the index).
* ext/ffmpeg/gstffmpegprotocol.c: (gst_ffmpegdata_peek),
(gst_ffmpegdata_read), (gst_ffmpegdata_seek):
Add support for size querying.
Original commit message from CVS:
* ext/ffmpeg/Makefile.am:
* ext/ffmpeg/gstffmpeg.c: (plugin_init):
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_base_init),
(gst_ffmpegdemux_init), (gst_ffmpegdemux_stream_from_pad),
(gst_ffmpegdemux_handle_seek), (gst_ffmpegdemux_src_event),
(gst_ffmpegdemux_src_query_list), (gst_ffmpegdemux_src_query),
(gst_ffmpegdemux_add), (gst_ffmpegdemux_open),
(gst_ffmpegdemux_loop), (gst_ffmpegdemux_sink_activate),
(gst_ffmpegdemux_sink_activate_pull),
(gst_ffmpegdemux_change_state), (gst_ffmpegdemux_register):
* ext/ffmpeg/gstffmpegprotocol.c: (gst_ffmpegdata_open),
(gst_ffmpegdata_peek), (gst_ffmpegdata_read),
(gst_ffmpegdata_write), (gst_ffmpegdata_seek),
(gst_ffmpegdata_close):
Welcome the demuxers to 0.10 :)
For the time being the demuxers work only pull-based.
Seeking, querying works.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c:
Make type explicit.
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_release_buffer),
(gst_ffmpegdec_frame), (gst_ffmpegdec_chain),
(gst_ffmpegdec_change_state):
When we provide a buffer and get a valid return value (data was
read), but no output (have-data==0), then we need to reuse this
same output buffer, because it may be used for caching output
data. Fixes#307353.
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_loop):
Timestamp fix.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_loop):
timestamp is a GstClockTime, not a double..
fix#309149 and maybe others bug
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_open),
(gst_ffmpegdec_handle_event):
Fix breakage of testsuite by adding keyframe syncing; changed
to sync only for some streamtypes (MPEG-2/-4, for now).
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_add):
Add tags.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps),
(gst_ffmpeg_caps_to_codecid):
Add AMR mimetypes (since those are recognized in ffmpeg).
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_frame),
(gst_ffmpegdec_register):
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_register):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_register):
Register law codecs too, add time debug. Use ANY for template caps
since we don't necessarily have available decoders for recognized
formats.
Original commit message from CVS:
* TODO:
Update since some tasks are done.
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_formatid_to_caps):
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_type_find):
Add correct mimetype for FLI video, restore correct typefind
priority.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_init),
(gst_ffmpegdec_query), (gst_ffmpegdec_event), (gst_ffmpegdec_open),
(gst_ffmpegdec_negotiate), (gst_ffmpegdec_chain):
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_init),
(gst_ffmpegdemux_close), (gst_ffmpegdemux_src_query),
(gst_ffmpegdemux_loop):
Add simple query functions. Seeking is only cosmetic, it's not
actually filled in yet (in ffmpegdec).
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_type_find):
Decrease certainty a bit so we use flxdec for .flx videos.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_register):
Don't cripple rank as a hack to not register typefind functions.
Instead, just use a boolean for that. This makes ID3-tagged MP3
files play even if mad is not installed.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps),
(gst_ffmpeg_caps_to_codecid), (gst_ffmpeg_get_codecid_longname),
(gst_ffmpeg_avpicture_fill):
Add new codecIDs from new snapshot (Thu Dec 16 11:27:18 2004).
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_src_event),
(gst_ffmpegdemux_type_find), (gst_ffmpegdemux_loop):
* ext/ffmpeg/gstffmpegmux.c: (gst_ffmpegmux_loop):
API updates.
* ext/ffmpeg/gstffmpegprotocol.c: (gst_ffmpegdata_read),
(gst_ffmpegdata_seek):
A slightly smarter approach, but it still needs lost of work to
be reliable in any way...
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_loop):
Fix timestamps. Quite some of the gaming formats work now. :).
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps),
(gst_ffmpeg_caps_with_codecid):
WMV extradata (make #152798 work). J-frames are only available
in the bitstream if the J-frame bit has been set in the extradata.
If not (or if extradata is not provided), the movie won't play.
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_close),
(gst_ffmpegdec_connect), (gst_ffmpegdec_chain):
Only close ffmpeg if privdata was allocated (else it segfaults).
Autodetect encoding bugs and workaround it. Don't copy data if
decoding failed.
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_loop):
Use read_frame() instead of read_packet() which is obsolete.
Original commit message from CVS:
* ext/ffmpeg/gstffmpeg.c: (plugin_init):
Remove redundant calls which are called internally already.
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_caps_with_codecid):
Indenting...
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_loop),
(gst_ffmpegdemux_register):
Remove typefind functions and rank for demuxers for any demuxer
that we already have in GStreamer.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_src_event),
(gst_ffmpegdemux_src_query), (gst_ffmpegdemux_src_convert),
(gst_ffmpegdemux_open), (gst_ffmpegdemux_loop):
* ext/ffmpeg/gstffmpegprotocol.c: (gst_ffmpegdata_open),
(gst_ffmpegdata_read), (gst_ffmpegdata_write),
(gst_ffmpegdata_seek):
Fix up demuxer. Works now for all my files that I tried, even
quicktime. I basically hack around the super-oversimplistic file
I/O handling in ffmpeg in several ways together, and I also hack
around the fairly annoying EOS-will-pause-the-source-element in
GStreamer itself.
This code is not pretty.
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_type_find):
reduce maximum probability in typefinding to POSSIBLE. prevents
misidentification of my mp3s as video/mpeg
Original commit message from CVS:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_base_init):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_base_init):
Fix category.
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_class_init),
(gst_ffmpegdemux_init), (gst_ffmpegdemux_close):
Slightly rewrite the init function, old one was ugly. Also remove
dispose function since it's obsolete, the change_state function
cleans up already. Also fixes segfault in test-app
gstreamer/tests/instantiate/create
Original commit message from CVS:
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_codecid_to_caps):
* ext/ffmpeg/gstffmpegcodecmap.h:
* ext/ffmpeg/gstffmpegdec.c: (gst_ffmpegdec_chain),
(gst_ffmpegdec_register):
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_add),
(gst_ffmpegdemux_register):
* ext/ffmpeg/gstffmpegenc.c: (gst_ffmpegenc_connect),
(gst_ffmpegenc_register):
* ext/ffmpeg/gstffmpegmux.c: (gst_ffmpegmux_register):
Add encode property to codec mapper, so we can get better-fitting
caps. Also use this everywhere. ;). Lastly, add some new codecIDs
to our list of recognized ones.
Original commit message from CVS:
* configure.ac:
Brackets cause autoconf errors. So remove them.
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_add):
Actually use the event/convert/query functions. D'oh.
Original commit message from CVS:
* HACKING:
Add some basic documentation on how our wrapping works.
* TODO:
Add a list of things that could be worked on or that need doing.
* configure.ac:
Update snapshot.
* ext/ffmpeg/Makefile.am:
Changne .la links. See below (autotools patch).
* ext/ffmpeg/gstffmpeg.c: (plugin_init):
Enable demuxers. See below (gstffmpegdemux.c).
* ext/ffmpeg/gstffmpegcodecmap.c: (gst_ffmpeg_formatid_to_caps):
Realmedia caused a crash - fix that.
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_averror),
(gst_ffmpegdemux_base_init), (gst_ffmpegdemux_init),
(gst_ffmpegdemux_close), (gst_ffmpegdemux_dispose),
(gst_ffmpegdemux_stream_from_pad),
(gst_ffmpegdemux_src_event_mask), (gst_ffmpegdemux_src_event),
(gst_ffmpegdemux_src_format_list),
(gst_ffmpegdemux_src_query_list), (gst_ffmpegdemux_src_query),
(gst_ffmpegdemux_src_convert), (gst_ffmpegdemux_add),
(gst_ffmpegdemux_open), (gst_ffmpegdemux_loop),
(gst_ffmpegdemux_change_state), (gst_ffmpegdemux_register):
Right. OK, so I fixed up the demuxing and have it basically-working,
and the best way to get some more people to test it is to actually
enable it. I'm not sure if we want this for 0.8.0, but we can at
least give it a try. I've tested avi, matroska and mpeg, all appear
to work. The cool thing is that this gives us instant support for
several exotic formats that we'd never care about ourselves. Again,
this needs more testing for it to still be enabled in 0.8.0, but I
want to give it a try...
* ext/ffmpeg/gstffmpegmux.c: (gst_ffmpegmux_base_init),
(gst_ffmpegmux_init), (gst_ffmpegmux_request_new_pad),
(gst_ffmpegmux_connect), (gst_ffmpegmux_loop),
(gst_ffmpegmux_register):
Add some fixups that I use locally. Make it work in the case of
MPEG encoding, but the muxer is still not in shape to be enabled.
* ext/ffmpeg/gstffmpegprotocol.c: (gst_ffmpegdata_open),
(gst_ffmpegdata_read), (gst_ffmpegdata_write),
(gst_ffmpegdata_seek), (gst_ffmpegdata_close):
Some small fixups that crept into it while it was disabled for the
last few years. Basically works.
* gst-libs/ext/ffmpeg/Makefile.am:
Instead of having our local-autotoolized version, I patch the ffmpeg
source to be fully autotoolized. That means a simple SUBDIRS here
is now enough.
* gst-libs/ext/ffmpeg/Tag:
Version update.
* gst-libs/ext/ffmpeg/patch/autotools.diff:
Autotoolize ffmpeg. Needs to be sent to ffmpeg-devel@...
* gst-libs/ext/ffmpeg/patch/disableinstalllibs.diff:
Don't install their libs.
* gst-libs/ext/ffmpeg/patch/disablemmx.diff:
Don't use MMX. It cannot ocmpile using PIC.
* gst-libs/ext/ffmpeg/patch/disabletools.diff:
Don't compile/install their tools, we don't use them.
* gst-libs/ext/ffmpeg/patch/functions.diff:
Prevent symbol conflicts.
* gst-libs/ext/ffmpeg/patch/matroska.diff:
Add a matroska demuxer. Needs to be sent to ffmpeg-devel@...
Original commit message from CVS:
Make sure set_explicit_caps() is called before adding pad.
* ext/ffmpeg/gstffmpegdemux.c: (gst_ffmpegdemux_loop):
* gst/id3/gstid3types.c: (gst_id3types_loop):
* gst/mpegstream/gstmpegdemux.c: (gst_mpeg_demux_parse_syshead):
* gst/realmedia/rmdemux.c: (gst_rmdemux_add_stream):
Original commit message from CVS:
merge TYPEFIND branch. Major changes:
- totally reworked type(find) system
- all typefind functions are in gst/typefind now
- more typefind functions then before
- some plugins might fail to compile now because I don't have them installed and they
a) require bytestream or
b) haven't had their typefind fixed.
Please fix those plugins and put the typefind functions into gst/typefind if they don't have dependencies
Original commit message from CVS:
New typefind system:
* bytestream is now part of the core
* all plugins have been modified to use this new typefind system
* asf typefinding added
* mpeg video stream typefiding removed because it's broken
* duplicate typefind entries removed
* extra id3 typefinding added, because we've seen 4 types of files
(riff/wav, flac, vorbis, mp3) with id3 headers and each of these needs
to work. Instead, I've added an id3 element and let it redo typefiding
after the id3 header. this needs a hack because spider only typefinds
once. We can remove this hack once spider supports multiple typefinds.
* with all this, mp3 typefinding is semi-rewritten
* id3 typefinding in flac/vorbis is removed, it's no longer needed
* fixed spider and gst-typefind to use this, too.
* Other general cleanups
Original commit message from CVS:
compatibility fix for new GST_DEBUG stuff.
Includes fixes for missing includes for config.h and unistd.h
I only ensured for plugins I can build that they work, so if some of them are still broken, you gotta fix them yourselves unfortunately.
Original commit message from CVS:
Hm, I actually misunderstood this protocol thing, since ffmpeg provides an read/write indicator, too... Here's a semi-rewrite that handles writings too
Original commit message from CVS:
Some modifications to make the demuxers work, plus moving typefinding over from separate plugin to the demuxers themselves
Original commit message from CVS:
Release early.. try to wrap more ffmpeg stuff, like muxers and demuxers.
This thing needs a fix in ffmpeg (which I will send to the ffmpeg authors
eventually)
