platform/upstream/gstreamer.git
11 years agoandroidmedia: fix access to invalid buffers in the decoding loop
Andoni Morales Alastruey [Fri, 18 Oct 2013 09:54:12 +0000 (11:54 +0200)]
androidmedia: fix access to invalid buffers in the decoding loop

Flushing the decoder invalidates all buffers, so it should be done
after quiting the decoding loop. Otherwise we can jump into
"failed_release" and stop everything

https://bugzilla.gnome.org/show_bug.cgi?id=711156

11 years agoh264parse: Unify value setting
Edward Hervey [Mon, 28 Oct 2013 08:33:47 +0000 (09:33 +0100)]
h264parse: Unify value setting

the initial par_n = par_d = 0; was always overwritten since the switch/case
handles all values

And remove the 0 case (it's the same handling as default)

11 years agoh264parse: Avoid ending up with undefined values
Edward Hervey [Mon, 28 Oct 2013 08:33:16 +0000 (09:33 +0100)]
h264parse: Avoid ending up with undefined values

If caps aren't fixed, we'd end up with undefined values for format
and align.

11 years agotsdemux: Demote GST_ERROR to GST_WARNING
Edward Hervey [Fri, 11 Oct 2013 09:23:16 +0000 (11:23 +0200)]
tsdemux: Demote GST_ERROR to GST_WARNING

discontinuities are not errors per-se and we handle them correctly.

11 years agohlsdemux: fix memory leak in gst_hls_demux_get_next_fragment
Alex Ashley [Fri, 25 Oct 2013 10:11:30 +0000 (11:11 +0100)]
hlsdemux: fix memory leak in gst_hls_demux_get_next_fragment

This patch fixes three memory leaks in hlsdemux, one that occurs
during normal operation and two that occur during error conditions.

The gst_hls_demux_get_next_fragment function calls
gst_fragment_get_buffer which increments the reference count
on the buffer but gst_hls_demux_get_next_fragment never calls unref on
the buffer. This means that the reference count for each downloaded
fragment never gets to zero and so its memory is never released.

This patch adds a call to gst_buffer_unref after the flags have been
updated on the buffer.

There is a leak-on-error in gst_hls_demux_decrypt_fragment if it fails
to download the key file. If the key fails to download, null is
returned without doing an unref on the encrypted fragment. The
semantics of gst_hls_demux_decrypt_fragment is that it takes ownership
of the encrypted fragment and releases it before returning.

There is a leak-on-error in gst_hls_src_buf_to_utf8_playlist in the
unlikely event that the gst_buffer_map fails. In the "happy path"
operation of gst_hls_src_buf_to_utf8_playlist the buffer gets an unref
before the function returns, therefore the error condition must do the
same.

https://bugzilla.gnome.org/show_bug.cgi?id=710881

11 years agompegtsmux: fix uninitialized variable warning
Matthieu Bouron [Wed, 23 Oct 2013 16:11:20 +0000 (17:11 +0100)]
mpegtsmux: fix uninitialized variable warning

https://bugzilla.gnome.org/show_bug.cgi?id=710759

11 years agompegtsdemux: fix uninitialized variable warning
Matthieu Bouron [Wed, 23 Oct 2013 15:58:49 +0000 (16:58 +0100)]
mpegtsdemux: fix uninitialized variable warning

https://bugzilla.gnome.org/show_bug.cgi?id=710759

11 years agoandroidmedia: make gst_amc_avc_profile_from_string recognize alt name
Chen Jie [Fri, 18 Oct 2013 02:12:56 +0000 (10:12 +0800)]
androidmedia: make gst_amc_avc_profile_from_string recognize alt name

https://bugzilla.gnome.org/show_bug.cgi?id=710433

11 years agopitch: Fix inverted condition in setcaps
Olivier Crête [Sat, 12 Oct 2013 20:04:25 +0000 (16:04 -0400)]
pitch: Fix inverted condition in setcaps

11 years agompegtspacketizer: Skip packets with corrupted length
Edward Hervey [Fri, 11 Oct 2013 08:32:52 +0000 (10:32 +0200)]
mpegtspacketizer: Skip packets with corrupted length

This is similar to what we do for packets with payload

11 years agotsdemux: Refactor language tag extraction
Edward Hervey [Fri, 11 Oct 2013 08:29:20 +0000 (10:29 +0200)]
tsdemux: Refactor language tag extraction

subtitling descriptor needs to be used with a different method to extract
language codes.

11 years agompegts: Add support for parsing DVB Subtitling descriptor (0x59)
Edward Hervey [Fri, 11 Oct 2013 08:21:25 +0000 (10:21 +0200)]
mpegts: Add support for parsing DVB Subtitling descriptor (0x59)

Essentially works in the same vein as the iso639 descriptor code

11 years agompegts: Fix assertion fault in ISO 639 parsing
Jesper Larsen [Thu, 10 Oct 2013 20:46:48 +0000 (22:46 +0200)]
mpegts: Fix assertion fault in ISO 639 parsing

https://bugzilla.gnome.org/show_bug.cgi?id=709180

11 years agospeed: Fix some bytes-per-frame vs. bytes-per-sample mixups
Sebastian Dröge [Wed, 9 Oct 2013 15:18:41 +0000 (17:18 +0200)]
speed: Fix some bytes-per-frame vs. bytes-per-sample mixups

Otherwise we will crash when handling multiple channels and
miscalculate timestamps.

https://bugzilla.gnome.org/show_bug.cgi?id=709734

11 years agoconfigure: Make sure to try window-system-agnostic EGL if nothing was auto-detected
Sebastian Dröge [Tue, 8 Oct 2013 09:10:49 +0000 (11:10 +0200)]
configure: Make sure to try window-system-agnostic EGL if nothing was auto-detected

11 years agompegpsdemux: log time format as name instead of the number
Stefan Sauer [Tue, 8 Oct 2013 05:56:11 +0000 (07:56 +0200)]
mpegpsdemux: log time format as name instead of the number

11 years agoegl: Allow disabling gstreamer-egl library
Gilles Dartiguelongue [Sun, 6 Oct 2013 15:37:38 +0000 (17:37 +0200)]
egl: Allow disabling gstreamer-egl library

https://bugzilla.gnome.org/show_bug.cgi?id=709531

11 years agotsdemux: Use mpegts-lib for ISO 639 language tags
Jesper Larsen [Sat, 5 Oct 2013 12:45:33 +0000 (14:45 +0200)]
tsdemux: Use mpegts-lib for ISO 639 language tags

https://bugzilla.gnome.org/show_bug.cgi?id=709180

11 years agotsdemux: Return descriptor in get_descriptor
Jesper Larsen [Sat, 5 Oct 2013 12:45:32 +0000 (14:45 +0200)]
tsdemux: Return descriptor in get_descriptor

https://bugzilla.gnome.org/show_bug.cgi?id=709180

11 years agompegts: Add ISO 639 parsing functions
Jesper Larsen [Sat, 5 Oct 2013 12:45:31 +0000 (14:45 +0200)]
mpegts: Add ISO 639 parsing functions

https://bugzilla.gnome.org/show_bug.cgi?id=709180

11 years agodvdspu: handle sticky dvd events
Matej Knopp [Fri, 4 Oct 2013 22:00:52 +0000 (00:00 +0200)]
dvdspu: handle sticky dvd events

palette-changed events are now sticky.

https://bugzilla.gnome.org/show_bug.cgi?id=709454

11 years agohlsdemux: Small improvement on always valid if condition
Thiago Santos [Thu, 3 Oct 2013 10:57:17 +0000 (07:57 -0300)]
hlsdemux: Small improvement on always valid if condition

No need to check for !cancelled as the above if guarantees it
to be true

11 years agohlsdemux: Do not call _stop holding the updates lock
Thiago Santos [Thu, 3 Oct 2013 10:56:21 +0000 (07:56 -0300)]
hlsdemux: Do not call _stop holding the updates lock

It will cause a deadlock and the calers for _get_next_fragment
will already call _stop if required when _get_next_fragment fails.

Fixes #690148

11 years agodashdemux: Improve if cause readability
Greg Rutz [Wed, 12 Jun 2013 16:50:51 +0000 (10:50 -0600)]
dashdemux: Improve if cause readability

Fixed up the error-handling code when downloading fragments.
Modifed the error-handling code to use positive logic when
testing for cancellation of the download loop.

https://bugzilla.gnome.org/show_bug.cgi?id=701404

11 years agodashdemux: actually return true on latency query
Thiago Santos [Thu, 26 Sep 2013 20:05:24 +0000 (17:05 -0300)]
dashdemux: actually return true on latency query

Do not forget to set the return for latency query

11 years agodashdemux: stop fetching live fragments that don't yet exist
Alex Ashley [Thu, 26 Sep 2013 19:13:33 +0000 (16:13 -0300)]
dashdemux: stop fetching live fragments that don't yet exist

There is an issue for live streams where download_loop will keep
downloading segments until it gets a 404 error for a segment
that has not yet been published. This is a problem because this
request for a segment that doesn't exist will propagate all the
way back to the origin server(s). This means that dashdemux causes
extra load on the origin server(s) for segments that aren't yet
available.

This patch uses availabilityStartTime, period
and the host's idea of UTC to decide if a fragment is available to
be requested from an HTTP server and filter out requests for fragments
that are not yet available.

https://bugzilla.gnome.org/show_bug.cgi?id=701404

11 years agoh264parse: Fix conversion to AVC if a SPS/PPS not from the first indizes is used
Sebastian Dröge [Thu, 3 Oct 2013 11:26:06 +0000 (13:26 +0200)]
h264parse: Fix conversion to AVC if a SPS/PPS not from the first indizes is used

https://bugzilla.gnome.org/show_bug.cgi?id=707974

11 years agoopenjpeg: Port OpenJPEG plugin to the 2.0 API
Sebastian Dröge [Wed, 2 Oct 2013 08:55:30 +0000 (10:55 +0200)]
openjpeg: Port OpenJPEG plugin to the 2.0 API

11 years agovtenc: Add support for I420
Dominik Röttsches [Wed, 2 Oct 2013 02:49:43 +0000 (05:49 +0300)]
vtenc: Add support for I420

https://bugzilla.gnome.org/show_bug.cgi?id=709241

11 years agovtenc: Use correct strides, etc from the GstVideoFrame
Dominik Röttsches [Wed, 2 Oct 2013 02:49:34 +0000 (05:49 +0300)]
vtenc: Use correct strides, etc from the GstVideoFrame

https://bugzilla.gnome.org/show_bug.cgi?id=706211

11 years agocodecparsers: h264: store quantization matrices in zig-zag order
Sreerenj Balachandran [Tue, 20 Aug 2013 14:03:38 +0000 (17:03 +0300)]
codecparsers: h264: store quantization matrices in zig-zag order

Quantizer matrices are encoded in zigzag scan order in the bitstream,
so always parse it as it is.

https://bugzilla.gnome.org/show_bug.cgi?id=708629

11 years agowin32: update with the new file libgstcodecparsers.def
Sreerenj Balachandran [Wed, 2 Oct 2013 08:40:05 +0000 (11:40 +0300)]
win32: update with the new file libgstcodecparsers.def

https://bugzilla.gnome.org/show_bug.cgi?id=708629

11 years agocodecparsers: h264: add helpers to convert quantization matrices
Sreerenj Balachandran [Wed, 2 Oct 2013 08:24:58 +0000 (11:24 +0300)]
codecparsers: h264: add helpers to convert quantization matrices

Add utility functions to convert quantization matrices from zig-zag scan
order into raster scan order and vice-versa

https://bugzilla.gnome.org/show_bug.cgi?id=708629

11 years agoAdd gstreamer-mpegts pkg-config file.
Brendan Long [Mon, 30 Sep 2013 21:32:55 +0000 (15:32 -0600)]
Add gstreamer-mpegts pkg-config file.

https://bugzilla.gnome.org/show_bug.cgi?id=709145

11 years agoopencv: don't unref NULL caps in finalize
Tim-Philipp Müller [Sun, 29 Sep 2013 22:43:23 +0000 (23:43 +0100)]
opencv: don't unref NULL caps in finalize

Fixes gst-inspect-1.0 -a and generic states unit test.

11 years agoiosassetsrc: Port to 1.0 and fix some bugs on the way
Sebastian Dröge [Thu, 26 Sep 2013 11:37:09 +0000 (13:37 +0200)]
iosassetsrc: Port to 1.0 and fix some bugs on the way

11 years agoiosassetsrc: Add new element from the GStreamer SDK project
Andoni Morales Alastruey [Thu, 26 Sep 2013 11:21:59 +0000 (13:21 +0200)]
iosassetsrc: Add new element from the GStreamer SDK project

11 years agoeglglessink: Port iOS part to 1.0
Sebastian Dröge [Mon, 23 Sep 2013 12:36:40 +0000 (14:36 +0200)]
eglglessink: Port iOS part to 1.0

11 years agoconfigure: Fix iOS check for eglglessink
Sebastian Dröge [Mon, 23 Sep 2013 11:20:46 +0000 (13:20 +0200)]
configure: Fix iOS check for eglglessink

11 years agoapplemedia: avfvideosrc is not ported to 1.0 yet
Sebastian Dröge [Mon, 23 Sep 2013 09:35:56 +0000 (11:35 +0200)]
applemedia: avfvideosrc is not ported to 1.0 yet

11 years agoliveadder: round when calculation length from duration
Michael Olbrich [Wed, 18 Sep 2013 15:51:26 +0000 (17:51 +0200)]
liveadder: round when calculation length from duration

liveadder sometimes calculates the offsets incorrectly before adding. The
resulting errors can easily be heard when mixing silence with a sine.
I'm not sure what the exact conditions are to trigger this, but it definitively
happens when the buffers of two streams have a different duration and buffer
length and duration don't match exactly for one stream because of rounding
errors (e.g.  duration=0:00:00.021333333)

I have to admit, I got lost in the math somewhere but it seems that not
rounding in gst_live_adder_length_from_duration() causes 1 sample overlaps in
consecutive buffers from the same stream.
When using gst_util_uint64_scale_int_round() instead of just truncating the
sine sound correctly again.

https://bugzilla.gnome.org/show_bug.cgi?id=708345

11 years agohlssink: Use floats for the EXTINF duration values
Javier Jardón [Wed, 25 Sep 2013 14:22:12 +0000 (15:22 +0100)]
hlssink: Use floats for the EXTINF duration values

https://bugzilla.gnome.org/show_bug.cgi?id=708851

11 years agohlssink: Write EXT-X-VERSION tag in the playlist file
Javier Jardón [Wed, 25 Sep 2013 14:18:39 +0000 (15:18 +0100)]
hlssink: Write EXT-X-VERSION tag in the playlist file

https://bugzilla.gnome.org/show_bug.cgi?id=708851

11 years agocheck: Update uvch264demux for behaviour change
Edward Hervey [Thu, 19 Sep 2013 12:53:46 +0000 (14:53 +0200)]
check: Update uvch264demux for behaviour change

Since 006e7a3428bd097fc6e6c648332bff583dd434f9 we no longer error out on
incomplete aux segment.

11 years agompegts: Add SCTE/EISS sections and table types
Steve Maynard [Fri, 27 Sep 2013 14:45:00 +0000 (14:45 +0000)]
mpegts: Add SCTE/EISS sections and table types

https://bugzilla.gnome.org/show_bug.cgi?id=708688

11 years agotsdemux: Wait for valid PCR/offset obvervations
Edward Hervey [Mon, 29 Jul 2013 06:10:07 +0000 (08:10 +0200)]
tsdemux: Wait for valid PCR/offset obvervations

It is quite possible that we might get PTS/DTS before the first
PCR/Offset observation.

In order to end up with valid timestamp we wait until at least one
stream was able to get a proper running-time for any PTS/DTS.
Until then, we queue up the pending buffers to push out.

Once we see a first valid timestamp, we re-evaluate the amount of
running-time elapsed (based on returned inital running-time and amount
of data/DTS queued up) for any given stream.

Taking the biggest amount of elapsed time, we set that on the packetizer
as the initial offset and recalculate all pending buffers running-time
PTS/DTS.

Note: The buffer queueing system can also be used later on for the
dvb fast start proposal (where we queue up all stream packets before
seeing PAT/PMT and then push them once we know if they belong to the
chosen program).

11 years agoexamples: Add descriptor dumping
Edward Hervey [Sun, 7 Jul 2013 13:04:18 +0000 (15:04 +0200)]
examples: Add descriptor dumping

11 years agompegtsdemux: New PCR<=>Offset estimation code
Edward Hervey [Fri, 26 Jul 2013 05:54:30 +0000 (07:54 +0200)]
mpegtsdemux: New PCR<=>Offset estimation code

This allows:
* Better duration estimation
* More accurate PCR location
* Overall more accurate running-time location and calculation

Location and values of PCR are recorded in groups (PCROffsetGroup)
with notable PCR/Offset observations in them (when bitrate changed
for example). PCR and offset are stored as 32bit values to
reduce memory usage (they are differences against that group's
first_{pcr|offset}.

Those groups each contain a global PCR offset (pcr_offset) which
indicates how far in the stream that group is.

Whenever new PCR values are observed, we store them in a sliding
window estimator (PCROffsetGroupCurrent).

When a reset/wrapover/gap is detected, we close the current group with
current values and start a new one (the pcr_offset of that new group
is also calculated).

When a notable change in bitrate is observed (+/- 10%), we record
new values in the current group. This is a compromise between
storing all PCR/offset observations and none, while at the same time
providing better information for running-time<=>offset calculation
in VBR streams.

Whenever a new non-contiguous group is start (due to seeking for example)
we re-evaluate the pcr_offset of each groups. This allows detecting as
quickly as possible PCR wrapover/reset.

When wanting to find the offset of a certain running-time, one can
iterate the groups by looking at the pcr_offset (which in essence *is*
the running-time of that group in the overall stream).
Once a group (or neighbouring groups if the running-time is between two
groups) is found, once can use the recorded values to find the most
accurate offset.

Right now this code is only used in pull-mode , but could also
be activated later on for any seekable stream, like live timeshift
with queue2.

Future improvements:
* some heuristics to "compress" the stored values in groups so as to keep
  the memory usage down while still keeping a decent amount of notable
  points.
* After a seek compare expected and obtained PCR/Offset and if the
  difference is too big, re-calculate position with newly observed
  values and seek to that more accurate position.

Note that this code will *not* provide keyframe-accurate seeking, but
will allow a much more accurate PCR/running-time/offset location on
any random stream.
For past (observed) values it will be as accurate as can be.
For future values it will be better than the current situation.
Finally the more you seek, the more accurate your positioning will be.

11 years agompegtspacketizer: No longer use a private struct
Edward Hervey [Fri, 5 Jul 2013 15:11:46 +0000 (17:11 +0200)]
mpegtspacketizer: No longer use a private struct

These are not public headers, it just adds complexity for no reason

11 years agohlsdemux: Fix dereferencing of NULL pointer
Alex Ashley [Thu, 26 Sep 2013 15:51:25 +0000 (16:51 +0100)]
hlsdemux: Fix dereferencing of NULL pointer

On some live HLS streams, gst_hls_demux_switch_playlist causes
assertion failures because it tried to dereference a NULL fragment.
This is because g_queue_peek_tail sometimes was returning NULL and
this case was not being checked.

This patch does two things:
* move the g_queue_peek_tail inside the semaphore protection
* check if q_queue_peek_tail returns NULL

https://bugzilla.gnome.org/show_bug.cgi?id=708849

11 years agompegtsparse: Queue buffers until we have enough to know the caps
Sebastian Dröge [Fri, 27 Sep 2013 14:10:54 +0000 (16:10 +0200)]
mpegtsparse: Queue buffers until we have enough to know the caps

https://bugzilla.gnome.org/show_bug.cgi?id=708222

11 years agompegtspacketizer: rework TS packet sync and extraction
Arnaud Vrac [Mon, 16 Sep 2013 09:46:27 +0000 (11:46 +0200)]
mpegtspacketizer: rework TS packet sync and extraction

The previous code could enter an infinite loop because the adapter state
could get out of sync with its mapped data state after sync was lost.
The code was pretty confusing so it's been rewritten to be clearer.

The easiest way to reproduce the infinite loop is to use the breakmydata
element before tsdemux to trigger a resync.

https://bugzilla.gnome.org/show_bug.cgi?id=708161

11 years agotsdemux: fix buffer overflow
Arnaud Vrac [Mon, 16 Sep 2013 09:42:48 +0000 (11:42 +0200)]
tsdemux: fix buffer overflow

This can happen with a corrupt TS file, found with breakmydata element
plugged before tsdemux.

https://bugzilla.gnome.org/show_bug.cgi?id=708161

11 years agoconfigure: remove AS_SCRUB_INCLUDE
Tim-Philipp Müller [Wed, 25 Sep 2013 22:57:42 +0000 (23:57 +0100)]
configure: remove AS_SCRUB_INCLUDE

Shouldn't be needed any more.

https://bugzilla.gnome.org/show_bug.cgi?id=707658

11 years agobluez: Fix compilation on big endian systems
Sebastian Dröge [Wed, 25 Sep 2013 18:25:03 +0000 (20:25 +0200)]
bluez: Fix compilation on big endian systems

11 years agoAutomatic update of common submodule
Tim-Philipp Müller [Tue, 24 Sep 2013 17:31:10 +0000 (18:31 +0100)]
Automatic update of common submodule

From 6b03ba7 to 865aa20

11 years agoconfigure: Actually use 1.3.0.1 as version to make configure happy
Sebastian Dröge [Tue, 24 Sep 2013 13:05:28 +0000 (15:05 +0200)]
configure: Actually use 1.3.0.1 as version to make configure happy

11 years agoBack to development
Sebastian Dröge [Tue, 24 Sep 2013 13:00:27 +0000 (15:00 +0200)]
Back to development

11 years agoRelease 1.2.0
Sebastian Dröge [Tue, 24 Sep 2013 12:37:58 +0000 (14:37 +0200)]
Release 1.2.0

11 years agoUpdate .po files
Sebastian Dröge [Tue, 24 Sep 2013 12:37:29 +0000 (14:37 +0200)]
Update .po files

11 years agosdpdemux: Change rank to NONE until it can be autoplugged properly
Sebastian Dröge [Mon, 23 Sep 2013 14:18:43 +0000 (16:18 +0200)]
sdpdemux: Change rank to NONE until it can be autoplugged properly

https://bugzilla.gnome.org/show_bug.cgi?id=702495

11 years agoAutomatic update of common submodule
Edward Hervey [Fri, 20 Sep 2013 14:17:14 +0000 (16:17 +0200)]
Automatic update of common submodule

From b613661 to 6b03ba7

11 years agoaudiofxbad: Change plugin name to audiofxbad from audiochannelmix
Sebastian Dröge [Thu, 19 Sep 2013 18:17:01 +0000 (20:17 +0200)]
audiofxbad: Change plugin name to audiofxbad from audiochannelmix

11 years agodash/smoothstreaming: Remove unused debug category with the same name
Sebastian Dröge [Thu, 19 Sep 2013 18:09:26 +0000 (20:09 +0200)]
dash/smoothstreaming: Remove unused debug category with the same name

It caused static linking to fail.

11 years agoAutomatic update of common submodule
Tim-Philipp Müller [Thu, 19 Sep 2013 17:43:28 +0000 (18:43 +0100)]
Automatic update of common submodule

From 74a6857 to b613661

11 years agoAutomatic update of common submodule
Tim-Philipp Müller [Thu, 19 Sep 2013 16:36:03 +0000 (17:36 +0100)]
Automatic update of common submodule

From 098c0d7 to 74a6857

11 years agoeglglessink: Fix static linking for Android by passing --tag=CC to libtool
Sebastian Dröge [Thu, 19 Sep 2013 14:44:48 +0000 (16:44 +0200)]
eglglessink: Fix static linking for Android by passing --tag=CC to libtool

11 years agoBack to development
Sebastian Dröge [Thu, 19 Sep 2013 09:35:01 +0000 (11:35 +0200)]
Back to development

11 years agoRelease 1.1.90
Sebastian Dröge [Thu, 19 Sep 2013 08:50:47 +0000 (10:50 +0200)]
Release 1.1.90

11 years agoUpdate .po files
Sebastian Dröge [Thu, 19 Sep 2013 08:46:27 +0000 (10:46 +0200)]
Update .po files

11 years agopo: Update translations
Sebastian Dröge [Thu, 19 Sep 2013 07:46:34 +0000 (09:46 +0200)]
po: Update translations

11 years agoeglglessink: Update for new GstContext API
Sebastian Dröge [Wed, 18 Sep 2013 20:36:51 +0000 (22:36 +0200)]
eglglessink: Update for new GstContext API

11 years agoegl: Update for new GstContext API
Sebastian Dröge [Wed, 18 Sep 2013 20:24:07 +0000 (22:24 +0200)]
egl: Update for new GstContext API

11 years agohlsdemux: Reset GstUriDownloader cancellation when restarting to play
Olivier Crête [Tue, 17 Sep 2013 21:41:39 +0000 (17:41 -0400)]
hlsdemux: Reset GstUriDownloader cancellation when restarting to play

11 years agompegtspacketizer: Correct condition check for current next indicator
Sudip Jain [Sun, 15 Sep 2013 11:39:16 +0000 (17:09 +0530)]
mpegtspacketizer: Correct condition check for current next indicator

https://bugzilla.gnome.org/show_bug.cgi?id=708106

11 years agorawparse: fix event order
Wim Taymans [Thu, 12 Sep 2013 10:03:45 +0000 (12:03 +0200)]
rawparse: fix event order

Delay forwarding the segment event until we pushed caps.
Send STREAM_START in pull mode.

11 years agoamcvideodec: Don't put the level restrictions on the sinkpad caps
Sebastian Dröge [Thu, 12 Sep 2013 11:21:37 +0000 (13:21 +0200)]
amcvideodec: Don't put the level restrictions on the sinkpad caps

They tend to be inaccurate and having them in the sinkpad caps
prevents playback of files that would otherwise play fine.

11 years agotsdemux: respect seqnums on seeks
Thiago Santos [Tue, 10 Sep 2013 21:51:52 +0000 (18:51 -0300)]
tsdemux: respect seqnums on seeks

Pass the seqnum to other events that are consequence of the
original seek event

11 years agoopencv: We also work with OpenCV 2.4.6.1
Sebastian Dröge [Tue, 10 Sep 2013 09:37:55 +0000 (11:37 +0200)]
opencv: We also work with OpenCV 2.4.6.1

11 years agomodplug: Specify directory when including stdafx.h
Iain Lane [Tue, 10 Sep 2013 08:48:32 +0000 (08:48 +0000)]
modplug: Specify directory when including stdafx.h

modplug stopped exposing their directory in their pcfile, meaining
consumers accessing the headers directly fail to build.

http://sourceforge.net/p/modplug-xmms/git/ci/75e9b166982ed637b59ef7cbc1835a09f768923e/

11 years agouvch264src: Fix small mem leak
Olivier Crête [Mon, 9 Sep 2013 17:14:13 +0000 (13:14 -0400)]
uvch264src: Fix small mem leak

11 years agoh264parse: don't update src caps if only codec_data differs
Matej Knopp [Sat, 7 Sep 2013 21:09:31 +0000 (23:09 +0200)]
h264parse: don't update src caps if only codec_data differs

https://bugzilla.gnome.org/show_bug.cgi?id=705333

11 years agouvch264src: Let the caps be NULL if there is no peer
Olivier Crête [Thu, 5 Sep 2013 23:23:10 +0000 (19:23 -0400)]
uvch264src: Let the caps be NULL if there is no peer

11 years agouvch264src: No need to drop segment events
Olivier Crête [Thu, 5 Sep 2013 23:22:51 +0000 (19:22 -0400)]
uvch264src: No need to drop segment events

In 1.0, segment events are sticky and not additive, no need to prevent
their accumulation.

11 years agodecklinksrc: send caps event before segment
Tim-Philipp Müller [Thu, 5 Sep 2013 18:10:31 +0000 (19:10 +0100)]
decklinksrc: send caps event before segment

11 years agoeglglessink: change pool->sink->last_buffer to pool->last_buffer
Julien Isorce [Wed, 4 Sep 2013 09:58:24 +0000 (10:58 +0100)]
eglglessink: change pool->sink->last_buffer to pool->last_buffer

So that GstEGLImageBufferPool does not depend on GstEglGlesSink
The goal is still to move it into gstegl lib

11 years agoeglglessink: buffer pool does not need to maintain a ref on the display
Julien Isorce [Wed, 4 Sep 2013 09:56:12 +0000 (10:56 +0100)]
eglglessink: buffer pool does not need to maintain a ref on the display

Because it does not use it and also it may not know it when
we create the pool

11 years agoeglglessink: add GstEGLImageBufferPoolSendBlockingAllocate callback
Julien Isorce [Wed, 4 Sep 2013 09:52:51 +0000 (10:52 +0100)]
eglglessink: add GstEGLImageBufferPoolSendBlockingAllocate callback

The goal here is to prepare GstEGLBufferPool to be moved into
gstegl lib. So it has to not depend on 'gst_eglglessink_queue_object'

11 years agoeglglessink: prepare gst_egl_adaptation_allocate_eglimage to be moved
Julien Isorce [Wed, 4 Sep 2013 09:48:34 +0000 (10:48 +0100)]
eglglessink: prepare gst_egl_adaptation_allocate_eglimage to be moved

into gstegl lib or splited between gstegl lib and gstgl lib
because it both depends on egl and gl

So it has to not depend on GstEglAdaptationContext

11 years agoh264parse: Add support for stream-format=avc3
Alex Ashley [Fri, 30 Aug 2013 12:54:40 +0000 (13:54 +0100)]
h264parse: Add support for stream-format=avc3

When outputting in AVC3 stream format, the codec_data should not
contain any SPS or PPS, because they are embedded inside the stream.

In case of avc->bytestream h264parse will push the SPS and PPS from
codec_data downstream at the start of the stream, at intervals
controlled by "config-interval" and when there is a codec_data change.

In the case of avc3->bytstream h264parse detects that there is
already SPS/PPS in the stream and sets h264parse->push_codec to FALSE.
Therefore avc3->bytstream was already supported, except for the stream
type.

In the case of bystream->avc h264parse will generate codec_data caps
from the parsed SPS/PPS in the stream. However it does not remove these
SPS/PPS from the stream. bytestream->avc3 is the same as bytestream->avc
except that the codec_data must not have any SPS/PPS in it.

|--------------+-------------+-------------------|
|stream-format | SPS in-band | SPS in codec_data |
|--------------+-------------+-------------------|
| avc          | maybe       | always            |
|--------------+-------------+-------------------|
| avc3         | always      | never             |
|--------------+-------------+-------------------|

Amendment 2 of ISO/IEC 14496-15 (AVC file format) is defining a new
structure for fragmented MP4 called "avc3". The principal difference
between AVC1 and AVC3 is the location of the codec initialisation
data (e.g. SPS, PPS). In AVC1 this data is placed in the initial MOOV box
(moov.trak.mdia.minf.stbl.stsd.avc1) but in AVC3 this data goes in the
first sample of every fragment.

https://bugzilla.gnome.org/show_bug.cgi?id=702004

11 years agosoundtouch: Fix compilation with soundtouch 1.4.0
Sebastian Dröge [Tue, 3 Sep 2013 15:36:08 +0000 (17:36 +0200)]
soundtouch: Fix compilation with soundtouch 1.4.0

It used FLOAT_SAMPLES/INTEGER_SAMPLES #defines instead of ones properly
prefixed with a namespace.

https://bugzilla.gnome.org/show_bug.cgi?id=707390

11 years ago.gitignore: ignore new win32 config.h-new
Tim-Philipp Müller [Tue, 3 Sep 2013 10:14:43 +0000 (11:14 +0100)]
.gitignore: ignore new win32 config.h-new

11 years agoamcvideodec: Port tiled NV12 conversion to 1.0
Sebastian Dröge [Tue, 3 Sep 2013 09:05:02 +0000 (11:05 +0200)]
amcvideodec: Port tiled NV12 conversion to 1.0

11 years agoamcvideodec: Add FIXME comment
Sebastian Dröge [Tue, 3 Sep 2013 08:56:46 +0000 (10:56 +0200)]
amcvideodec: Add FIXME comment

11 years agoandroidmedia: add support for a new qualcomm colorspace
Jorge Luis Zapata [Mon, 12 Aug 2013 10:16:16 +0000 (10:16 +0000)]
androidmedia: add support for a new qualcomm colorspace

11 years agompegpsdemux: minor clean-up
Tim-Philipp Müller [Mon, 2 Sep 2013 22:28:38 +0000 (23:28 +0100)]
mpegpsdemux: minor clean-up

11 years agompegdemux: send events on pads that are not linked
Matej Knopp [Mon, 2 Sep 2013 20:54:32 +0000 (22:54 +0200)]
mpegdemux: send events on pads that are not linked

Someone might be waiting for certain events with a probe.

https://bugzilla.gnome.org/show_bug.cgi?id=707317

11 years agod3dvideosink: disable buffer pools
Andoni Morales Alastruey [Wed, 28 Aug 2013 18:07:48 +0000 (20:07 +0200)]
d3dvideosink: disable buffer pools

On a device lost, all the surfaces allocated in the
device need to be released before resetting the device,
which can't be done for the allocated buffers.

https://bugzilla.gnome.org/show_bug.cgi?id=706566

11 years agoChange how win32/common/config.h is updated
Tim-Philipp Müller [Mon, 2 Sep 2013 12:40:37 +0000 (13:40 +0100)]
Change how win32/common/config.h is updated

Generate win32/common/config.h-new directly from config.h.in,
using shell variables in configure and some hard-coded information.
Change top-level makefile so that 'make win32-update' copies the
generated file to win32/common/config.h, which we keep in source
control.  It's kept in source control so that the git tree is
buildable from VS.

This change is similar to the one recently applied to GStreamer
and gst-plugins-good.  The previous config.h file in -bad was in
pretty bad shape, so unlike core and base, I didn't attempt to
leave it strictly the same, but fixed it as necessary.  Needs
testing I cannot do myself.

https://bugzilla.gnome.org/show_bug.cgi?id=569015

11 years agotests: ignore new aiffparse test binary
Tim-Philipp Müller [Mon, 2 Sep 2013 11:19:12 +0000 (12:19 +0100)]
tests: ignore new aiffparse test binary