platform/upstream/gstreamer.git
12 years agopbutils: fix installer detail string version number
Tim-Philipp Müller [Thu, 25 Oct 2012 13:41:22 +0000 (14:41 +0100)]
pbutils: fix installer detail string version number

Should still be '1.0' not '1.1'. Fixs pbutils unit test.

12 years agoaudioresample: Use auto sinc table mode by default
Sebastian Dröge [Tue, 23 Oct 2012 09:16:57 +0000 (11:16 +0200)]
audioresample: Use auto sinc table mode by default

12 years agoaudioresample: added ARM NEON support
Carlos Rafael Giani [Mon, 15 Oct 2012 20:07:22 +0000 (22:07 +0200)]
audioresample: added ARM NEON support

This adds ARM NEON accelerated code paths for 16-bit integer
and 32-bit floating point samples.

It is a modified combination of patches #3 and #5 from Jyri Sarha
( http://lists.xiph.org/pipermail/speex-dev/2011-September/008240.html &
http://lists.xiph.org/pipermail/speex-dev/2011-September/008238.html )

Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
12 years agoaudioresample: changed inner_product_single semantics
Carlos Rafael Giani [Mon, 15 Oct 2012 20:21:14 +0000 (22:21 +0200)]
audioresample: changed inner_product_single semantics

This is an adaptation of patch #3 from Jyri Sarha
( http://lists.xiph.org/pipermail/speex-dev/2011-September/008240.html ),
but without the NEON optimizations (these come in a separate commit).
The idea is to replace SATURATE32(PSHR32(x, shift), a) operations with a
combined SATURATE32PSHR(x, shift, a) macro that can be optimized for
specific platforms (and also avoids rare rounding errors).

Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
12 years agoaudioresample: sinc filter performance improvements
Carlos Rafael Giani [Sun, 7 Oct 2012 01:00:52 +0000 (03:00 +0200)]
audioresample: sinc filter performance improvements

Original idea comes from Jyri Sarha
( http://lists.xiph.org/pipermail/speex-dev/2011-September/008243.html ).
Patch was discovered by Branislav Katreniak
( branislav.katreniak@streamunlimited.com ) for StreamUnlimited
( http://streamunlimited.com/ ). Tests showed up to 5x speed increase in
the resampler in the 44.1<->48kHz case.
I added the sinc-filter-mode and sinc-filter-auto-threshold properties
and the auto mode threshold tests, and adapted the code to GStreamer 1.0.

Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
12 years agoBack to feature development
Tim-Philipp Müller [Thu, 25 Oct 2012 11:19:46 +0000 (12:19 +0100)]
Back to feature development

12 years agoRelease 1.0.2
Tim-Philipp Müller [Wed, 24 Oct 2012 23:54:24 +0000 (00:54 +0100)]
Release 1.0.2

12 years agoaudiodecoder: track forced decoding state
Mark Nauwelaerts [Wed, 24 Oct 2012 12:05:56 +0000 (14:05 +0200)]
audiodecoder: track forced decoding state

12 years agostreamsynchronizer: Also send a GAP event to let audio sinks start their clock in...
Sebastian Dröge [Wed, 24 Oct 2012 11:34:15 +0000 (13:34 +0200)]
streamsynchronizer: Also send a GAP event to let audio sinks start their clock in case they did not have enough data yet

12 years agostreamsynchronizer: Use correct timestamp/duration for the GAP events
Sebastian Dröge [Wed, 24 Oct 2012 11:29:45 +0000 (13:29 +0200)]
streamsynchronizer: Use correct timestamp/duration for the GAP events

12 years agoRevert "gst: Add better support for static plugins"
Sebastian Dröge [Wed, 24 Oct 2012 11:26:22 +0000 (13:26 +0200)]
Revert "gst: Add better support for static plugins"

This reverts commit d2d79e3bc2a02ec57258e504b031f7e2d3729ea2,
which was accidentially pushed.

12 years agostreamsynchronizer: Send GAP events to advance streams
Sebastian Dröge [Wed, 24 Oct 2012 11:25:19 +0000 (13:25 +0200)]
streamsynchronizer: Send GAP events to advance streams

12 years agogst: Add better support for static plugins
Sebastian Dröge [Wed, 24 Oct 2012 10:10:44 +0000 (12:10 +0200)]
gst: Add better support for static plugins

12 years agoaudiobasesink: Add explanation to the GAP event handling code
Sebastian Dröge [Wed, 24 Oct 2012 09:22:29 +0000 (11:22 +0200)]
audiobasesink: Add explanation to the GAP event handling code

12 years agostreamsynchronizer: Create a GAP event with a sensible timestamp
Sebastian Dröge [Wed, 24 Oct 2012 07:57:23 +0000 (09:57 +0200)]
streamsynchronizer: Create a GAP event with a sensible timestamp

12 years agoaudiobasesink: Properly handle GAP events
Sebastian Dröge [Wed, 24 Oct 2012 09:16:54 +0000 (11:16 +0200)]
audiobasesink: Properly handle GAP events

These are now converted into silence buffers if they have
a duration or cause the ringbuffer and clock to be started
if they don't have a duration.

Fixes bug #685273.

12 years agostreamsynchronizer: Also propagate return value of pushing GAP event upstream
Sebastian Dröge [Tue, 23 Oct 2012 16:16:26 +0000 (18:16 +0200)]
streamsynchronizer: Also propagate return value of pushing GAP event upstream

12 years agostreamsynchronizer: Return TRUE from the EOS handler
Sebastian Dröge [Tue, 23 Oct 2012 15:37:46 +0000 (17:37 +0200)]
streamsynchronizer: Return TRUE from the EOS handler

12 years agovorbistag: add mapping for 'ALBUM ARTIST' with space
Tim-Philipp Müller [Tue, 23 Oct 2012 14:56:10 +0000 (15:56 +0100)]
vorbistag: add mapping for 'ALBUM ARTIST' with space

As found in sample file for bug #684701.

12 years agotcp: sys/socket.h is needed for getsockname() and similar functions
Sebastian Dröge [Mon, 22 Oct 2012 13:44:16 +0000 (15:44 +0200)]
tcp: sys/socket.h is needed for getsockname() and similar functions

12 years agoriff: add bpp to caps for msvideo
Wim Taymans [Mon, 22 Oct 2012 08:30:16 +0000 (10:30 +0200)]
riff: add bpp to caps for msvideo

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=686298

12 years agovideoconvert: add more debug
Wim Taymans [Mon, 22 Oct 2012 07:44:20 +0000 (09:44 +0200)]
videoconvert: add more debug

12 years agotag: remove unnecessary g_type_init() call from mklicensestable tool
Tim-Philipp Müller [Sat, 20 Oct 2012 11:59:11 +0000 (12:59 +0100)]
tag: remove unnecessary g_type_init() call from mklicensestable tool

https://bugzilla.gnome.org/show_bug.cgi?id=686456

12 years agoalsasink: fix caps leak in acceptcaps function
Tim-Philipp Müller [Sat, 20 Oct 2012 10:38:55 +0000 (11:38 +0100)]
alsasink: fix caps leak in acceptcaps function

https://bugzilla.gnome.org/show_bug.cgi?id=681192

12 years agoaudiodecoder: don't leak message strings when error is not fatal
Tim-Philipp Müller [Sat, 20 Oct 2012 10:38:10 +0000 (11:38 +0100)]
audiodecoder: don't leak message strings when error is not fatal

https://bugzilla.gnome.org/show_bug.cgi?id=681192

12 years agovideodecoder: don't leak message strings when error is not fatal
Tim-Philipp Müller [Sat, 20 Oct 2012 10:37:33 +0000 (11:37 +0100)]
videodecoder: don't leak message strings when error is not fatal

12 years agotcpserver{sink,src}: improve docs and property strings
Tim-Philipp Müller [Fri, 19 Oct 2012 17:29:00 +0000 (18:29 +0100)]
tcpserver{sink,src}: improve docs and property strings

And some minor clean-ups.

12 years agotcpserver{sink,src}: add 'current-port' property and signal actually used port
Alexandre Relange [Wed, 17 Oct 2012 10:19:56 +0000 (12:19 +0200)]
tcpserver{sink,src}: add 'current-port' property and signal actually used port

Useful when port=0 (use random available port) was requested.

https://bugzilla.gnome.org/show_bug.cgi?id=580093

12 years agoaudioconvert: enhance transforming caps
Mark Nauwelaerts [Thu, 18 Oct 2012 20:13:09 +0000 (22:13 +0200)]
audioconvert: enhance transforming caps

... so as to preserve input format precision,
and preferably not convert at all.

12 years agovorbistag: fix 'TODO' on image tag parsing
Thiago Santos [Thu, 18 Oct 2012 15:02:00 +0000 (12:02 -0300)]
vorbistag: fix 'TODO' on image tag parsing

Image tag now uses GstSample that has the buffer and caps
associated with it.

12 years agoalsa: if no formats in native endianness could be detected, try non-native endianness...
Tim-Philipp Müller [Wed, 17 Oct 2012 23:39:42 +0000 (00:39 +0100)]
alsa: if no formats in native endianness could be detected, try non-native endianness as well

This can happen, e.g. when using an USB sound card on
a big-endian device

https://bugzilla.gnome.org/show_bug.cgi?id=680904

12 years agoalsa: fix supported format detection
Tim-Philipp Müller [Wed, 17 Oct 2012 23:04:06 +0000 (00:04 +0100)]
alsa: fix supported format detection

The format probing code was assuming there'd be one caps
structure for each separate width/depth combination like
we did in 0.10 all over the place: for one, we'd query
unsigned/signed formats together for the same width/height,
and we'd add the entire current structure to the probed
caps when we find a format is supported. Now that we have
all raw formats in a single structure, this is all not going
to work so well any more. We added the entire structure with
all possible formats to the caps if we support just one format.

Fix probing so that we only return the list of actually
supported raw audio formats (with native endianness) from
get_caps().

12 years agoaudiocdsrc: mention TOCs in docs
Tim-Philipp Müller [Wed, 17 Oct 2012 18:59:57 +0000 (19:59 +0100)]
audiocdsrc: mention TOCs in docs

12 years agotheora, app: use gst_element_class_set_static_metadata()
Tim-Philipp Müller [Wed, 17 Oct 2012 15:54:14 +0000 (16:54 +0100)]
theora, app: use gst_element_class_set_static_metadata()

Avoids string copies.

12 years agovideodecoder: return NULL from _allocate_output_buffer() if alloc fails
Tim-Philipp Müller [Wed, 17 Oct 2012 09:55:01 +0000 (10:55 +0100)]
videodecoder: return NULL from _allocate_output_buffer() if alloc fails

.. instead of garbage pointer. Also log failure in debug log.
Should've returned the flow return like _allocate_output_frame().

https://bugzilla.gnome.org/show_bug.cgi?id=683098

12 years agoriff-media: fix palette extraction some more
Tim-Philipp Müller [Tue, 16 Oct 2012 10:48:32 +0000 (11:48 +0100)]
riff-media: fix palette extraction some more

We still need to make sure the palette is always at least 1024
bytes.

12 years agoriff: create palette_data buffer correctly
Tim-Philipp Müller [Mon, 15 Oct 2012 23:55:56 +0000 (00:55 +0100)]
riff: create palette_data buffer correctly

gst_buffer_copy_into() will append to any existing
memory region, so don't create a buffer and alloc
some memory, but just create an empty buffer and
let _copy_into() append the memory we want. Fixes
the palette being 2048 bytes with the first half
being filled with garbage.

https://bugzilla.gnome.org/show_bug.cgi?id=686046

12 years agoaudio: properly handle clipping of empty buffer
Mark Nauwelaerts [Mon, 15 Oct 2012 16:47:30 +0000 (18:47 +0200)]
audio: properly handle clipping of empty buffer

12 years agovideotestsrc: make and copy palette
Wim Taymans [Mon, 15 Oct 2012 14:33:24 +0000 (16:33 +0200)]
videotestsrc: make and copy palette

12 years agovideoconvert: actually copy the palette
Wim Taymans [Mon, 15 Oct 2012 14:32:25 +0000 (16:32 +0200)]
videoconvert: actually copy the palette

Copy the default palette in the destination buffer too.

12 years agodocs: fix RGB8P format description docs
Wim Taymans [Mon, 15 Oct 2012 13:50:44 +0000 (15:50 +0200)]
docs: fix RGB8P format description docs

12 years agodecodebin2: Fix group switching algorithm
David Corvoysier [Thu, 11 Oct 2012 09:36:54 +0000 (11:36 +0200)]
decodebin2: Fix group switching algorithm

There were two issues with the previous decodebin2 group switching algorithm:

Issue 1: It operated with no memory of what has been drained or not, leading to
multiple checks for chains/groups that were already drained.

Issue 2: When receiving an EOS, it only detected that a higher-level chain
was drained if it contained the pad receiving the EOS.

The following modifications have been applied:
- a new drained property has been added to GstDecodeChain
- both drained properties of chain/group are set as soon as they are detected
- the algorithm now tests agains these values

See https://bugzilla.gnome.org/show_bug.cgi?id=685938

12 years agortsprange: fix formatting and parsing of range floating-point values
Tim-Philipp Müller [Thu, 20 Sep 2012 00:07:08 +0000 (01:07 +0100)]
rtsprange: fix formatting and parsing of range floating-point values

Other locales might use a comma instead of a floating point
for floats, which might lead to parsing errors.

https://bugzilla.gnome.org/show_bug.cgi?id=684411

12 years agodocs: update for RGB8_PALETTED -> RGB8P
Tim-Philipp Müller [Fri, 12 Oct 2012 20:36:49 +0000 (21:36 +0100)]
docs: update for RGB8_PALETTED -> RGB8P

12 years agoriff: 8-bit paletted video is format RGB8P, not RGB8_PALETTED
Tim-Philipp Müller [Fri, 12 Oct 2012 20:31:25 +0000 (21:31 +0100)]
riff: 8-bit paletted video is format RGB8P, not RGB8_PALETTED

https://bugzilla.gnome.org/show_bug.cgi?id=686046

12 years agoaudiodecoder: set of base_ts for segment formats other than time
Josep Torra [Thu, 11 Oct 2012 10:54:39 +0000 (12:54 +0200)]
audiodecoder: set of base_ts for segment formats other than time

Fixes setting of converted segment start as base_ts when estimate rate
is allowed.

12 years agoaudiodecoder: Don't unref caps twice
Sebastian Dröge [Wed, 10 Oct 2012 13:49:46 +0000 (15:49 +0200)]
audiodecoder: Don't unref caps twice

Thanks to Josep Torra for noticing.

12 years agovideodecoder: finetune missing timestamp estimating
Mark Nauwelaerts [Wed, 10 Oct 2012 13:04:07 +0000 (15:04 +0200)]
videodecoder: finetune missing timestamp estimating

Monitor for reordered output timestamps, and then avoid oldest DTS
as PTS approach, and try for an oldest PTS as out PTS approach,
if at least all valid PTS available.

Avoids bogus estimating upon sparse available input PTS, and tries
to handle all-keyframe input, or input PTS which are actually DTS.

12 years agoplaysinkconvertbin: Change GST_WARNING to GST_INFO
Sebastian Dröge [Wed, 10 Oct 2012 09:50:12 +0000 (11:50 +0200)]
playsinkconvertbin: Change GST_WARNING to GST_INFO

It's not a problem if we have no converters, this only means
that none were requested at this point.

12 years agoivorbisdec: Rename debug category to prevent symbol conflict when using static linking
Sebastian Dröge [Tue, 9 Oct 2012 11:07:38 +0000 (13:07 +0200)]
ivorbisdec: Rename debug category to prevent symbol conflict when using static linking

12 years agodocs: playbin2 -> playbin
Wim Taymans [Tue, 9 Oct 2012 10:18:01 +0000 (12:18 +0200)]
docs: playbin2 -> playbin

12 years agotests: fix audio caps
Wim Taymans [Tue, 9 Oct 2012 10:17:42 +0000 (12:17 +0200)]
tests: fix audio caps

12 years agoaudio/video: update documentation for vfunc's that require chaining up
Andoni Morales Alastruey [Mon, 8 Oct 2012 10:43:03 +0000 (12:43 +0200)]
audio/video: update documentation for vfunc's that require chaining up

12 years agoconfigure: Reintroduced xmmintrin.h/emmintrin.h header checks
Carlos Rafael Giani [Sun, 7 Oct 2012 00:58:05 +0000 (02:58 +0200)]
configure: Reintroduced xmmintrin.h/emmintrin.h header checks

The audio resampler needs these for the SSE/SSE2 code paths

Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
12 years agovideo: small docs fix
Wim Taymans [Mon, 8 Oct 2012 07:21:16 +0000 (09:21 +0200)]
video: small docs fix

12 years agotests: fix video overlay_composition_premultiplied_alpha test on big-endian machines
Tim-Philipp Müller [Sun, 7 Oct 2012 18:46:45 +0000 (19:46 +0100)]
tests: fix video overlay_composition_premultiplied_alpha test on big-endian machines

The unit test was checking for alpha at the wrong position.

12 years agoBack to development (bug fixing)
Tim-Philipp Müller [Sun, 7 Oct 2012 15:52:27 +0000 (16:52 +0100)]
Back to development (bug fixing)

12 years agoRelease 1.0.1
Tim-Philipp Müller [Sun, 7 Oct 2012 14:11:10 +0000 (15:11 +0100)]
Release 1.0.1

12 years agotests: fix ABI struct headers for x86
Tim-Philipp Müller [Sun, 7 Oct 2012 12:34:06 +0000 (13:34 +0100)]
tests: fix ABI struct headers for x86

Not caused by anything we changed recently as
far as I can tell.

12 years agotests: add ABI structs header for 32-bit powerpc
Tim-Philipp Müller [Sun, 7 Oct 2012 12:13:37 +0000 (13:13 +0100)]
tests: add ABI structs header for 32-bit powerpc

12 years agotests: skip adder test_live_seeking test while it's unreliable
Tim-Philipp Müller [Sat, 6 Oct 2012 14:32:55 +0000 (15:32 +0100)]
tests: skip adder test_live_seeking test while it's unreliable

Was an issue in 0.10 as well.

https://bugzilla.gnome.org/show_bug.cgi?id=617418

12 years agoAutomatic update of common submodule
Tim-Philipp Müller [Sat, 6 Oct 2012 13:56:06 +0000 (14:56 +0100)]
Automatic update of common submodule

From 6c0b52c to 6bb6951

12 years agotests: fix test-effect-switch
Wim Taymans [Fri, 5 Oct 2012 08:59:30 +0000 (10:59 +0200)]
tests: fix test-effect-switch

Make it into an example of how to dynamically change an element
in a playing pipeline using pad blocking.

12 years agoaudioencoder: make stop() vfunc also optional
Tim-Philipp Müller [Thu, 4 Oct 2012 12:40:32 +0000 (13:40 +0100)]
audioencoder: make stop() vfunc also optional

Just change default value, since we also don't want to fail
if we want to deactivate and aren't active or want to activate
and are already active.

https://bugzilla.gnome.org/show_bug.cgi?id=685490

12 years agoaudioencoder: don't fail if the start vfunc is not implemented
Andoni Morales Alastruey [Thu, 4 Oct 2012 12:05:13 +0000 (14:05 +0200)]
audioencoder: don't fail if the start vfunc is not implemented

Fix behaviour to match documentation and decoder class behaviour.

https://bugzilla.gnome.org/show_bug.cgi?id=685490

12 years agotests: don't stop on just warnings
Wim Taymans [Thu, 4 Oct 2012 10:15:39 +0000 (12:15 +0200)]
tests: don't stop on just warnings

12 years agotests: fix scale test for 1.0
Wim Taymans [Thu, 4 Oct 2012 09:12:42 +0000 (11:12 +0200)]
tests: fix scale test for 1.0

It needs a basetransform patch that makes it prefer the order of
the caps property instead of passthrough.

12 years agoMerge branch 'master' of ssh://git.freedesktop.org/git/gstreamer/gst-plugins-base
Michael Smith [Wed, 3 Oct 2012 17:45:26 +0000 (10:45 -0700)]
Merge branch 'master' of ssh://git.freedesktop.org/git/gstreamer/gst-plugins-base

12 years agometa registration: use g_once functions to register these threadsafely.
Michael Smith [Wed, 3 Oct 2012 17:44:59 +0000 (10:44 -0700)]
meta registration: use g_once functions to register these threadsafely.

12 years agoplayback: class_ref() some types so we can create multiple playback elements at the...
Tim-Philipp Müller [Wed, 3 Oct 2012 10:37:33 +0000 (11:37 +0100)]
playback: class_ref() some types so we can create multiple playback elements at the same time

Should fix "cannot register existing type `GstPlaybinSelectorPad'" warnings
and subsequent errors when creating multiple players at the same time.

Conflicts:
gst/playback/gststreamselector.c

12 years agovideodecoder: Fix unused variable compiler warning if debugging is disabled
Sebastian Dröge [Tue, 2 Oct 2012 07:29:27 +0000 (09:29 +0200)]
videodecoder: Fix unused variable compiler warning if debugging is disabled

12 years agortsp: mark url argument of gst_rtsp_url_parse() as out arg
Sebastian Pölsterl [Mon, 1 Oct 2012 19:31:39 +0000 (21:31 +0200)]
rtsp: mark url argument of gst_rtsp_url_parse() as out arg

https://bugzilla.gnome.org/show_bug.cgi?id=685242

12 years agovideodecoder: Also use the object lock to protect the output_state
Olivier Crête [Sat, 29 Sep 2012 00:07:43 +0000 (20:07 -0400)]
videodecoder: Also use the object lock to protect the output_state

Hold both the stream and the object lock to modify the output_state,
this way it can be safely modified while hold either one or the other.

Also, only hold the object lock in the query

https://bugzilla.gnome.org/show_bug.cgi?id=684832

12 years agodocs: update for 1.0
Wim Taymans [Mon, 1 Oct 2012 09:58:36 +0000 (11:58 +0200)]
docs: update for 1.0

12 years agoencodebin: muxer sink pad is not always a request pad
Alban Browaeys [Sat, 29 Sep 2012 22:31:21 +0000 (00:31 +0200)]
encodebin: muxer sink pad is not always a request pad

GstId3Mux sink pad is an always (static) pad. Thus releasing it
as if a request pad triggers:
(sound-juicer:11826): GStreamer-CRITICAL **:
gst_element_release_request_pad: assertion `GST_PAD_PAD_TEMPLATE (pad)
== NULL || GST_PAD_TEMPLATE_PRESENCE (GST_PAD_PAD_TEMPLATE (pad)) ==
GST_PAD_REQUEST' failed

https://bugzilla.gnome.org/show_bug.cgi?id=685110

12 years agoappsrc: fix max-latency property getter
Tim-Philipp Müller [Sat, 29 Sep 2012 20:42:46 +0000 (21:42 +0100)]
appsrc: fix max-latency property getter

Was returning the min-latency value.

12 years agoPurge all references to liboil
Tim-Philipp Müller [Sat, 29 Sep 2012 10:46:56 +0000 (11:46 +0100)]
Purge all references to liboil

And remove unused ffmpegcolorspace tests in the process.

https://bugzilla.gnome.org/show_bug.cgi?id=673285

12 years agovideo{de,en}coder: fix missing timestamp estimating
Mark Nauwelaerts [Fri, 28 Sep 2012 11:59:24 +0000 (13:59 +0200)]
video{de,en}coder: fix missing timestamp estimating

... by having some more timestamp tracking in a private frame field.
Not doing so would lead to (a.o.) losing the needed minimum timestamp in
an earlier sent frame.

12 years agobasetextoverlay: Correctly handle empty text buffers
Sebastian Dröge [Thu, 27 Sep 2012 10:40:51 +0000 (12:40 +0200)]
basetextoverlay: Correctly handle empty text buffers

12 years agovideodecoder: use oldest frame DTS to estimate missing outgoing PTS
Mark Nauwelaerts [Thu, 27 Sep 2012 09:31:34 +0000 (11:31 +0200)]
videodecoder: use oldest frame DTS to estimate missing outgoing PTS

12 years agovideoencoder: use oldest frame PTS to estimate missing outgoing DTS
Mark Nauwelaerts [Wed, 26 Sep 2012 14:31:27 +0000 (16:31 +0200)]
videoencoder: use oldest frame PTS to estimate missing outgoing DTS

12 years agovideoencoder: incoming buffer DTS is irrelevant
Mark Nauwelaerts [Wed, 26 Sep 2012 14:22:56 +0000 (16:22 +0200)]
videoencoder: incoming buffer DTS is irrelevant

... and bogus anyway if PTS != DTS

12 years agotest: fix for new-sample signature
Wim Taymans [Wed, 26 Sep 2012 11:22:09 +0000 (13:22 +0200)]
test: fix for new-sample signature

The new-sample signal expects a GstFlowReturn as a result.
Add support for external subtitles as well.

12 years agovideoencoder: clip input buffers to current input segment
Mark Nauwelaerts [Tue, 25 Sep 2012 15:19:15 +0000 (17:19 +0200)]
videoencoder: clip input buffers to current input segment

... rather than to output segment, which will only be set
to current input segment if some output is produced
(coming from non-clipped input).

Also fixup debug message.

12 years agovideoconvert: Set correct plugin metadata
Sebastian Dröge [Tue, 25 Sep 2012 11:16:45 +0000 (13:16 +0200)]
videoconvert: Set correct plugin metadata

12 years agoBack to development (bug fixing)
Tim-Philipp Müller [Mon, 24 Sep 2012 15:38:35 +0000 (16:38 +0100)]
Back to development (bug fixing)

12 years agoRelease 1.0.0
Tim-Philipp Müller [Mon, 24 Sep 2012 12:35:05 +0000 (13:35 +0100)]
Release 1.0.0

12 years agovideodecoder: don't take STREAM_LOCK on upstream events
Tim-Philipp Müller [Mon, 24 Sep 2012 09:16:09 +0000 (10:16 +0100)]
videodecoder: don't take STREAM_LOCK on upstream events

Don't try to take STREAM_LOCK on upstream events such as QOS.
Protect qos-related variables with object lock instead. Fixes
possible deadlock when shutting down in certain situations.

https://bugzilla.gnome.org/show_bug.cgi?id=684658

12 years agovideotestsrc: keep track of the correct running time after renegotiations
Thiago Santos [Wed, 29 Aug 2012 19:02:11 +0000 (16:02 -0300)]
videotestsrc: keep track of the correct running time after renegotiations

Need to store the old running time and frame numbers when renegotiating and
start from 0 again when a new caps is set, preventing that framerate changes
cause timestamping issues.

For example, if a stream pushed 10 buffers on framerate=2/1, its
running time will be 5s. If a new framerate of 1/1 is set, it would
make the running time go to 10s as it would count those 10 buffers
as being sent on this new framerate.

Fixes camerbin unit test.

https://bugzilla.gnome.org/show_bug.cgi?id=682973

12 years agoadder: send stream-start event, and send caps event after stream-start
Tim-Philipp Müller [Sun, 23 Sep 2012 12:31:17 +0000 (13:31 +0100)]
adder: send stream-start event, and send caps event after stream-start

Delay sending of caps event so that it is sent only after
the stream-start event.

12 years agooggmux: send stream-start event
Tim-Philipp Müller [Sun, 23 Sep 2012 12:27:27 +0000 (13:27 +0100)]
oggmux: send stream-start event

12 years agoAutomatic update of common submodule
Tim-Philipp Müller [Sat, 22 Sep 2012 15:07:35 +0000 (16:07 +0100)]
Automatic update of common submodule

From 4f962f7 to 6c0b52c

12 years agooggmux: fix up previous commit
Tim-Philipp Müller [Fri, 21 Sep 2012 15:10:27 +0000 (16:10 +0100)]
oggmux: fix up previous commit

Was missing the header file change.

12 years agooggmux: send a segment event at the beginning
Tim-Philipp Müller [Fri, 21 Sep 2012 14:58:07 +0000 (15:58 +0100)]
oggmux: send a segment event at the beginning

12 years agovideodecoder: Update comments about forwarding/not-forwarding serialized events immed...
Sebastian Dröge [Thu, 20 Sep 2012 08:03:32 +0000 (10:03 +0200)]
videodecoder: Update comments about forwarding/not-forwarding serialized events immediately

12 years agovideodecoder: Protect all accesses to priv->output_frame with the stream lock
Olivier Crête [Thu, 20 Sep 2012 01:16:01 +0000 (21:16 -0400)]
videodecoder: Protect all accesses to priv->output_frame with the stream lock

Fixes segfault as queries/events can happen after a reset

12 years agotests: port playbin-text example to 1.0 api
Andreas Frisch [Wed, 19 Sep 2012 15:29:01 +0000 (17:29 +0200)]
tests: port playbin-text example to 1.0 api

https://bugzilla.gnome.org/show_bug.cgi?id=684084

12 years agoaudio: Explicitly specify endianness for IEC 61937 payloading
Arun Raghavan [Wed, 19 Sep 2012 03:22:45 +0000 (08:52 +0530)]
audio: Explicitly specify endianness for IEC 61937 payloading

This is required since some systems (DirectSound and OS X) manage the
final byte order themselves.

https://bugzilla.gnome.org/show_bug.cgi?id=678021

12 years agoaudioresample: mark semi-unused variable
Mark Nauwelaerts [Tue, 18 Sep 2012 11:16:39 +0000 (13:16 +0200)]
audioresample: mark semi-unused variable

../../../gst-plugins-base/gst/audioresample/gstaudioresample.c: In function 'gst_audio_resample_dump_drain':
../../../gst-plugins-base/gst/audioresample/gstaudioresample.c:729:9: warning: variable 'in_len' set but not used [-Wunused-but-set-variable]

12 years agoRelease 0.11.99
Tim-Philipp Müller [Mon, 17 Sep 2012 16:57:19 +0000 (17:57 +0100)]
Release 0.11.99