platform/upstream/gstreamer.git
10 years agortsp: Link to ws2_32 on Windows
Руслан Ижбулатов [Sun, 4 May 2014 15:54:54 +0000 (15:54 +0000)]
rtsp: Link to ws2_32 on Windows

Needed for getsockname and setsockopt

https://bugzilla.gnome.org/show_bug.cgi?id=729514

10 years agoMake X11 detection more precise
Руслан Ижбулатов [Sun, 4 May 2014 15:54:06 +0000 (15:54 +0000)]
Make X11 detection more precise

Don't be content with just X11/Xlib.h, check for X11/XKBlib.h as well.
This prevents false positives (for example, from partial X11 headers
installed by tcl/tk).

https://bugzilla.gnome.org/show_bug.cgi?id=729513

10 years agotests: fix printf format compiler warning in playback test on win32
Руслан Ижбулатов [Sun, 4 May 2014 15:57:35 +0000 (15:57 +0000)]
tests: fix printf format compiler warning in playback test on win32

https://bugzilla.gnome.org/show_bug.cgi?id=729515

10 years agoAdd new unit test binary to .gitignore
Tim-Philipp Müller [Sun, 4 May 2014 17:14:54 +0000 (18:14 +0100)]
Add new unit test binary to .gitignore

10 years agodiscoverer: Add APIs to simply get installer details for missing plugins
Thibault Saunier [Tue, 14 Jan 2014 14:39:55 +0000 (15:39 +0100)]
discoverer: Add APIs to simply get installer details for missing plugins

Currently the API is far from optimal and the user has to work around
our badly defined API to simply install missing plugins.

API:
  new:
    gst_discoverer_info_get_missing_elements_installer_details

  deprecated:
    gst_discoverer_info_get_misc
    gst_discoverer_stream_info_get_misc

https://bugzilla.gnome.org/show_bug.cgi?id=720596

10 years agoBack to development
Sebastian Dröge [Sat, 3 May 2014 18:48:27 +0000 (20:48 +0200)]
Back to development

10 years agotextoverlay: Link unit test with the local version of the library, not an installed one
Sebastian Dröge [Sat, 3 May 2014 16:57:38 +0000 (18:57 +0200)]
textoverlay: Link unit test with the local version of the library, not an installed one

10 years agoRelease 1.3.1
Sebastian Dröge [Sat, 3 May 2014 15:50:10 +0000 (17:50 +0200)]
Release 1.3.1

10 years agoUpdate .po files
Sebastian Dröge [Sat, 3 May 2014 15:48:04 +0000 (17:48 +0200)]
Update .po files

10 years agopo: Update translations
Sebastian Dröge [Sat, 3 May 2014 15:22:10 +0000 (17:22 +0200)]
po: Update translations

10 years agortpbasepayload: Implement reconfigure event & renegotiation without subclass
Olivier Crête [Fri, 2 May 2014 23:09:59 +0000 (19:09 -0400)]
rtpbasepayload: Implement reconfigure event & renegotiation without subclass

Implement the reconfigure event, also do correct downstream caps negotiation
if the subclass doesn't implementy set_caps.

https://bugzilla.gnome.org/show_bug.cgi?id=725361

10 years agotests/check/libs/rtpbasepayload.c: Run gst-indent
Olivier Crête [Fri, 2 May 2014 23:09:44 +0000 (19:09 -0400)]
tests/check/libs/rtpbasepayload.c: Run gst-indent

https://bugzilla.gnome.org/show_bug.cgi?id=725361

10 years agoAutomatic update of common submodule
Sebastian Dröge [Sat, 3 May 2014 08:14:51 +0000 (10:14 +0200)]
Automatic update of common submodule

From bcb1518 to 211fa5f

10 years agortpbasepayload: Save the PT after fixating
Olivier Crête [Fri, 2 May 2014 22:30:16 +0000 (18:30 -0400)]
rtpbasepayload: Save the PT after fixating

10 years agortspdefs: remove outdated comments
Tim-Philipp Müller [Fri, 2 May 2014 18:36:34 +0000 (19:36 +0100)]
rtspdefs: remove outdated comments

10 years agortpbuffer: avoid underflow in size calculation
Vincent Penquerc'h [Fri, 2 May 2014 14:09:35 +0000 (15:09 +0100)]
rtpbuffer: avoid underflow in size calculation

10 years agovideodecoder: do not parse caps for not using it
Thiago Santos [Thu, 1 May 2014 22:31:09 +0000 (19:31 -0300)]
videodecoder: do not parse caps for not using it

Saving some cpu

10 years agortpbasepayload: restrict initial random sequence number to be <= 32767
John Bassett [Fri, 3 Jan 2014 10:06:22 +0000 (11:06 +0100)]
rtpbasepayload: restrict initial random sequence number to be <= 32767

In order to prevent SRTP roll over counter issues the initial sequence
number is restricted to <= 32767. This is recommended by RFC 4568 section 6.4.

10 years agosdp: Add some more gobject-introspection annotations for bindings
Sebastian Dröge [Thu, 1 May 2014 13:11:04 +0000 (15:11 +0200)]
sdp: Add some more gobject-introspection annotations for bindings

https://bugzilla.gnome.org/show_bug.cgi?id=729123

10 years agoplaybin: Don't block on non-serialized events
Sebastian Dröge [Thu, 1 May 2014 11:15:57 +0000 (13:15 +0200)]
playbin: Don't block on non-serialized events

https://bugzilla.gnome.org/show_bug.cgi?id=729321

10 years agoplaysink: Don't block on non-serialized events
Sebastian Dröge [Thu, 1 May 2014 11:08:24 +0000 (13:08 +0200)]
playsink: Don't block on non-serialized events

https://bugzilla.gnome.org/show_bug.cgi?id=729321

10 years agoplaysinkconvertbin: Don't block on non-serialized events
Sebastian Dröge [Thu, 1 May 2014 11:06:53 +0000 (13:06 +0200)]
playsinkconvertbin: Don't block on non-serialized events

https://bugzilla.gnome.org/show_bug.cgi?id=729321

10 years agosubtitleoverlay: Don't block on non-serialized events
Sebastian Dröge [Thu, 1 May 2014 11:05:05 +0000 (13:05 +0200)]
subtitleoverlay: Don't block on non-serialized events

https://bugzilla.gnome.org/show_bug.cgi?id=729321

10 years agortcpbuffer: check claimed data size against available size
Vincent Penquerc'h [Wed, 30 Apr 2014 10:06:27 +0000 (11:06 +0100)]
rtcpbuffer: check claimed data size against available size

Coverity 1208773

10 years agortspconnection: Empty queue when flush.
Göran Jönsson [Wed, 23 Apr 2014 06:06:36 +0000 (08:06 +0200)]
rtspconnection: Empty queue when flush.

Empty the watchs queue when calling
gst_rtsp_watch_set_flushing with flushing variabel is TRUE.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=728772

10 years agortspconnection: Add more tests
Ognyan Tonchev [Sun, 16 Mar 2014 15:09:36 +0000 (16:09 +0100)]
rtspconnection: Add more tests

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=728907

10 years agovideotestsrc: fix undefined behaviour of left-shift
Luis de Bethencourt [Tue, 29 Apr 2014 14:15:47 +0000 (10:15 -0400)]
videotestsrc: fix undefined behaviour of left-shift

With a small type for the color values being left-shifted, the result is
undefined and it could potentially overflow.

https://bugzilla.gnome.org/show_bug.cgi?id=729195

10 years agowin32: fix export files again
Tim-Philipp Müller [Tue, 29 Apr 2014 09:59:02 +0000 (10:59 +0100)]
win32: fix export files again

Revert unintended parts of d8a0927930a87a2eb60d4c98cb3fea8aed911b27

10 years agoAdd mikey.h file
Christian Fredrik Kalager Schaller [Tue, 29 Apr 2014 09:39:18 +0000 (11:39 +0200)]
Add mikey.h file

10 years agoaudiodecoder: Make caps writable before fixating
Haakon Sporsheim [Tue, 29 Apr 2014 07:58:21 +0000 (09:58 +0200)]
audiodecoder: Make caps writable before fixating

https://bugzilla.gnome.org/show_bug.cgi?id=729114

10 years agosdpmessage: Add array length annotation to gst_sdp_message_parse_buffer
Sebastian Dröge [Tue, 29 Apr 2014 07:54:18 +0000 (09:54 +0200)]
sdpmessage: Add array length annotation to gst_sdp_message_parse_buffer

https://bugzilla.gnome.org/show_bug.cgi?id=729123

10 years agortpbuffer: fix memory leak when gst_rtp_buffer_map fails
Stian Selnes [Tue, 29 Apr 2014 06:46:02 +0000 (08:46 +0200)]
rtpbuffer: fix memory leak when gst_rtp_buffer_map fails

Make sure rtp->data[3] is set before jumping to error path.

https://bugzilla.gnome.org/show_bug.cgi?id=729117

10 years agogst-play: add option to supply media files from playlist file
Ravi Kiran K N [Mon, 28 Apr 2014 13:17:06 +0000 (18:47 +0530)]
gst-play: add option to supply media files from playlist file

https://bugzilla.gnome.org/show_bug.cgi?id=728845

10 years agogiobasesink: we mustn't change the format of a query response
Tim-Philipp Müller [Sat, 26 Apr 2014 23:49:01 +0000 (00:49 +0100)]
giobasesink: we mustn't change the format of a query response

Not even in the DEFAULT case. That's bad 0.10 behaviour, no caller
is ever going to check the format of the response.

10 years agoplaybin: add nick for soft colorbalance play flag to fix gst-inspect
Tim-Philipp Müller [Sat, 26 Apr 2014 23:25:16 +0000 (00:25 +0100)]
playbin: add nick for soft colorbalance play flag to fix gst-inspect

Fix gst-inspect-1.0 playbin criticals when printing the
flags, which was caused by a missing nick name for one
of the flags.

10 years agodocs: remove outdated and pointless 'Last reviewed' lines from docs
Tim-Philipp Müller [Sat, 26 Apr 2014 22:26:09 +0000 (23:26 +0100)]
docs: remove outdated and pointless 'Last reviewed' lines from docs

They are very confusing for people, and more often than not
also just not very accurate. Seeing 'last reviewed: 2005' in
your docs is not very confidence-inspiring. Let's just remove
those comments.

10 years agogiobasesink: Implement handling of the SEEKING query
Sebastian Dröge [Fri, 25 Apr 2014 15:32:59 +0000 (17:32 +0200)]
giobasesink: Implement handling of the SEEKING query

10 years agoaudiodecoder: Plug caps leaks
Edward Hervey [Fri, 25 Apr 2014 09:30:37 +0000 (11:30 +0200)]
audiodecoder: Plug caps leaks

We were returning in various places without unreffing the caps, and
we were also leaking (overwriting) the caps we got from _get_current_caps()

Spotted by Haakon Sporsheim in #gstreamer

10 years agoaudioresample: Don't left-shift into the sign bit, instead use unsigned integers
Sebastian Dröge [Tue, 22 Apr 2014 16:28:10 +0000 (18:28 +0200)]
audioresample: Don't left-shift into the sign bit, instead use unsigned integers

10 years agotag: exif: avoid adding empty strings
Thiago Santos [Tue, 22 Apr 2014 03:21:01 +0000 (00:21 -0300)]
tag: exif: avoid adding empty strings

Fixes assertion with some jpeg files

10 years agoplay: Improve pipeline states
Wim Taymans [Mon, 21 Apr 2014 13:35:32 +0000 (15:35 +0200)]
play: Improve pipeline states

First set the pipeline to the PAUSED state to check if we are dealing
with a live pipeline or not. Then move to the desired state.

If we don't do this, it is possible that we receive a BUFFERING message
before we know that the pipeline is live and we would set the pipeline
to PAUSED and deadlock.

10 years agoplay: Update buffering state for live pipelines
Wim Taymans [Mon, 21 Apr 2014 13:33:10 +0000 (15:33 +0200)]
play: Update buffering state for live pipelines

Update the buffering variable, even for live pipelines so that we don't
print \n for each buffering message.

10 years agovideoframe: Initialise GstVideoFrame to zeroes if mapping fails
Sebastian Dröge [Wed, 16 Apr 2014 17:53:14 +0000 (19:53 +0200)]
videoframe: Initialise GstVideoFrame to zeroes if mapping fails

This should allow for more meaningful errors. Dereferencing NULL
is more useful information than dereferencing a random address
happened to be on the stack.

10 years agoexiftag: catch buffer mapping failure
Vincent Penquerc'h [Wed, 16 Apr 2014 10:43:40 +0000 (11:43 +0100)]
exiftag: catch buffer mapping failure

Might be what caused:

Coverity 1139734

10 years agoaudioresample: Fix memory leaks in test
Sebastian Dröge [Tue, 15 Apr 2014 17:17:06 +0000 (19:17 +0200)]
audioresample: Fix memory leaks in test

10 years agoaudioresample: Fix up indention
Sebastian Dröge [Tue, 15 Apr 2014 17:16:44 +0000 (19:16 +0200)]
audioresample: Fix up indention

10 years agoaudioresample: Fix out of bounds memory accesses
Sebastian Dröge [Tue, 15 Apr 2014 17:16:18 +0000 (19:16 +0200)]
audioresample: Fix out of bounds memory accesses

10 years agopango: Make static caps actually static to fix a memory leak
Sebastian Dröge [Tue, 15 Apr 2014 11:57:08 +0000 (13:57 +0200)]
pango: Make static caps actually static to fix a memory leak

10 years agovideotestsrc: Fix memory leak in test
Sebastian Dröge [Tue, 15 Apr 2014 11:54:45 +0000 (13:54 +0200)]
videotestsrc: Fix memory leak in test

10 years agoencodebin: Fix memory leak in test
Sebastian Dröge [Tue, 15 Apr 2014 11:48:46 +0000 (13:48 +0200)]
encodebin: Fix memory leak in test

10 years agoencoding-profile: Free preset name in finalize
Sebastian Dröge [Tue, 15 Apr 2014 11:48:17 +0000 (13:48 +0200)]
encoding-profile: Free preset name in finalize

10 years agooggmux: Clear Ogg streams before initing them
Sebastian Dröge [Tue, 15 Apr 2014 11:39:39 +0000 (13:39 +0200)]
oggmux: Clear Ogg streams before initing them

They might've been inited before, in which case we leak
memory when initing them again without clearing.

10 years agoaudioconvert: Fix leaks in unit test
Sebastian Dröge [Tue, 15 Apr 2014 11:03:34 +0000 (13:03 +0200)]
audioconvert: Fix leaks in unit test

10 years agovideoencoder/decoder: Fix memory leaks in the tests
Sebastian Dröge [Tue, 15 Apr 2014 09:55:22 +0000 (11:55 +0200)]
videoencoder/decoder: Fix memory leaks in the tests

10 years agoaudiodecoder: Actually allocate enough memory for 64 bits, not just 32 bits
Sebastian Dröge [Tue, 15 Apr 2014 09:53:43 +0000 (11:53 +0200)]
audiodecoder: Actually allocate enough memory for 64 bits, not just 32 bits

Also fix a memory leak.

10 years agoaudioencoder: Fix memory leaks in unit test
Sebastian Dröge [Tue, 15 Apr 2014 09:43:41 +0000 (11:43 +0200)]
audioencoder: Fix memory leaks in unit test

10 years agortp: Fix GBytes memory leak in test
Sebastian Dröge [Tue, 15 Apr 2014 08:29:12 +0000 (10:29 +0200)]
rtp: Fix GBytes memory leak in test

10 years agortpbasedepay: add stats property
Wim Taymans [Sat, 12 Apr 2014 05:10:36 +0000 (07:10 +0200)]
rtpbasedepay: add stats property

Add a stats property that holds a structure with all the current
values of the depayloader.

See https://bugzilla.gnome.org/show_bug.cgi?id=646577

10 years agortpbasepayload: update docs
Wim Taymans [Sat, 12 Apr 2014 04:43:24 +0000 (06:43 +0200)]
rtpbasepayload: update docs

10 years agortpbasepayload: add current timestamp and seqnum offset to stats
Wim Taymans [Sat, 12 Apr 2014 04:27:36 +0000 (06:27 +0200)]
rtpbasepayload: add current timestamp and seqnum offset to stats

Expose the current timestamp and seqnum offset in the stats

See https://bugzilla.gnome.org/show_bug.cgi?id=646577

10 years agotextrender: push segment event after caps event
Josep Torra [Fri, 11 Apr 2014 08:24:10 +0000 (10:24 +0200)]
textrender: push segment event after caps event

Fixes warning "Sticky event misordering, got 'segment' before 'caps'".

10 years agooggstream: use G_GUINT64_CONSTANT instead of ll suffix
Vincent Penquerc'h [Thu, 10 Apr 2014 15:08:29 +0000 (16:08 +0100)]
oggstream: use G_GUINT64_CONSTANT instead of ll suffix

Thanks slomo for pointing out it's not standard.

10 years agoxvimage: remove dead code
Vincent Penquerc'h [Thu, 10 Apr 2014 14:55:57 +0000 (15:55 +0100)]
xvimage: remove dead code

matching_attr can not be NULL here, we've tested that away a few
lines beforehand.

Coverity 1139655

10 years agovideotestsrc: bail out on unsupported caps
Vincent Penquerc'h [Thu, 10 Apr 2014 14:51:05 +0000 (15:51 +0100)]
videotestsrc: bail out on unsupported caps

This avoids using uninitialized data (and properly rejects caps).

Coverity 1139898

10 years agotypefind: remove pointless checks for data being NULL
Vincent Penquerc'h [Thu, 10 Apr 2014 14:16:03 +0000 (15:16 +0100)]
typefind: remove pointless checks for data being NULL

It was already checked in an early out, and as it's only
incremented for at most the size of the passed buffer, it
can only become NULL in an address wraparound.

While there, don't cast away const on a pointer.

Coverity 1139845

10 years agodecodebin: consider "no demuxer" case to not have dynamic pads
Vincent Penquerc'h [Thu, 10 Apr 2014 12:34:58 +0000 (13:34 +0100)]
decodebin: consider "no demuxer" case to not have dynamic pads

This fixes a possible NULL dereference.

Coverity 1195146

10 years agoencodebin: guard against gst_pad_get_peer returning NULL
Vincent Penquerc'h [Thu, 10 Apr 2014 12:28:30 +0000 (13:28 +0100)]
encodebin: guard against gst_pad_get_peer returning NULL

If it does, the pad may be leaked if it's a request pad, though.

Coverity 1139799

10 years agoencodebin: guard against pathological NULL dereference
Vincent Penquerc'h [Thu, 10 Apr 2014 12:26:42 +0000 (13:26 +0100)]
encodebin: guard against pathological NULL dereference

Coverity 1139798

10 years agoaudioresample: reject 0 denominator when creating resampler
Vincent Penquerc'h [Thu, 10 Apr 2014 11:32:24 +0000 (12:32 +0100)]
audioresample: reject 0 denominator when creating resampler

Coverity 119514011951391195138

10 years agovideo-overlay-composition: guard against NULL pointer dereference on error
Vincent Penquerc'h [Thu, 10 Apr 2014 11:14:48 +0000 (12:14 +0100)]
video-overlay-composition: guard against NULL pointer dereference on error

If gst_video_overlay_rectangle_apply_global_alpha is called with
a rectangle with unsuitable alpha, expanding the alpha plane will
fail, and thus lead to dereferencing a NULL src pointer. It's not
certain this will happen in practice, as the function is static
and callers might ensure suitable alpha before calling, but there
is no apparent explicit such check.
Add prologue asserts for proper alpha to explicitely prevent this.

Coverity 1139707

10 years agovideometa: fix texture_type memcpy size
Vincent Penquerc'h [Thu, 10 Apr 2014 11:10:47 +0000 (12:10 +0100)]
videometa: fix texture_type memcpy size

Coverity 11395891139588

10 years agosdpmessage: fix multi statement macros
Vincent Penquerc'h [Thu, 10 Apr 2014 10:19:26 +0000 (11:19 +0100)]
sdpmessage: fix multi statement macros

Wasn't playing nice with an if statement below.

Coverity 1139767

10 years agoaudiocdsrc: guard aginst overflow
Vincent Penquerc'h [Thu, 10 Apr 2014 10:14:25 +0000 (11:14 +0100)]
audiocdsrc: guard aginst overflow

An audio CD may contain about a tenth of the samples 32 bit can
represent, so it doesn't seem likely this will be hit in practice.

Coverity 1139805

10 years agopbutils: descriptions: default to systemstream=false for partial video/mpeg caps
Tim-Philipp Müller [Thu, 10 Apr 2014 11:30:50 +0000 (12:30 +0100)]
pbutils: descriptions: default to systemstream=false for partial video/mpeg caps

Assume systemstream=false for video/mpeg caps where that field
is missing.

10 years agoaudiobasesink: avoid possible sample count overflow
Vincent Penquerc'h [Thu, 10 Apr 2014 09:57:53 +0000 (10:57 +0100)]
audiobasesink: avoid possible sample count overflow

At 48 kHz, 2<<31 samples is reached before 13 hours so it
sounds plausible this would be hit.

Coverity 11398001139801

10 years agotheoraenc: fix comparison to unset timestamp
Vincent Penquerc'h [Thu, 10 Apr 2014 09:45:21 +0000 (10:45 +0100)]
theoraenc: fix comparison to unset timestamp

Also rejects negative timestamps that aren't GST_CLOCK_TIME_NONE.

Coverity 1139797

10 years agooggstream: fix a few left shifts operations on 32 bits cast to 64 bits
Vincent Penquerc'h [Thu, 10 Apr 2014 09:33:46 +0000 (10:33 +0100)]
oggstream: fix a few left shifts operations on 32 bits cast to 64 bits

This should not cause any actual bug since Theora and Daala have
a maximum shift of 31, and a packet duration of 2^31 seems very
implausible. But it fixes:

Coverity 113980411398031139802

10 years agooggstream: remove NULL test after dereference
Vincent Penquerc'h [Thu, 10 Apr 2014 09:29:34 +0000 (10:29 +0100)]
oggstream: remove NULL test after dereference

And add NULLness asserts at top of function. The only call
to this passes local variable pointers, so non NULL.

Coverity 206375

10 years agooggmux: test for failure to return tag
Vincent Penquerc'h [Thu, 10 Apr 2014 09:25:46 +0000 (10:25 +0100)]
oggmux: test for failure to return tag

It should really not happen unless the tag list it corrupt,
but the API returns a failure code so we may as well use it.

Coverity 1139595

10 years agooggdemux: do not dereference NULL pad in warning message
Vincent Penquerc'h [Thu, 10 Apr 2014 09:22:43 +0000 (10:22 +0100)]
oggdemux: do not dereference NULL pad in warning message

Coverity 1197695

10 years agovideo-event: Update the running times in the force-keyunit events from the pad offsets
Sebastian Dröge [Thu, 10 Apr 2014 07:18:05 +0000 (09:18 +0200)]
video-event: Update the running times in the force-keyunit events from the pad offsets

10 years agodecodebin: In adaptive streaming mode, only have a fixed buffer limit for the non...
Sebastian Dröge [Wed, 9 Apr 2014 14:03:15 +0000 (16:03 +0200)]
decodebin: In adaptive streaming mode, only have a fixed buffer limit for the non-buffering multiqueue

10 years agosdp: guard against address parse errors.
Wim Taymans [Tue, 8 Apr 2014 13:43:50 +0000 (15:43 +0200)]
sdp: guard against address parse errors.

10 years agoadder: rework the logic to check if eos has to be sent.
Mathieu Duponchelle [Tue, 25 Mar 2014 16:11:34 +0000 (17:11 +0100)]
adder: rework the logic to check if eos has to be sent.

Checking the size available was incorrect, and the infos
for per-pad EOS are available.

Same logic as audiomixer.

fixes: https://bugzilla.gnome.org/show_bug.cgi?id=727025

10 years agoaudioringbuffer: parse channels field from compressed audio caps
Josep Torra [Tue, 8 Apr 2014 10:46:21 +0000 (12:46 +0200)]
audioringbuffer: parse channels field from compressed audio caps

Also parse channels as an optional field in the caps for compressed
audio formats.

10 years agosubtitleoverlay: Consider all caps for overlays, not just the first.
Jan Schmidt [Sun, 6 Apr 2014 12:26:20 +0000 (22:26 +1000)]
subtitleoverlay: Consider all caps for overlays, not just the first.

Check all supported caps on the overlay video pad, not just the
first of (possibly) many.

10 years agotools: update gst-play-1.0 man page
Tim-Philipp Müller [Sat, 5 Apr 2014 12:25:46 +0000 (13:25 +0100)]
tools: update gst-play-1.0 man page

10 years agovideodecoder: do not deactivate the bufferpool, just unref
Thiago Santos [Wed, 2 Apr 2014 10:20:43 +0000 (07:20 -0300)]
videodecoder: do not deactivate the bufferpool, just unref

Videodecoder does late renegotiation, it will wait for the next
buffer before renegotiating its caps and bufferpool. It might happen
that downstream element switched from passthrough to non-passthrough
and sent a reconfigure upstream (that caused this renegotiation).
This downstream element will ask the video sink below for the bufferpool
with an allocation query and will get the same bufferpool that
videodecoder is holding, too.

When renegotiating, if videodecoder deactivates its bufferpool it
might be deactivating the bufferpool that some element downstream
is using and cause the pipeline to fail.

https://bugzilla.gnome.org/show_bug.cgi?id=727498

10 years agoaudiobasesink: clip start samples to match clipped start time
Vincent Penquerc'h [Mon, 24 Feb 2014 16:17:05 +0000 (11:17 -0500)]
audiobasesink: clip start samples to match clipped start time

Clock slaving can clip start time to zero, giving us a shorted
duration than we originally got. To keep in sync, we must then
discard the samples falling before that zero timestamp.

This possibly fixes random distortion caused by constant PA
underflows which are never resynced.

10 years agomikey: Fix the KEMAC payload
Wim Taymans [Fri, 4 Apr 2014 15:36:04 +0000 (17:36 +0200)]
mikey: Fix the KEMAC payload

The KEMAC payload actually needs to have subpayloads and the key should
go into the KEY_DATA subpayload. Add support for subpayloads and
implement the KEY_DATA payload.
Add some pointers to the conversion functions that allow us to add
encryption and decryption later.

10 years agoplaybin: Drop reference to any source element in NULL state
Jan Schmidt [Thu, 3 Apr 2014 15:14:50 +0000 (02:14 +1100)]
playbin: Drop reference to any source element in NULL state

Drop the reference instead of waiting for either finalize(), or
for a new source when reused. Everyone else already forgot about
the old source.

10 years agortspconnection: Added gst_rtsp_watch_set_flushing to list.
Göran Jönsson [Tue, 1 Apr 2014 08:38:23 +0000 (10:38 +0200)]
rtspconnection: Added gst_rtsp_watch_set_flushing to list.

Added gst_rtsp_watch_set_flushing to list in file
libgstrtsp.def

10 years agovideodecoder: Always drain the decoder after a discont group in reverse playback...
Sebastian Dröge [Sun, 30 Mar 2014 16:26:59 +0000 (18:26 +0200)]
videodecoder: Always drain the decoder after a discont group in reverse playback mode

10 years agovideodecoder: Flush the decoder once per discont group, not once per keyframe
Sebastian Dröge [Sun, 30 Mar 2014 15:54:11 +0000 (17:54 +0200)]
videodecoder: Flush the decoder once per discont group, not once per keyframe

10 years agovideodecoder: Handle reverse playback with multiple GOPs per discont group properly
Sebastian Dröge [Sun, 30 Mar 2014 15:54:11 +0000 (17:54 +0200)]
videodecoder: Handle reverse playback with multiple GOPs per discont group properly

baseparse will reverse each GOP for us already, so the segment events can
be after our keyframe. Make sure to get it and all other relevant sticky
events before starting to decode.

10 years agovideodecoder: Log event types of events that are pushed downstream
Sebastian Dröge [Sat, 29 Mar 2014 09:23:05 +0000 (10:23 +0100)]
videodecoder: Log event types of events that are pushed downstream

10 years agovideodecoder: In reverse playback mode we need to finish the subclass after passing...
Sebastian Dröge [Thu, 27 Mar 2014 19:15:01 +0000 (20:15 +0100)]
videodecoder: In reverse playback mode we need to finish the subclass after passing all frames to it

10 years agortspconnection: add flush method
Wim Taymans [Fri, 28 Mar 2014 08:32:20 +0000 (09:32 +0100)]
rtspconnection: add flush method

Add a method to set/unset the flushing state that makes _wait_backlog()
unlock.

See https://bugzilla.gnome.org/show_bug.cgi?id=725898

10 years agoximagesink: only extrapolate alpha mask for 32-bit depth
Nicolas Dufresne [Thu, 27 Mar 2014 20:43:10 +0000 (16:43 -0400)]
ximagesink: only extrapolate alpha mask for 32-bit depth

Instead of passing bogus alpha mask values when there's no alpha.

https://bugzilla.gnome.org/show_bug.cgi?id=727188

10 years agomikey: fix return values of g_return_*
Wim Taymans [Tue, 25 Mar 2014 10:14:51 +0000 (11:14 +0100)]
mikey: fix return values of g_return_*