platform/upstream/gstreamer.git
9 years agomatroska: read: fix tag list memory leak
Vineeth TM [Fri, 21 Aug 2015 04:19:02 +0000 (13:19 +0900)]
matroska: read: fix tag list memory leak

gst_toc_entry_merge_tags makes a new ref of the taglist, so it should
be unref'ed as soon as the tags are merged to the tocentry

https://bugzilla.gnome.org/show_bug.cgi?id=753904

9 years agowavpackdec: fix taglist memory leak
Vineeth TM [Fri, 21 Aug 2015 03:20:59 +0000 (12:20 +0900)]
wavpackdec: fix taglist memory leak

When passing the taglist to gst_audio_decoder_merge_tags, the reference is increased
by audiodecoder and the caller should free the taglist being passed.

https://bugzilla.gnome.org/show_bug.cgi?id=753903

9 years agov4l2transform: fix pad closing
Jean-Michel Hautbois [Thu, 20 Aug 2015 12:45:33 +0000 (14:45 +0200)]
v4l2transform: fix pad closing

Signed-off-by: Jean-Michel Hautbois <jean-michel.hautbois@veo-labs.com>
https://bugzilla.gnome.org/show_bug.cgi?id=753875

9 years agoRelease 1.5.90
Sebastian Dröge [Wed, 19 Aug 2015 10:29:53 +0000 (13:29 +0300)]
Release 1.5.90

9 years agoUpdate .po files
Sebastian Dröge [Wed, 19 Aug 2015 09:47:42 +0000 (12:47 +0300)]
Update .po files

9 years agopo: Update translations
Sebastian Dröge [Wed, 19 Aug 2015 08:29:55 +0000 (11:29 +0300)]
po: Update translations

9 years agomultifilesrc: fix regression with starting from index set via index property
Tim-Philipp Müller [Thu, 13 Aug 2015 16:29:58 +0000 (17:29 +0100)]
multifilesrc: fix regression with starting from index set via index property

When we haven't started yet, set the start_index when we set the index property,
so that we start at the right index position after the initial seek. The index
property was never really meant to be for writing, but it used to work, so let's
support it for backwards compatibility.

https://bugzilla.gnome.org/show_bug.cgi?id=739472

9 years agoqtdemux: fix offset calculation when parsing CENC aux info
Alex Ashley [Tue, 18 Aug 2015 09:52:11 +0000 (10:52 +0100)]
qtdemux: fix offset calculation when parsing CENC aux info

Commit 7d7e54ce6863ff53e188d0276d2651b65082ffdb added support for
DASH common encryption, however commit
bb336840c0b0b02fa18dc4437ce0ded3d9142801 that went onto master
shortly before the CENC commit caused the calculation of the CENC
aux info offset to be incorrect.

The base_offset was being added if present, but if the base_offset
is relative to the start of the moof, the offset was being added twice.
The correct approach is to calculate the offset from the start of the
moof and use that offset when parsing the CENC aux info.

9 years agoflacenc: actually return true for accept-caps query handling
Thiago Santos [Mon, 17 Aug 2015 17:28:24 +0000 (14:28 -0300)]
flacenc: actually return true for accept-caps query handling

9 years agortp: copy metadata in the (de)payloaders which is missed before
Hyunjun Ko [Mon, 17 Aug 2015 05:07:10 +0000 (14:07 +0900)]
rtp: copy metadata in the (de)payloaders which is missed before

https://bugzilla.gnome.org/show_bug.cgi?id=753706

9 years agodirectsoundsink: allow specifying audio playback device
Dustin Spicuzza [Sun, 16 Aug 2015 19:21:51 +0000 (15:21 -0400)]
directsoundsink: allow specifying audio playback device

https://bugzilla.gnome.org/show_bug.cgi?id=753670

9 years agoflacenc: remove single entry if from loop
Thiago Santos [Sun, 16 Aug 2015 16:51:47 +0000 (13:51 -0300)]
flacenc: remove single entry if from loop

Iterate from the 2nd channel on and create the 1 channel struct
outside to make loop structure simpler and only slightly faster.

9 years agoflacenc: implement proper accept-caps
Thiago Santos [Sun, 16 Aug 2015 16:21:41 +0000 (13:21 -0300)]
flacenc: implement proper accept-caps

Should just compare with what can be immediatelly accepted by
the element. flacenc can't renegotiate so if it has a caps already
it should only accept if it is that caps otherwise just use the
template caps

9 years agoflacenc: improve sink pad template caps
Thiago Santos [Sun, 16 Aug 2015 16:03:36 +0000 (13:03 -0300)]
flacenc: improve sink pad template caps

Removes the need for custom caps query handling and makes it more
correct from the beginning on the template. It is a bit uglier
to read because there is 1 entry per channel but makes code easier
to maintain.

9 years agoy4mencode: fix gst-launch version in documentation
Thiago Santos [Sun, 16 Aug 2015 15:41:56 +0000 (12:41 -0300)]
y4mencode: fix gst-launch version in documentation

9 years agoaudioencoders: use template subset check for accept-caps
Thiago Santos [Sun, 16 Aug 2015 01:32:21 +0000 (22:32 -0300)]
audioencoders: use template subset check for accept-caps

It is faster than doing a query that propagates downstream and
should be enough

Elements: speexenc, wavpackenc, mulawenc, alawenc

9 years agovideoencoders: use template subset check for accept-caps
Thiago Santos [Sun, 16 Aug 2015 01:29:41 +0000 (22:29 -0300)]
videoencoders: use template subset check for accept-caps

It is faster than doing a query that propagates downstream and
should be enough

Elements: jpegenc, pngenc, vp8enc, vp9enc, y4menc

9 years agompegaudioparse: use new baseparse API to fix tag handling
Tim-Philipp Müller [Sun, 16 Aug 2015 16:21:24 +0000 (17:21 +0100)]
mpegaudioparse: use new baseparse API to fix tag handling

https://bugzilla.gnome.org/show_bug.cgi?id=679768

9 years agoaudioparsers: use new base parse API to fix tag handling
Olivier Crête [Tue, 17 Mar 2015 21:50:37 +0000 (17:50 -0400)]
audioparsers: use new base parse API to fix tag handling

https://bugzilla.gnome.org/show_bug.cgi?id=679768

9 years agoflacparse: use new baseparse API and fix tag handling
Tim-Philipp Müller [Sun, 16 Aug 2015 13:37:53 +0000 (14:37 +0100)]
flacparse: use new baseparse API and fix tag handling

https://bugzilla.gnome.org/show_bug.cgi?id=679768

9 years agoqtdemux: Use signed integer type to be able to check for negative subtraction results
Sebastian Dröge [Sun, 16 Aug 2015 11:04:02 +0000 (13:04 +0200)]
qtdemux: Use signed integer type to be able to check for negative subtraction results

CID 1315829

9 years agortpvorbisdepay: remove dead code
Luis de Bethencourt [Sun, 16 Aug 2015 10:50:34 +0000 (11:50 +0100)]
rtpvorbisdepay: remove dead code

payload_buffer must be NULL in ignore_reserved. Check will always be false.

Introduced by b1089fb5207697ba26edb4ff66ed0f465c6df3cf

CID #1316476

9 years agoalawenc: port to AudioEncoder base class
Thiago Santos [Sun, 16 Aug 2015 01:45:53 +0000 (22:45 -0300)]
alawenc: port to AudioEncoder base class

9 years agoaudiodecoders: use default pad accept-caps handling
Thiago Santos [Sat, 15 Aug 2015 12:16:23 +0000 (09:16 -0300)]
audiodecoders: use default pad accept-caps handling

Avoids useless check of downstream caps when handling an
accept-caps query

Elements: flacdec, speexdec, wavpackdec, mulawdec, alawdec

9 years agovideodecoders: use default pad accept-caps handling
Thiago Santos [Sat, 15 Aug 2015 11:49:57 +0000 (08:49 -0300)]
videodecoders: use default pad accept-caps handling

Avoids useless check of downstream caps when handling an
accept-caps query

Elements: jpegdec, pngdec, vp8dec, vp9dec

9 years agoalawdec: make error handling a bit nicer
Thiago Santos [Sat, 15 Aug 2015 14:31:04 +0000 (11:31 -0300)]
alawdec: make error handling a bit nicer

Print the element along with the debug to make it easier to trace
the failures

9 years agoalawdec: port to audiodecoder base class
Thiago Santos [Sat, 15 Aug 2015 14:04:16 +0000 (11:04 -0300)]
alawdec: port to audiodecoder base class

mulawdec was already ported, alawdec was left behind.

9 years agoqtdemux: only look for more samples in moofs in pull-mode
Thiago Santos [Sat, 15 Aug 2015 13:34:14 +0000 (10:34 -0300)]
qtdemux: only look for more samples in moofs in pull-mode

For playback of some fragmented formats with qtdemux it will
try to look for the next moof after finishing one but it is only
possible for pull-mode. For playback of streaming fragmented formats
such as DASH it should just not try to look for another moof but
instead wait for more data.

https://bugzilla.gnome.org/show_bug.cgi?id=752602

https://bugzilla.gnome.org/show_bug.cgi?id=752603

9 years agodcaparse: Don't look for a second syncword
Sebastian Dröge [Sat, 15 Aug 2015 10:58:50 +0000 (12:58 +0200)]
dcaparse: Don't look for a second syncword

There are streams out there that consistently contain garbage between
every frame so we never ever find a second consecutive syncword.

See https://bugzilla.gnome.org/show_bug.cgi?id=738237

9 years agovp8enc, vp9enc: reset multipass file index when stopping encoder
Tim-Philipp Müller [Sat, 15 Aug 2015 10:12:05 +0000 (11:12 +0100)]
vp8enc, vp9enc: reset multipass file index when stopping encoder

Fixes multipass encoding when re-using the same element/pipeline
for subsequent encoding runs.

https://bugzilla.gnome.org/show_bug.cgi?id=747728

9 years agovp9enc: provide support for multiple pass cache files
Tim-Philipp Müller [Sat, 15 Aug 2015 10:09:42 +0000 (11:09 +0100)]
vp9enc: provide support for multiple pass cache files

Some files may provide different caps insight of one stream. Since
vp9enc support caps reinit, we should support cache reinit too.
If more then file cache file will be created, the naming will be:
cache cache.1 cache.2 ...

Based on patch by: Oleksij Rempel <linux@rempel-privat.de>

https://bugzilla.gnome.org/show_bug.cgi?id=747728

9 years agotests: aacparse: use caps query instead of accept-caps
Thiago Santos [Fri, 14 Aug 2015 14:41:42 +0000 (11:41 -0300)]
tests: aacparse: use caps query instead of accept-caps

The accept-caps query just does a shallow check at the current
element while at this test we want it to also look at downstream.
So use caps query there.

https://bugzilla.gnome.org/show_bug.cgi?id=753623

9 years agoaudioparsers: enable accept-template flag
Thiago Santos [Fri, 14 Aug 2015 14:40:22 +0000 (11:40 -0300)]
audioparsers: enable accept-template flag

Do a quick check with the pad template caps as it is enough. Users
should have figured the appropriate full caps on a previous caps query

https://bugzilla.gnome.org/show_bug.cgi?id=753623

9 years agortspsrc: send the User-Agent header
George Kiagiadakis [Fri, 14 Aug 2015 13:46:53 +0000 (15:46 +0200)]
rtspsrc: send the User-Agent header

Sometimes it is useful to know this information on the
server side. Other popular implementations (vlc, ffmpeg, ...)
also send this header on every message.

This includes a new "user-agent" property that the user
can set to use a custom User-Agent string. The default
is "GStreamer/<version>"

https://bugzilla.gnome.org/show_bug.cgi?id=750101

9 years agortspsrc: wrap gst_rtsp_message_init_request in a local function
George Kiagiadakis [Fri, 14 Aug 2015 13:42:42 +0000 (15:42 +0200)]
rtspsrc: wrap gst_rtsp_message_init_request in a local function

This will allow adding common request initialization, like the
user agent string, in just one place.

9 years agoaudioecho: make sure buffer gets reallocated if max_delay changes
Prashant Gotarne [Fri, 14 Aug 2015 04:06:09 +0000 (09:36 +0530)]
audioecho: make sure buffer gets reallocated if max_delay changes

https://bugzilla.gnome.org/show_bug.cgi?id=753490

9 years agovp8enc: provide support for multiple pass cache files
Oleksij Rempel [Thu, 9 Jul 2015 07:51:26 +0000 (09:51 +0200)]
vp8enc: provide support for multiple pass cache files

Some files may provide different caps insight of one stream. Since vp8enc
support caps reinit, we should support cache reinit too.
If more then file cache file will be created, the naming will be:
cache
cache.1
cache.2
...

https://bugzilla.gnome.org/show_bug.cgi?id=747728

9 years agortpmp4gdepay: fix timestamps for RTP packets with multiple AUs
Ramiro Polla [Wed, 15 Apr 2015 20:51:51 +0000 (22:51 +0200)]
rtpmp4gdepay: fix timestamps for RTP packets with multiple AUs

Use constantDuration to calculate the timestamp of non-first AU in the
RTP packet.

If constantDuration is not present in the MIME parameters, its value
must be calculated based on the timing information from two consecutive
RTP packets with AU-Index equal to 0.

https://bugzilla.gnome.org/show_bug.cgi?id=747881

9 years agosouphttpsrc: remove unnecessary if, g_free is null safe
Thiago Santos [Fri, 14 Aug 2015 09:43:13 +0000 (06:43 -0300)]
souphttpsrc: remove unnecessary if, g_free is null safe

9 years agosouphttpsrc: add property to set HTTP method
Alex Ashley [Fri, 14 Aug 2015 07:33:56 +0000 (08:33 +0100)]
souphttpsrc: add property to set HTTP method

To allow souphttpsrc to be use HTTP methods other than GET
(e.g. HEAD), add a "method" property that is a string. If this
property is not set, GET is used.

https://bugzilla.gnome.org/show_bug.cgi?id=752413

9 years agocheck: Rename states unit test
Edward Hervey [Fri, 14 Aug 2015 09:13:01 +0000 (11:13 +0200)]
check: Rename states unit test

Makes it easier to differentiate from other modules states unit test

9 years agogoom: Rename get_type() function of base class to prevent symbol conflicts
Sebastian Dröge [Fri, 14 Aug 2015 07:21:25 +0000 (09:21 +0200)]
goom: Rename get_type() function of base class to prevent symbol conflicts

This is a problem when statically linking.

9 years agortpjitterbuffer: Keep the DTS estimate if we got no DTS after a jitterbuffer reset
Sebastian Dröge [Thu, 13 Aug 2015 14:32:55 +0000 (16:32 +0200)]
rtpjitterbuffer: Keep the DTS estimate if we got no DTS after a jitterbuffer reset

Otherwise we will just output buffers without timestamps after a reset if no
timestamps are provided by upstream, e.g. when using RTSP over TCP.

https://bugzilla.gnome.org/show_bug.cgi?id=749536

9 years agomatroska: Remove unused variable
Ravi Kiran K N [Wed, 12 Aug 2015 11:46:01 +0000 (17:16 +0530)]
matroska: Remove unused variable

https://bugzilla.gnome.org/show_bug.cgi?id=753556

9 years agortp: Copy metadata in the (de)payloader, but only the relevant ones
Sebastian Dröge [Tue, 4 Aug 2015 17:59:17 +0000 (20:59 +0300)]
rtp: Copy metadata in the (de)payloader, but only the relevant ones

The payloader didn't copy anything so far, the depayloader copied every
possible meta. Let's make it consistent and just copy all metas without
tags or with only the video tag.

https://bugzilla.gnome.org/show_bug.cgi?id=751774

9 years agoqtdemux: fix small typo in comment
Thiago Santos [Mon, 10 Aug 2015 21:20:15 +0000 (18:20 -0300)]
qtdemux: fix small typo in comment

9 years agogoom2k1/doc: Fixup previous commit
Nicolas Dufresne [Mon, 10 Aug 2015 20:19:18 +0000 (16:19 -0400)]
goom2k1/doc: Fixup previous commit

9 years agogoom2k1/doc: Use GstGoom2k1 namespace
Nicolas Dufresne [Mon, 10 Aug 2015 19:55:19 +0000 (15:55 -0400)]
goom2k1/doc: Use GstGoom2k1 namespace

The doc generator isn't happy when we have class name clash. Simply
use it's own namespace.

9 years agoaudioecho: removed unused variable in set_property
Prashant Gotarne [Mon, 10 Aug 2015 11:40:42 +0000 (17:10 +0530)]
audioecho: removed unused variable in set_property

unused local variable 'delay' is removed.

https://bugzilla.gnome.org/show_bug.cgi?id=753450

9 years agoqtdemux: fix suboptimal queue iteration code
Tim-Philipp Müller [Mon, 10 Aug 2015 11:45:27 +0000 (12:45 +0100)]
qtdemux: fix suboptimal queue iteration code

9 years agoqtdemux: don't use glib 2.44-only API
Tim-Philipp Müller [Sun, 9 Aug 2015 16:25:45 +0000 (17:25 +0100)]
qtdemux: don't use glib 2.44-only API

9 years agoqtdemux: add support for ISOBMFF Common Encryption
Alex Ashley [Wed, 29 Jul 2015 13:14:50 +0000 (14:14 +0100)]
qtdemux: add support for ISOBMFF Common Encryption

This commit adds support for ISOBMFF Common Encryption (cenc), as
defined in ISO/IEC 23001-7. It uses a GstProtection event to
pass the contents of PSSH boxes to downstream decryptor elements
and attached GstProtectionMeta to each sample.

https://bugzilla.gnome.org/show_bug.cgi?id=705991

9 years agortph264depay: checking if depay has sps/pps nals before insertion
Hyunjun Ko [Mon, 10 Aug 2015 05:13:50 +0000 (14:13 +0900)]
rtph264depay: checking if depay has sps/pps nals before insertion

https://bugzilla.gnome.org/show_bug.cgi?id=753430

9 years agomatroskamux: fix outdated comment
Tim-Philipp Müller [Sat, 8 Aug 2015 15:44:49 +0000 (16:44 +0100)]
matroskamux: fix outdated comment

The default behaviour was changed in the 0.10 -> 1.x
transition, but the comment was not updated.

9 years agortptheorapay: If flushing a packet failed, go out of the loop immediately
Sebastian Dröge [Sat, 8 Aug 2015 15:42:22 +0000 (17:42 +0200)]
rtptheorapay: If flushing a packet failed, go out of the loop immediately

9 years agortpvorbispay: If flushing a packet failed, go out of the loop immediately
Sebastian Dröge [Sat, 8 Aug 2015 15:41:02 +0000 (17:41 +0200)]
rtpvorbispay: If flushing a packet failed, go out of the loop immediately

9 years agortptheorapay: Extract pixel format from the ident header to put it into the sampling...
Sebastian Dröge [Sat, 8 Aug 2015 15:34:50 +0000 (17:34 +0200)]
rtptheorapay: Extract pixel format from the ident header to put it into the sampling field of the caps

We always put 4:2:0 into the caps before, which obviously is wrong for 4:2:2
and 4:4:4 formats.

9 years agortpklv(de)pay: add "RTP" in the klass string
George Kiagiadakis [Thu, 6 Aug 2015 15:46:13 +0000 (17:46 +0200)]
rtpklv(de)pay: add "RTP" in the klass string

GstRTSPMedia uses this classification to detect the real payloader
inside a dynpay bin and asserts if it doesn't find it, therefore
it is required

https://bugzilla.gnome.org/show_bug.cgi?id=753325

9 years agotests: rtpaux: use a dynamic pt in the test
Thiago Santos [Wed, 5 Aug 2015 14:13:09 +0000 (11:13 -0300)]
tests: rtpaux: use a dynamic pt in the test

1) Tests that using dynamic PT instead of the default ones work
2) If we ever decide to change the codec here we don't need to
   worry about change the PT for the default one of the new codec
   in the test

https://bugzilla.gnome.org/show_bug.cgi?id=746445

9 years agortprtxsend: print valid type where guint32 is expected
Hyunjun Ko [Wed, 5 Aug 2015 01:53:15 +0000 (10:53 +0900)]
rtprtxsend: print valid type where guint32 is expected

https://bugzilla.gnome.org/show_bug.cgi?id=746445

9 years agortppayload: set standard payload type as default
Hyunjun Ko [Thu, 6 Aug 2015 02:33:37 +0000 (11:33 +0900)]
rtppayload: set standard payload type as default

Initialize the PT to the default value of the codec and check if
it is still the default before declaring the pt to be dynamic or
not when setting the caps.

Also use the PT constants from the rtp lib when possible

https://bugzilla.gnome.org/show_bug.cgi?id=747965

9 years agoqtdemux: store the moof-offset also for push mode
Thiago Santos [Sun, 26 Jul 2015 15:07:56 +0000 (12:07 -0300)]
qtdemux: store the moof-offset also for push mode

It will be used in some cases for getting the correct offsets
from trun atoms.

https://bugzilla.gnome.org/show_bug.cgi?id=752603

9 years agoqtdemux: handle default-base-is-moof flag
Thiago Santos [Sun, 26 Jul 2015 05:09:24 +0000 (02:09 -0300)]
qtdemux: handle default-base-is-moof flag

Handle the flag from the tfhd that signals the base offset to
start from the moof atom

https://bugzilla.gnome.org/show_bug.cgi?id=752603

9 years agomatroskademux: Preserve forward referenced track tags
Glen Diener [Thu, 30 Jul 2015 00:54:35 +0000 (18:54 -0600)]
matroskademux: Preserve forward referenced track tags

https://bugzilla.gnome.org/show_bug.cgi?id=752850

9 years agotests: rtpaux: fix test failure
Thiago Santos [Tue, 4 Aug 2015 21:07:35 +0000 (18:07 -0300)]
tests: rtpaux: fix test failure

The RTP PT for alaw is 8.
Less than 50 packets are received in the length of this test so it
would never drop a buffer or would drop only the last buffer and
it would fail sometimes when the received wouldn't receive the
retransmission packet in time.

https://bugzilla.gnome.org/show_bug.cgi?id=746445

9 years agortpstreamdepay: Only allow activation in push mode
Sebastian Dröge [Tue, 4 Aug 2015 17:59:17 +0000 (20:59 +0300)]
rtpstreamdepay: Only allow activation in push mode

We need a proper caps event from upstream with the full RTP caps as we can't
create caps ourselves from thin air. Fixes usage of rtpstreamdepay after e.g.
a filesrc or any other element that supports pull mode.

https://bugzilla.gnome.org/show_bug.cgi?id=753066

9 years agosoup: fix typo in translated string
Tim-Philipp Müller [Tue, 4 Aug 2015 15:28:17 +0000 (16:28 +0100)]
soup: fix typo in translated string

https://bugzilla.gnome.org/show_bug.cgi?id=753240

9 years agortph264depay: Put the profile and level into the caps
Sebastian Dröge [Tue, 4 Aug 2015 09:25:46 +0000 (12:25 +0300)]
rtph264depay: Put the profile and level into the caps

9 years agortph264depay: Only update the srcpad caps if something else than the codec_data changed
Sebastian Dröge [Tue, 4 Aug 2015 09:09:12 +0000 (12:09 +0300)]
rtph264depay: Only update the srcpad caps if something else than the codec_data changed

h264parse does the same, let's keep the behaviour consistent. As we now
include the codec_data inside the stream too here, this causes less caps
renegotiation.

9 years agortph264depay: PPS replaces and old PPS if it has the same id, independent of SPS id
Sebastian Dröge [Tue, 4 Aug 2015 08:48:27 +0000 (11:48 +0300)]
rtph264depay: PPS replaces and old PPS if it has the same id, independent of SPS id

The spec says:

When a picture parameter set NAL unit with a particular value of
pic_parameter_set_id is received, its content replaces the content of the
previous picture parameter set NAL unit, in decoding order, with the same
value of pic_parameter_set_id (when a previous picture parameter set NAL unit
with the same value of pic_parameter_set_id was present in the bitstream).

9 years agosplitmuxsink: remove extra \n at debug message
Thiago Santos [Mon, 3 Aug 2015 16:45:59 +0000 (13:45 -0300)]
splitmuxsink: remove extra \n at debug message

9 years agosplitmuxsink: prevent deadlock when states change too fast
Thiago Santos [Mon, 3 Aug 2015 16:42:20 +0000 (13:42 -0300)]
splitmuxsink: prevent deadlock when states change too fast

If the GOP is completed, pads have to start gathering for the
next one but it is possible that the the state might go to
COLLECTING_GOP_START and back to WAITING_GOP_COMPLETE before the
thread has a chance to wake up and proceed, leaving it trapped in
the check_completed_gop loop and deadlocking the other threads
waiting for it to advance.

To solve it, this patch also checks that tha input running time
hasn't changed to prevent this scenario.

9 years agortph264depay: Insert SPS/PPS NALs into the stream
Sebastian Dröge [Mon, 3 Aug 2015 14:55:01 +0000 (17:55 +0300)]
rtph264depay: Insert SPS/PPS NALs into the stream

h264parse does the same and this fixes decoding of some streams with 32 SPS
(or 256 PPS). It is allowed to have SPS ID 0 to 31 (or PPS ID 0 to 255), but
the field in the codec_data for the number of SPS or PPS is only 5 (or 8) bit.
As such, 32 SPS (or 256 PPS) are interpreted as 0 everywhere.

This looks like a mistake in the part of the spec about the codec_data.

9 years agosouphttpsrc: handle empty http proxy string
Eunhae Choi [Thu, 30 Jul 2015 02:29:27 +0000 (11:29 +0900)]
souphttpsrc: handle empty http proxy string

1) If the system http_proxy environment variable is not set
or set to an empty string, we must not set proxy to avoid
http connection error.

2) In case of proxy property setting, if user want to clear
the proxy setting, they should be able to set it to NULL or
an empty string again, so this is fixed too.

3) Check if the proxy string was parsed correctly.

https://bugzilla.gnome.org/show_bug.cgi?id=752866

9 years agodvdemux: remove unused variable
Ravi Kiran K N [Wed, 29 Jul 2015 10:16:20 +0000 (15:46 +0530)]
dvdemux: remove unused variable

Remove unused variable 'framecount' from dvdemux

https://bugzilla.gnome.org/show_bug.cgi?id=753008

9 years agortspsrc: assertion error due to wrong condition check
Vineeth TM [Thu, 30 Jul 2015 06:32:09 +0000 (15:32 +0900)]
rtspsrc: assertion error due to wrong condition check

In media to caps function, reserved_keys array is being used for variable i,
leading to GLib-CRITICAL **: g_ascii_strcasecmp: assertion 's1 != NULL' failed
changed it to variable j

https://bugzilla.gnome.org/show_bug.cgi?id=753009

9 years agortpmp4vdepay: rtpbuffer is being unref'ed twice
Vineeth TM [Thu, 30 Jul 2015 06:21:20 +0000 (15:21 +0900)]
rtpmp4vdepay: rtpbuffer is being unref'ed twice

process_rtp_packet doesn't transfer the rtp buffer to mp4v_process_depay
the refernce should not be removed here

https://bugzilla.gnome.org/show_bug.cgi?id=753042

9 years agortspsrc: Strip keys from the fmtp that we use internally in our caps
Sebastian Dröge [Wed, 29 Jul 2015 10:26:46 +0000 (11:26 +0100)]
rtspsrc: Strip keys from the fmtp that we use internally in our caps

Skip keys from the fmtp, which we already use ourselves for the
caps. Some software is adding random things like clock-rate into
the fmtp, and we would otherwise here set a string-typed clock-rate
in the caps... and thus fail to create valid RTP caps

https://bugzilla.gnome.org/show_bug.cgi?id=753009

9 years agosplitmuxsink: Support mpegtsmux as a muxer.
Jan Schmidt [Wed, 29 Jul 2015 09:28:33 +0000 (19:28 +1000)]
splitmuxsink: Support mpegtsmux as a muxer.

As a fallback, look for a pad template sink_%d on
the muxer when requesting pads, to support mpegtsmux

https://bugzilla.gnome.org/show_bug.cgi?id=752999

9 years agosplitmuxsrc: Use a separate lock to delay typefind.
Jan Schmidt [Wed, 24 Jun 2015 15:35:27 +0000 (01:35 +1000)]
splitmuxsrc: Use a separate lock to delay typefind.

Don't hold the main splitmux part lock over
the parent state change function, as it prevents
posting error messages that happen. Since the purpose
is to prevent typefinding from proceeding, use a
separate mutex just for that.

9 years agomatroska: fix memory leak
Vineeth TM [Wed, 29 Jul 2015 04:43:50 +0000 (13:43 +0900)]
matroska: fix memory leak

After adding to tag list, key_val is not being free'd
resulting in memory leak

https://bugzilla.gnome.org/show_bug.cgi?id=752992

9 years agoqtdemux: fix 16-bit PCM audio advertised with 'raw ' fourcc
Manasa Athreya [Mon, 27 Jul 2015 04:34:14 +0000 (13:34 +0900)]
qtdemux: fix 16-bit PCM audio advertised with 'raw ' fourcc

'NONE' and 'raw ' fourcc don't always contain U8 audio, it can
be more bits as well, in which case it's just like 'twos'.

https://bugzilla.gnome.org/show_bug.cgi?id=752613

9 years agov4l2: Allow framerate to be large then 100pfs
Dimitrios Katsaros [Fri, 24 Jul 2015 13:10:05 +0000 (15:10 +0200)]
v4l2: Allow framerate to be large then 100pfs

This limit was arbitrary. We still fixate near 100pfs for compatibility.

https://bugzilla.gnome.org/show_bug.cgi?id=752825

9 years agoavidemux: Stop without posting error on flushing
Olivier Crête [Sat, 25 Jul 2015 07:25:28 +0000 (03:25 -0400)]
avidemux: Stop without posting error on flushing

This could just be a normal pipeline shutdown.

9 years agov4l2bufferpool: set GST_BUFFER_COPY_FLAGS to copy flags also
Hyunjun Ko [Thu, 23 Jul 2015 06:00:08 +0000 (15:00 +0900)]
v4l2bufferpool: set GST_BUFFER_COPY_FLAGS to copy flags also

https://bugzilla.gnome.org/show_bug.cgi?id=752618

9 years agotests: add minmal matroskademux test for subtitle output
Tim-Philipp Müller [Thu, 16 Jul 2015 17:09:30 +0000 (18:09 +0100)]
tests: add minmal matroskademux test for subtitle output

Some of the subtitle chunks will have embedded
NUL-terminators (last three), some don't (first three),
some will have markup, some won't, some will be valid
UTF-8 (all but last), some won't (last stanza).

https://bugzilla.gnome.org/show_bug.cgi?id=752421

9 years agomatroskademux: fix for subtitle buffers with NUL terminators
Dimitrios Christidis [Thu, 16 Jul 2015 15:49:26 +0000 (18:49 +0300)]
matroskademux: fix for subtitle buffers with NUL terminators

Commit 45892ec8 created a regression where g_utf8_validate() would fail
if the subtitle buffer had a NUL terminator as part of the data.

https://bugzilla.gnome.org/show_bug.cgi?id=752421

9 years agortpvp8depay: Check available bytes before copy
Stian Selnes [Tue, 21 Jul 2015 11:31:05 +0000 (13:31 +0200)]
rtpvp8depay: Check available bytes before copy

Need to check that the number of bytes we want to copy from the adapter
actually is available and handle the error case gracefully. This error
may happen if malformed packets are received and we don't have a
complete frame.

https://bugzilla.gnome.org/show_bug.cgi?id=752663

9 years agoqtdemux: Support subtitle when track subtype is fourcc_subt
Paul Hyunil [Thu, 16 Jul 2015 00:32:36 +0000 (09:32 +0900)]
qtdemux: Support subtitle when track subtype is fourcc_subt

https://bugzilla.gnome.org/show_bug.cgi?id=752655

9 years agov4l2bufferpool: Set timestamp when queue buffer.
Song Bing [Mon, 20 Jul 2015 08:59:40 +0000 (16:59 +0800)]
v4l2bufferpool: Set timestamp when queue buffer.

Should set timestamp when queue buffer.

https://bugzilla.gnome.org/show_bug.cgi?id=752618

9 years agortpmux: handle different ssrc's on sinkpads
Havard Graff [Thu, 16 Jul 2015 13:12:17 +0000 (15:12 +0200)]
rtpmux: handle different ssrc's on sinkpads

Do this by not putting the ssrc from the src pads in the caps used to
probe other sinkpads, and then  intersecting with it later.

https://bugzilla.gnome.org/show_bug.cgi?id=752491

9 years agoUpdate mailing list address from sourceforge to freedesktop
Tim-Philipp Müller [Thu, 16 Jul 2015 16:19:03 +0000 (17:19 +0100)]
Update mailing list address from sourceforge to freedesktop

9 years agomatroskademux: fix trailing '*' displayed with some text subtitles
Dimitrios Christidis [Wed, 15 Jul 2015 10:44:52 +0000 (13:44 +0300)]
matroskademux: fix trailing '*' displayed with some text subtitles

The subtitle buffer we push out should not include a NUL terminator
as part of the data, we just add such a terminator for safety, but
it should not be included in the buffer size.

A NUL terminator is not valid UTF-8, so checks will fail if it's
included in the size, and the NUL will be replaced by the fallback
character specified when converting, i.e. '*'.

https://bugzilla.gnome.org/show_bug.cgi?id=752421

9 years agopulse: add properties to GstDevice
Wim Taymans [Wed, 15 Jul 2015 16:23:05 +0000 (18:23 +0200)]
pulse: add properties to GstDevice

Add the extra properties we get from pulse to the GstDevice we expose
with the device monitor

9 years agoaudiofx: Fix typo in example pipelines
Ravi Kiran K N [Wed, 15 Jul 2015 11:50:20 +0000 (17:20 +0530)]
audiofx: Fix typo in example pipelines

Fix typo in example pipelines of audiowsincband and audioinvert.

https://bugzilla.gnome.org/show_bug.cgi?id=752416

9 years agosplitmuxsink: add a "format-location" signal that allows better control over filenames
George Kiagiadakis [Wed, 15 Apr 2015 16:27:04 +0000 (18:27 +0200)]
splitmuxsink: add a "format-location" signal that allows better control over filenames

In certain applications, splitting into files named after a base
location template and an incremental sequence number is not enough.

This signal gives more fine-grained control to the application to
decide how to name the files.

https://bugzilla.gnome.org/show_bug.cgi?id=750106

9 years agoosxaudiosrc: no resampling on OS X
Ilya Konstantinov [Wed, 15 Apr 2015 17:13:27 +0000 (20:13 +0300)]
osxaudiosrc: no resampling on OS X

Unlike Remote IO, AUHAL doesn't have built-in resampling
for sources -- confirmed by Core Audio engineer Doug Wyatt:
http://lists.apple.com/archives/coreaudio-api/2006/Sep/msg00088.html

https://bugzilla.gnome.org/show_bug.cgi?id=743758

9 years agoosxaudiosrc: avoid get_channel_layout
Ilya Konstantinov [Wed, 15 Apr 2015 15:29:14 +0000 (18:29 +0300)]
osxaudiosrc: avoid get_channel_layout

This only produces a warning and serves no purpose.

https://bugzilla.gnome.org/show_bug.cgi?id=743758

9 years agoosxaudio: Avoid making a duplicate structure in caps for mono/stereo case
Arun Raghavan [Tue, 7 Apr 2015 10:10:14 +0000 (15:40 +0530)]
osxaudio: Avoid making a duplicate structure in caps for mono/stereo case

For 1ch or 2ch devices, we just need to set the caps to allow both
options since CoreAudio will up/downmix appropriately.

Also fixes the condition for the 2ch case to be exact, rather than at
least 2 channels since the downmix will not take place in the >stereo
case.

9 years agoosxaudio: Don't set the format on an initialized AudioUnit
Arun Raghavan [Mon, 6 Apr 2015 10:52:34 +0000 (16:22 +0530)]
osxaudio: Don't set the format on an initialized AudioUnit

We need to initialize the AudioUnit early to be able to probe the
underlying device, but according to the AudioUnitInitialize() and
AudioUnitUninitialize() documentation, format changes should be done
while the AudioUnit is uninitialized. So we explicitly uninitialize the
AudioUnit during a format change and reinitialize it when we're done.