platform/upstream/gstreamer.git
13 years agoalsasrc: Fix some compilation errors
Sebastian Dröge [Mon, 9 May 2011 10:56:14 +0000 (12:56 +0200)]
alsasrc: Fix some compilation errors

13 years agoalsasrc: Improve timestamp accuracy
Pontus Oldberg [Mon, 9 May 2011 09:50:05 +0000 (11:50 +0200)]
alsasrc: Improve timestamp accuracy

Fixes bug #635256.

13 years agosubtitleoverlay: Use new, public ghostpad functions
Sebastian Dröge [Fri, 6 May 2011 15:01:53 +0000 (17:01 +0200)]
subtitleoverlay: Use new, public ghostpad functions

13 years agoxvimagesink: Use GST_BOILERPLATE
Sreerenj Balachandran [Tue, 3 May 2011 08:26:32 +0000 (11:26 +0300)]
xvimagesink: Use GST_BOILERPLATE

13 years agoBack to development
Tim-Philipp Müller [Sat, 14 May 2011 08:41:58 +0000 (09:41 +0100)]
Back to development

13 years agoRelease 0.10.34
Tim-Philipp Müller [Sat, 14 May 2011 00:00:38 +0000 (01:00 +0100)]
Release 0.10.34

13 years agoRelease 0.10.33
Tim-Philipp Müller [Tue, 10 May 2011 08:32:11 +0000 (09:32 +0100)]
Release 0.10.33

Highlights:

  - support for 16-bit-per-component video formats
  - playbin2 fixes and improvements for custom and non-raw sinks
  - oggmux muxes based on running time now
  - many other fixes and improvements

13 years ago0.10.32.4 pre-release
Tim-Philipp Müller [Sat, 30 Apr 2011 16:35:54 +0000 (17:35 +0100)]
0.10.32.4 pre-release

13 years agogst: update orc-generated disted C backup code to orc 0.4.14
Tim-Philipp Müller [Sat, 30 Apr 2011 16:21:28 +0000 (17:21 +0100)]
gst: update orc-generated disted C backup code to orc 0.4.14

13 years ago0.10.32.3 pre-release
Tim-Philipp Müller [Wed, 27 Apr 2011 11:09:33 +0000 (12:09 +0100)]
0.10.32.3 pre-release

13 years agovideoscale: Fix off-by-one error in previous commit
David Schleef [Mon, 25 Apr 2011 01:45:40 +0000 (18:45 -0700)]
videoscale: Fix off-by-one error in previous commit

Fix for 7c0b702e.  It helps to get your j+1's right.

13 years agovideoscale: Fix ARGB bilinear scaling
David Schleef [Mon, 25 Apr 2011 01:16:20 +0000 (18:16 -0700)]
videoscale: Fix ARGB bilinear scaling

Fixes #648548.  Orc generates bad code for
gst_videoscale_orc_resample_merge_bilinear_u32, so we'll use the
slightly slower two-stage process.  I'd fix Orc, but it's hard to
get excited about fixing a feature that I'm planning to deprecate
and replace.

13 years agovideoscale: hack to fix invalid reads in linear
David Schleef [Sat, 23 Apr 2011 20:42:23 +0000 (13:42 -0700)]
videoscale: hack to fix invalid reads in linear

https://bugzilla.gnome.org/show_bug.cgi?id=633837

13 years agovideoscale: protect 4tap from out-of-bounds reads
David Schleef [Sat, 23 Apr 2011 19:46:09 +0000 (12:46 -0700)]
videoscale: protect 4tap from out-of-bounds reads

https://bugzilla.gnome.org/show_bug.cgi?id=633837

13 years agoAutomatic update of common submodule
Tim-Philipp Müller [Sun, 24 Apr 2011 13:03:12 +0000 (14:03 +0100)]
Automatic update of common submodule

From c3cafe1 to 46dfcea

13 years agovideoscale: use simpler scaling method for small images
David Schleef [Sat, 23 Apr 2011 19:44:50 +0000 (12:44 -0700)]
videoscale: use simpler scaling method for small images

https://bugzilla.gnome.org/show_bug.cgi?id=633837

13 years agoaudioresample: fix unused-but-set-variable warnings with gcc 4.6
Marc Plano-Lesay [Thu, 14 Apr 2011 07:32:19 +0000 (09:32 +0200)]
audioresample: fix unused-but-set-variable warnings with gcc 4.6

https://bugzilla.gnome.org/show_bug.cgi?id=647294

13 years agotag: exif: register common tags from tag library
Víctor Manuel Jáquez Leal [Fri, 22 Apr 2011 11:55:20 +0000 (13:55 +0200)]
tag: exif: register common tags from tag library

Exif uses tags like  image-vertical-ppi or image-horizontal-ppi which are
registered in gst_tag_register_musicbrainz_tags(), but neither GstExifReader
nor GstExifWriter register them.

https://bugzilla.gnome.org/show_bug.cgi?id=648459

13 years agotag: update some FIXMEs for 0.11
Tim-Philipp Müller [Sun, 24 Apr 2011 11:16:47 +0000 (12:16 +0100)]
tag: update some FIXMEs for 0.11

13 years agotests: add unit test for basetransform/videoscale negotiation regression
Tim-Philipp Müller [Thu, 21 Apr 2011 13:11:49 +0000 (14:11 +0100)]
tests: add unit test for basetransform/videoscale negotiation regression

Turn Rene's test pipeline into a unit test.

https://bugzilla.gnome.org/show_bug.cgi?id=648220

13 years agoringbuffer: make sure to not start if the may_start flag is FALSE
Håvard Graff [Thu, 25 Nov 2010 16:01:53 +0000 (17:01 +0100)]
ringbuffer: make sure to not start if the may_start flag is FALSE

Fixes #635784

13 years agox(v)imagesink: If NULL caps are passed to buffer_alloc() do fallback allocation
Sebastian Dröge [Mon, 18 Apr 2011 09:24:57 +0000 (11:24 +0200)]
x(v)imagesink: If NULL caps are passed to buffer_alloc() do fallback allocation

Fixes bug #647857.

13 years agooggmux: Remove bus GSource to prevent a valgrind warning
Sebastian Dröge [Mon, 18 Apr 2011 08:19:52 +0000 (10:19 +0200)]
oggmux: Remove bus GSource to prevent a valgrind warning

13 years agogio: Remove the bus GSource from the main context
Sebastian Dröge [Mon, 18 Apr 2011 07:16:35 +0000 (09:16 +0200)]
gio: Remove the bus GSource from the main context

Prevents a valgrind warning about possibly leaked memory,
see bug #647763.

13 years agosdp: remove gst_init() for g-i scanner here again as well to avoid problems with...
Tim-Philipp Müller [Sun, 17 Apr 2011 18:33:04 +0000 (19:33 +0100)]
sdp: remove gst_init() for g-i scanner here again as well to avoid problems with -Wl,--as-needed

13 years agofft: remove gst_init() for g-i scanner again
Tim-Philipp Müller [Sun, 17 Apr 2011 16:59:40 +0000 (17:59 +0100)]
fft: remove gst_init() for g-i scanner again

libgstfft doesn't actually use any symbols from libgstreamer, so when
compiling with -Wl,--as-needed it won't even link to it, which can
cause failures with older versions of g-i that ignore the --pkg
arguments.

Should fix PPA build failure on Ubuntu Maverick

13 years agotextoverlay: Always hold the class-global pango mutex when using pango API
Sebastian Dröge [Sat, 16 Apr 2011 14:31:57 +0000 (16:31 +0200)]
textoverlay: Always hold the class-global pango mutex when using pango API

13 years ago{time,clock}overlay: Hold the class-global pango mutex when changing the pango context
Sebastian Dröge [Sat, 16 Apr 2011 14:23:47 +0000 (16:23 +0200)]
{time,clock}overlay: Hold the class-global pango mutex when changing the pango context

13 years ago{clock,time}overlay: Only set the global pango context options once in class_init
Sebastian Dröge [Sat, 16 Apr 2011 14:21:39 +0000 (16:21 +0200)]
{clock,time}overlay: Only set the global pango context options once in class_init

Instead of doing it over and over again when instantiating a new instance.

13 years agopango: Create a new pango context for every subclass
Sebastian Dröge [Sat, 16 Apr 2011 14:18:40 +0000 (16:18 +0200)]
pango: Create a new pango context for every subclass

timeoverlay/clockoverlay are setting some global options
on the context that shouldn't be used for the generic textoverlay.

13 years ago0.10.32.2 pre-release
Tim-Philipp Müller [Sat, 16 Apr 2011 15:03:56 +0000 (16:03 +0100)]
0.10.32.2 pre-release

13 years agogst: update disted orc backup code
Tim-Philipp Müller [Sat, 16 Apr 2011 14:58:21 +0000 (15:58 +0100)]
gst: update disted orc backup code

13 years agodocs: update documentation
Tim-Philipp Müller [Sat, 16 Apr 2011 14:50:05 +0000 (15:50 +0100)]
docs: update documentation

13 years agopo: update translations
Tim-Philipp Müller [Sat, 16 Apr 2011 14:42:04 +0000 (15:42 +0100)]
po: update translations

13 years agovorbis: add support for using tremolo on android
Thibault Saunier [Thu, 31 Mar 2011 17:56:00 +0000 (17:56 +0000)]
vorbis: add support for using tremolo on android

Tremolo is an ARM-optimised version of xiph's tremor library.

13 years agooggmux: prefer headers from caps to determine stream type
Vincent Penquerc'h [Fri, 15 Apr 2011 12:36:39 +0000 (13:36 +0100)]
oggmux: prefer headers from caps to determine stream type

Ogg mandates the first header packet must determine a stream's type.
However, some streams (such as VP8) do not include such a header
when muxed in other containers, and thus do not include this header
as a buffer, but only in caps. We thus use headers from caps when
available to determine a new stream's type.

https://bugzilla.gnome.org/show_bug.cgi?id=647856

13 years agolibs: gobject-introspection scanner doesn't need to scan or update plugin info
Tim-Philipp Müller [Sat, 16 Apr 2011 10:00:31 +0000 (11:00 +0100)]
libs: gobject-introspection scanner doesn't need to scan or update plugin info

Make sure the scanner doesn't load or introspect or check any plugins,
(especially not outside the build directory).

13 years agotests: list libs/struct*h files explicitly in Makefile.am
Tim-Philipp Müller [Fri, 15 Apr 2011 20:09:00 +0000 (21:09 +0100)]
tests: list libs/struct*h files explicitly in Makefile.am

Hopefully makes the gentoo buildbot happy again.

13 years agoplaybin2: avoid foregoing READY_TO_NULL when appropriate
Mark Nauwelaerts [Fri, 15 Apr 2011 09:11:52 +0000 (11:11 +0200)]
playbin2: avoid foregoing READY_TO_NULL when appropriate

13 years agoplaybin2: ensure proper PAUSED_TO_READY cleanup
Mark Nauwelaerts [Thu, 14 Apr 2011 20:13:21 +0000 (22:13 +0200)]
playbin2: ensure proper PAUSED_TO_READY cleanup

... since going async to PAUSED might fail, and never making it to PAUSED
subsequently skips going down to READY.

Fixes #647781.

13 years agoRevert "video: Remove the extensive checkings from switch"
David Schleef [Thu, 14 Apr 2011 19:42:20 +0000 (12:42 -0700)]
Revert "video: Remove the extensive checkings from switch"

This reverts commit 500d14c35c656890686574e1c041fb556df17056.

13 years agoencodebin: Unref encoding profiles after usage in the test
Sebastian Dröge [Thu, 14 Apr 2011 11:15:08 +0000 (13:15 +0200)]
encodebin: Unref encoding profiles after usage in the test

13 years agoencodebin: Release pads after setting the state to NULL in the unit test
Sebastian Dröge [Thu, 14 Apr 2011 10:55:00 +0000 (12:55 +0200)]
encodebin: Release pads after setting the state to NULL in the unit test

See bug #647756.

13 years agoencodebin: Set all elements to NULL and remove them from the bin when removing a...
Sebastian Dröge [Thu, 14 Apr 2011 10:23:10 +0000 (12:23 +0200)]
encodebin: Set all elements to NULL and remove them from the bin when removing a source group

13 years agovideo: Remove the extensive checkings from switch
Sreerenj Balachandran [Wed, 13 Apr 2011 21:26:34 +0000 (00:26 +0300)]
video: Remove the extensive checkings from switch

The default case handles them already

13 years agotests: tag: Fix typo
Thiago Santos [Thu, 14 Apr 2011 02:17:34 +0000 (23:17 -0300)]
tests: tag: Fix typo

13 years agotag: xmp: Adds mapping for GST_TAG_CAPTURING_EXPOSURE_COMPENSATION
Thiago Santos [Thu, 14 Apr 2011 02:17:14 +0000 (23:17 -0300)]
tag: xmp: Adds mapping for GST_TAG_CAPTURING_EXPOSURE_COMPENSATION

Adds mapping for GST_TAG_CAPTURING_EXPOSURE_COMPENSATION for xmp
library.

Includes unit tests.

13 years agotag: exif: Adds mapping for GST_TAG_CAPTURING_EXPOSURE_COMPENSATION
Thiago Santos [Thu, 14 Apr 2011 02:16:02 +0000 (23:16 -0300)]
tag: exif: Adds mapping for GST_TAG_CAPTURING_EXPOSURE_COMPENSATION

Adds mapping for GST_TAG_CAPTURING_EXPOSURE_COMPENSATION for exif
library.

Includes unit tests.

13 years agotag: Adds GST_TAG_CAPTURING_EXPOSURE_COMPENSATION
Thiago Santos [Thu, 14 Apr 2011 02:13:59 +0000 (23:13 -0300)]
tag: Adds GST_TAG_CAPTURING_EXPOSURE_COMPENSATION

Adds a new tag for indicating the used exposure compensation
level in EV used when capturing an image.

API: GST_TAG_CAPTURING_EXPOSURE_COMPENSATION

13 years agotests: fix unused-but-set-variable warnings with gcc 4.6
Tim-Philipp Müller [Wed, 13 Apr 2011 23:24:26 +0000 (00:24 +0100)]
tests: fix unused-but-set-variable warnings with gcc 4.6

https://bugzilla.gnome.org/show_bug.cgi?id=647294

13 years agoogg: fix unused-but-set-variable warnings with gcc 4.6
Tim-Philipp Müller [Wed, 13 Apr 2011 22:57:56 +0000 (23:57 +0100)]
ogg: fix unused-but-set-variable warnings with gcc 4.6

https://bugzilla.gnome.org/show_bug.cgi?id=647294

13 years agoffmpegcolorspace: fix unused-but-set-variable warnings with gcc 4.6
Tim-Philipp Müller [Wed, 13 Apr 2011 22:19:07 +0000 (23:19 +0100)]
ffmpegcolorspace: fix unused-but-set-variable warnings with gcc 4.6

https://bugzilla.gnome.org/show_bug.cgi?id=647294

13 years agotypefindfunctions: fix unused-but-set-variable warning with gcc 4.6
Tim-Philipp Müller [Wed, 13 Apr 2011 21:59:03 +0000 (22:59 +0100)]
typefindfunctions: fix unused-but-set-variable warning with gcc 4.6

We don't compare the bitrates of consecutive mp3 frames on purpose
here.

https://bugzilla.gnome.org/show_bug.cgi?id=647294

13 years agodocs: fix typo in video format docs
Tim-Philipp Müller [Wed, 13 Apr 2011 08:10:52 +0000 (09:10 +0100)]
docs: fix typo in video format docs

13 years agooggmux: fix uninitialised variable usage and element leak
Tim-Philipp Müller [Tue, 12 Apr 2011 11:41:06 +0000 (12:41 +0100)]
oggmux: fix uninitialised variable usage and element leak

gcc on OSX complains about ret being used uninitialized in
this function, and it is right. Don't leak element ref
when returning early because newsegment event is not in
TIME format.

13 years agomultifdsink: do check return values of fcntl() and fstat()
Tim-Philipp Müller [Tue, 12 Apr 2011 11:20:43 +0000 (12:20 +0100)]
multifdsink: do check return values of fcntl() and fstat()

https://bugzilla.gnome.org/show_bug.cgi?id=647294

13 years agofix unused-but-set-variable warnings with gcc 4.6
Marc Plano-Lesay [Sat, 9 Apr 2011 17:15:23 +0000 (19:15 +0200)]
fix unused-but-set-variable warnings with gcc 4.6

https://bugzilla.gnome.org/show_bug.cgi?id=647294

13 years agortsptranport: ensure valid int result when parsing ranges
Sreerenj Balachandran [Wed, 6 Apr 2011 19:57:41 +0000 (22:57 +0300)]
rtsptranport: ensure valid int result when parsing ranges

Specifically, make sure that the return value of strtol is falling in
between the range of G_MININT and G_MAXINT.

Fixes #646952.

13 years agoencoding-profile: fix unused-but-set-variable warnings with gcc 4.6
Bastien Nocera [Wed, 6 Apr 2011 15:27:54 +0000 (16:27 +0100)]
encoding-profile: fix unused-but-set-variable warnings with gcc 4.6

Top-level profiles don't have restrictions, only stream profiles,
so no need to serialise that here.

https://bugzilla.gnome.org/show_bug.cgi?id=646925

13 years agotests: dist all struct_*.h files for libs ABI test
Tim-Philipp Müller [Mon, 11 Apr 2011 13:29:35 +0000 (14:29 +0100)]
tests: dist all struct_*.h files for libs ABI test

Should fix distcheck on x86_64.

13 years agovideorate: empty caps have no structure to pick
Mark Nauwelaerts [Mon, 11 Apr 2011 13:02:38 +0000 (15:02 +0200)]
videorate: empty caps have no structure to pick

13 years agotag: fix typo in ID3 genres
Tim-Philipp Müller [Mon, 11 Apr 2011 09:06:53 +0000 (10:06 +0100)]
tag: fix typo in ID3 genres

psychadelic -> psychedelic. Spotted by Sébastien Wilmet.

https://bugzilla.gnome.org/show_bug.cgi?id=647399

13 years agoandroid: make it ready for androgenizer
Alessandro Decina [Thu, 27 Jan 2011 16:28:51 +0000 (17:28 +0100)]
android: make it ready for androgenizer

Remove the android/ top dir
Fixe the Makefile.am to be androgenized

To build gstreamer for android we are now using androgenizer which generates the
needed Android.mk files.
Androgenizer can be found here:
http://git.collabora.co.uk/?p=user/derek/androgenizer.git

13 years agoAdd new header file to spec file
Christian Fredrik Kalager Schaller [Sat, 9 Apr 2011 01:01:08 +0000 (02:01 +0100)]
Add new header file to spec file

13 years agortp: Unref events if the parent element disappeared or has no event handler implemented
Sebastian Dröge [Fri, 8 Apr 2011 13:10:02 +0000 (15:10 +0200)]
rtp: Unref events if the parent element disappeared or has no event handler implemented

13 years agortp: fix pad callbacks so they handle when parent goes away
Ole André Vadla Ravnås [Thu, 6 Jan 2011 17:20:58 +0000 (18:20 +0100)]
rtp: fix pad callbacks so they handle when parent goes away

1) We need to lock and get a strong ref to the parent, if still there.
2) If it has gone away, we need to handle that gracefully.

This is necessary in order to safely modify a running pipeline. Has been
observed when a streaming thread is doing a buffer_alloc() while an
application thread sends an event on a pad further downstream, and from
within a pad probe (holding STREAM_LOCK) carries out the pipeline plumbing
while the streaming thread has its buffer_alloc() in progress.

13 years agoaudioresample: Make src query MT-safe
Havard Graff [Sun, 20 Mar 2011 07:59:33 +0000 (08:59 +0100)]
audioresample: Make src query MT-safe

It is possible that the element might be going down while the event arrives

13 years agovorbisdec: Unref events if the parent element disappeared
Sebastian Dröge [Fri, 8 Apr 2011 13:00:58 +0000 (15:00 +0200)]
vorbisdec: Unref events if the parent element disappeared

13 years agovorbisdec: make upstream queries and events MT-safe
Havard Graff [Mon, 21 Mar 2011 15:03:16 +0000 (16:03 +0100)]
vorbisdec: make upstream queries and events MT-safe

13 years agortp: Remove unused variables
Bastien Nocera [Wed, 6 Apr 2011 15:25:37 +0000 (16:25 +0100)]
rtp: Remove unused variables

https://bugzilla.gnome.org/show_bug.cgi?id=646924

13 years agovideo: Fix creation of grayscale caps
Sebastian Dröge [Thu, 7 Apr 2011 08:06:53 +0000 (10:06 +0200)]
video: Fix creation of grayscale caps

The endianness was not set correctly before.

Fixes bug #646923.

13 years agodocs: Update interlaced video design document
Robert Swain [Wed, 6 Apr 2011 14:11:02 +0000 (16:11 +0200)]
docs: Update interlaced video design document

The RFF flag is to be reused for buffers in the telecine state to
indicate that the buffer contains only unneeded repeated fields that are
present in other buffers and as such this buffer can be dropped.

13 years agotheoraenc: refactor multipass file writing
Mark Nauwelaerts [Fri, 25 Mar 2011 15:59:51 +0000 (16:59 +0100)]
theoraenc: refactor multipass file writing

13 years agoaudioresample: minor simplification
Mark Nauwelaerts [Tue, 8 Feb 2011 13:02:20 +0000 (14:02 +0100)]
audioresample: minor simplification

... which avoids crashing in the off-chance that structure == NULL.

13 years agotests: add basic unit tests for discoverer
Stefan Kost [Tue, 5 Apr 2011 15:14:49 +0000 (18:14 +0300)]
tests: add basic unit tests for discoverer

13 years agortcpbuffer: Round to next 32bit word, not current 32bit word at end of SDES chunk
Pascal Buhler [Tue, 24 Aug 2010 11:14:33 +0000 (13:14 +0200)]
rtcpbuffer: Round to next 32bit word, not current 32bit word at end of SDES chunk

13 years agoxvimagesink: don't paint the window black when going to NULL
Stefan Kost [Tue, 5 Apr 2011 08:32:52 +0000 (11:32 +0300)]
xvimagesink: don't paint the window black when going to NULL

Leave dealing with the appearance of the window when we are not playing to the
applications. We anyway want to go to NULL as quickly as possible.

Fixes #635800

13 years agovideo: Fix YUV9 and YVU9 again
David Schleef [Mon, 4 Apr 2011 23:00:30 +0000 (16:00 -0700)]
video: Fix YUV9 and YVU9 again

13 years agotag: fix compiler warning on OSX
Tim-Philipp Müller [Mon, 4 Apr 2011 22:41:16 +0000 (23:41 +0100)]
tag: fix compiler warning on OSX

gstvorbistag.c: In function 'gst_tag_list_from_vorbiscomment_buffer':
gstvorbistag.c:371: warning: 'data' may be used uninitialized in this function

13 years agotests: ignore xmpwriter unit test binary
Tim-Philipp Müller [Mon, 4 Apr 2011 22:23:37 +0000 (23:23 +0100)]
tests: ignore xmpwriter unit test binary

13 years agotag: use gst/math-compat.h header.
Haakon Sporsheim [Mon, 4 Apr 2011 15:21:45 +0000 (17:21 +0200)]
tag: use gst/math-compat.h header.

https://bugzilla.gnome.org/show_bug.cgi?id=646744

13 years agotag: Remove constness to silence MS compiler.
Haakon Sporsheim [Mon, 4 Apr 2011 15:23:53 +0000 (17:23 +0200)]
tag: Remove constness to silence MS compiler.

https://bugzilla.gnome.org/show_bug.cgi?id=646744

13 years agotag: Explicit cast to GThreadFunc to silence MS compiler.
Haakon Sporsheim [Mon, 4 Apr 2011 15:23:13 +0000 (17:23 +0200)]
tag: Explicit cast to GThreadFunc to silence MS compiler.

https://bugzilla.gnome.org/show_bug.cgi?id=646744

13 years agoAutomatic update of common submodule
Stefan Kost [Mon, 4 Apr 2011 12:56:50 +0000 (15:56 +0300)]
Automatic update of common submodule

From 1ccbe09 to c3cafe1

13 years agortcpbuffer: fix invalid read in validation of padding in rtcp packet
Trond Andersen [Fri, 11 Mar 2011 09:41:11 +0000 (10:41 +0100)]
rtcpbuffer: fix invalid read in validation of padding in rtcp packet

13 years agobaseaudiosrc: Add src object lock around call to ringbuffer parse caps.
Stian Johansen [Wed, 23 Feb 2011 09:55:12 +0000 (10:55 +0100)]
baseaudiosrc: Add src object lock around call to ringbuffer parse caps.

A race was observed between query() and setcaps() where the latter would
change the ringbuffer spec while the former was performing operations
based this data.

13 years agobaseaudiosrc: protect against ringbuffer disappearing while in a query
Havard Graff [Sat, 22 Jan 2011 22:09:32 +0000 (23:09 +0100)]
baseaudiosrc: protect against ringbuffer disappearing while in a query

Observed a case where the src went to null-state during the query,
hence the spec pointer was no longer valid, and
gst_util_unit64_scale_int crashed (assertion `denom > 0´failed)

Add locking to make sure the ringbuffer can't disappear.

13 years agobaseaudiosink: don't allow aligning behind the read-segment
Havard Graff [Tue, 8 Feb 2011 17:27:43 +0000 (18:27 +0100)]
baseaudiosink: don't allow aligning behind the read-segment

Given a large enough drift-tolerance, one could end up in a situation
where one would keep aligning the written buffers behind the current
read-segment position. The result for the reader would be complete
silence, possible preceded by very choppy audio.

By checking the available headroom, one can determine if there is
room to do alignment, or if one should resort to a resync instead to get
the pointers back on track.

Also refactor the alignment-logic out of the render function for cleaner
code.

13 years agoRemove setting of plugindir from Makefiles
David Schleef [Fri, 1 Apr 2011 20:55:26 +0000 (13:55 -0700)]
Remove setting of plugindir from Makefiles

13 years agovideo: Fix height calculation for YUV9/YVU9
David Schleef [Thu, 24 Mar 2011 06:10:51 +0000 (23:10 -0700)]
video: Fix height calculation for YUV9/YVU9

13 years agooggmux: fix warning building in mac os x
Josep Torra [Fri, 1 Apr 2011 13:34:30 +0000 (15:34 +0200)]
oggmux: fix warning building in mac os x

13 years agotextoverlay: fix comparison is always false due to limited range of data type
Josep Torra [Fri, 1 Apr 2011 13:33:42 +0000 (15:33 +0200)]
textoverlay: fix comparison is always false due to limited range of data type

Perform calculation in a temp var with enough room as there's guarantee that
ret will be able to hold the result for example in _blit_AYUV.

13 years agovorbistag: Write GST_TAG_IMAGE and GST_TAG_PREVIEW_IMAGE as METADATA_BLOCK_PICTURE
Sebastian Dröge [Fri, 1 Apr 2011 10:52:05 +0000 (12:52 +0200)]
vorbistag: Write GST_TAG_IMAGE and GST_TAG_PREVIEW_IMAGE as METADATA_BLOCK_PICTURE

This is the official, standardized way of embedding images into
vorbiscomments now.

13 years agovorbistag: Add support for METADATA_BLOCK_PICTURE tags
Sebastian Dröge [Fri, 1 Apr 2011 10:28:28 +0000 (12:28 +0200)]
vorbistag: Add support for METADATA_BLOCK_PICTURE tags

This is the official, standardized way of embedding pictures
inside vorbiscomments now. Parsing code taken from flacparse
and slightly changed.

Fixes bug #635669.

13 years agovorbistag: Use g_base64_decode_inplace()
Sebastian Dröge [Fri, 1 Apr 2011 10:09:44 +0000 (12:09 +0200)]
vorbistag: Use g_base64_decode_inplace()

Instead of using the GLib base64 decoding functions manually to
do inplace base64 decoding. This makes the code easier to understand.

13 years agooggmux: Store the segment directly inside the pad
Sebastian Dröge [Fri, 1 Apr 2011 09:00:38 +0000 (11:00 +0200)]
oggmux: Store the segment directly inside the pad

Also initialize it always in TIME format. We require TIME segments
in oggmux anyway and drop newsegment events in other formats and
assume an open-ended segment starting at 0.

13 years agooggmux: Reset the segment on flush-stop events and when going back to READY
Sebastian Dröge [Fri, 1 Apr 2011 08:57:08 +0000 (10:57 +0200)]
oggmux: Reset the segment on flush-stop events and when going back to READY

13 years agooggmux: Use running time instead of timestamps
Thiago Santos [Thu, 3 Mar 2011 11:45:15 +0000 (08:45 -0300)]
oggmux: Use running time instead of timestamps

Theora and vorbis use running time (which is correct) for calculating
the granulepos for their ogg packets. Oggmux, however, used
timestamps to order the received buffers.

This patch makes it use the running time to compare buffer times
and also to timestamp pushed buffers.

Some bits of the code still use timestamps, but they are only
used to calculate durations, so it should be fine.

https://bugzilla.gnome.org/show_bug.cgi?id=643775

13 years agooggmux: Keep track of pad's segments
Thiago Santos [Wed, 16 Feb 2011 19:07:49 +0000 (16:07 -0300)]
oggmux: Keep track of pad's segments

https://bugzilla.gnome.org/show_bug.cgi?id=643775

13 years agotextoverlay: Add support for xBGR and RGBx
Sebastian Dröge [Fri, 1 Apr 2011 08:39:31 +0000 (10:39 +0200)]
textoverlay: Add support for xBGR and RGBx

Now all RGB variants are supported.