]> git.sesse.net Git - vlc/log
vlc
10 years agoconfigure: no deprecated libav features in use
Rémi Denis-Courmont [Thu, 8 Aug 2013 18:23:26 +0000 (21:23 +0300)]
configure: no deprecated libav features in use

10 years agoavcodec: use reference counting for audio decoding
Rémi Denis-Courmont [Thu, 8 Aug 2013 18:22:11 +0000 (21:22 +0300)]
avcodec: use reference counting for audio decoding

10 years agoavcodec: refer to AVFrame as a pointer, non-functional change
Rémi Denis-Courmont [Thu, 8 Aug 2013 18:18:37 +0000 (21:18 +0300)]
avcodec: refer to AVFrame as a pointer, non-functional change

10 years agoavcodec: avoid forward declaration
Rémi Denis-Courmont [Thu, 25 Jul 2013 16:57:48 +0000 (19:57 +0300)]
avcodec: avoid forward declaration

10 years agoContribs: update libpng to 1.6.3
Jean-Baptiste Kempf [Thu, 8 Aug 2013 15:55:57 +0000 (17:55 +0200)]
Contribs: update libpng to 1.6.3

10 years agoaout: rework aout_Interleave()
Rémi Denis-Courmont [Thu, 8 Aug 2013 17:17:59 +0000 (20:17 +0300)]
aout: rework aout_Interleave()

This no longer assumes that planes are contiguous. It should help with
libav reference counting, and also avoids interleaving unused padding.

10 years agoContribs: update ZVBI (!) to 0.2.34
Jean-Baptiste Kempf [Thu, 8 Aug 2013 16:34:33 +0000 (18:34 +0200)]
Contribs: update ZVBI (!) to 0.2.34

10 years agoContribs: update freetype to 2.5.0.1
Jean-Baptiste Kempf [Thu, 8 Aug 2013 15:46:40 +0000 (17:46 +0200)]
Contribs: update freetype to 2.5.0.1

10 years agoContribs: update Opus to 1.0.3
Jean-Baptiste Kempf [Thu, 8 Aug 2013 15:41:33 +0000 (17:41 +0200)]
Contribs: update Opus to 1.0.3

10 years agoGLwin32: fix init crash
Jean-Baptiste Kempf [Thu, 8 Aug 2013 14:26:52 +0000 (16:26 +0200)]
GLwin32: fix init crash

Regression from 5989169 and ticket #7558

Close #9124

10 years agoMap Voxware MetaSound Codec IDs
Jean-Baptiste Kempf [Thu, 8 Aug 2013 10:57:54 +0000 (12:57 +0200)]
Map Voxware MetaSound Codec IDs

10 years agoXiph: support TRACKNUMBER=xx/xx in vorbis comments
Jean-Baptiste Kempf [Thu, 8 Aug 2013 10:07:37 +0000 (12:07 +0200)]
Xiph: support TRACKNUMBER=xx/xx in vorbis comments

Yeah, yeah, such a clever idea, thx...
Let' not use TRACKTOTAL or TOTALTRACKS or TOTALTRACK or TRACKSTOTAL,
because you know, there is not enough options...
Oh, and let's not make ANY of the above official in the spec, because
then, people could follow the spec...
https://www.xiph.org/vorbis/doc/v-comment.html

Next time, please use mkv...

10 years agoAvcodec: correctly #ifdef VOXWARE CODEC_ID
Jean-Baptiste Kempf [Wed, 7 Aug 2013 09:18:31 +0000 (11:18 +0200)]
Avcodec: correctly #ifdef VOXWARE CODEC_ID

10 years agocodec: x264: avoid segfault with x264_encoder_delayed_frames
Francois Cartegnie [Tue, 6 Aug 2013 20:12:47 +0000 (22:12 +0200)]
codec: x264: avoid segfault with x264_encoder_delayed_frames

10 years agointeger: fix typo (fix #9127)
Rémi Denis-Courmont [Tue, 6 Aug 2013 19:31:47 +0000 (22:31 +0300)]
integer: fix typo (fix #9127)

10 years agoaraw: fix samples count on raw audio
Rémi Denis-Courmont [Tue, 6 Aug 2013 19:26:36 +0000 (22:26 +0300)]
araw: fix samples count on raw audio

(regression from 46f782a9f775156c1f9e53a39632e28e9854c6ef)

10 years agovmem: set video offset explicitly to zero
Rémi Denis-Courmont [Tue, 6 Aug 2013 19:26:00 +0000 (22:26 +0300)]
vmem: set video offset explicitly to zero

(since the callbacks do not know the actual offsets)

10 years agovmem: pass visible dimensions (refs #9147)
Rémi Denis-Courmont [Tue, 6 Aug 2013 18:21:08 +0000 (21:21 +0300)]
vmem: pass visible dimensions (refs #9147)

10 years agomacosx vout: fix pool leaks
David Fuhrmann [Tue, 6 Aug 2013 18:18:53 +0000 (20:18 +0200)]
macosx vout: fix pool leaks

10 years agomacosx: prevent possible double release for blackout windows
David Fuhrmann [Tue, 6 Aug 2013 18:08:55 +0000 (20:08 +0200)]
macosx: prevent possible double release for blackout windows

10 years agomacosx: add missing deletate for blackout window hiding animation (fixes #9106)
David Fuhrmann [Tue, 6 Aug 2013 17:59:05 +0000 (19:59 +0200)]
macosx: add missing deletate for blackout window hiding animation (fixes #9106)

10 years agoavcodec: fix sign warning
Rémi Denis-Courmont [Tue, 6 Aug 2013 17:10:59 +0000 (20:10 +0300)]
avcodec: fix sign warning

10 years agovdpau: udpate to latest libavcodec
Rémi Denis-Courmont [Tue, 6 Aug 2013 16:45:00 +0000 (19:45 +0300)]
vdpau: udpate to latest libavcodec

10 years agohwdummy: update to latest libavcodec
Rémi Denis-Courmont [Tue, 6 Aug 2013 16:44:50 +0000 (19:44 +0300)]
hwdummy: update to latest libavcodec

10 years agoqt4: use current file's path as suggested destination path
Adrian Yanes [Mon, 5 Aug 2013 21:30:26 +0000 (14:30 -0700)]
qt4: use current file's path as suggested destination path

closes #8981

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
10 years agoalsa: set period time before buffer time
Rémi Denis-Courmont [Mon, 5 Aug 2013 16:43:57 +0000 (19:43 +0300)]
alsa: set period time before buffer time

Currently, the period time is more critical, so constrain it first.
Both values are interdependant with some drivers.

10 years agoconfigure: require vdpau version 0.6
Rémi Denis-Courmont [Mon, 5 Aug 2013 16:16:33 +0000 (19:16 +0300)]
configure: require vdpau version 0.6

This is not fool proof, as the run-time version is not checked.
But versions 0.4/0.5 crash due to a bug in the libvdpau DRI2 code.

10 years agoRevert "vdpau: link against VDPAU_LIBS"
Rémi Denis-Courmont [Mon, 5 Aug 2013 16:17:34 +0000 (19:17 +0300)]
Revert "vdpau: link against VDPAU_LIBS"

This reverts commit a08cf819549e9af39ad95d22a64b0b94ff46f06e.

10 years agoaout: allow frequency up to 352.8 kHz
Rémi Denis-Courmont [Mon, 5 Aug 2013 15:58:26 +0000 (18:58 +0300)]
aout: allow frequency up to 352.8 kHz

This is the frequency for SACD Digital eXtreme Definition.

Higher sample rates exist, notably DSD's 2822.4kH but that is not PCM
and the bitrate computation would really risk 31-bits integer overflow.

10 years agoandroidsurface: Support cropping
Martin Storsjö [Mon, 29 Jul 2013 18:58:24 +0000 (21:58 +0300)]
androidsurface: Support cropping

Since b71c85b3d88b8d, the output from the avcodec video decoder
requires the vout to handle the cropping. This uses an updated
version of the jni_SetAndroidSurfaceSize function to be able to
pass the visible size separately from the full size.

Signed-off-by: Martin Storsjö <martin@martin.st>
10 years agoyoutube.lua: fix encoded characters in author name (web page)
Pierre Ynard [Mon, 5 Aug 2013 01:24:49 +0000 (03:24 +0200)]
youtube.lua: fix encoded characters in author name (web page)

Characters with accents are not encoded, but apostrophes happen

10 years agoyoutube.lua: fix encoded characters in author name (API)
Pierre Ynard [Mon, 5 Aug 2013 01:12:14 +0000 (03:12 +0200)]
youtube.lua: fix encoded characters in author name (API)

With google accounts it's happening now

10 years agoyoutube.lua: update author name extraction
Pierre Ynard [Sun, 4 Aug 2013 23:11:35 +0000 (01:11 +0200)]
youtube.lua: update author name extraction

10 years agoRTP access: fix dependency on rc.o when gcrypt is absent
Rafaël Carré [Sat, 3 Aug 2013 09:17:56 +0000 (11:17 +0200)]
RTP access: fix dependency on rc.o when gcrypt is absent

10 years agovout_ios2: take into account scale of attached screen
Gleb Pinigin [Fri, 2 Aug 2013 12:41:55 +0000 (19:41 +0700)]
vout_ios2: take into account scale of attached screen

As said in Apple documentation drawRect should not be implemented for view based on opengl es layer.
Instead contentScaleFactor should be changed manually if needed. Underlying opengl es layer will adjust its scale accordingly.

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>
10 years agoRelicense access outputs to LGPL
Jean-Baptiste Kempf [Fri, 2 Aug 2013 17:16:25 +0000 (19:16 +0200)]
Relicense access outputs to LGPL

This is exception of bonjour, because of jlj

10 years agoVisualizations: relicense to LGPLv2.1
Jean-Baptiste Kempf [Fri, 2 Aug 2013 15:46:27 +0000 (17:46 +0200)]
Visualizations: relicense to LGPLv2.1

Note that the visual plugin still depends on a GPL'd fft.
Note that the new vsxu plugin is still GPL (depending on a GPL library
anyway)

10 years agoRevert "Fix scope alignment"
Jean-Baptiste Kempf [Fri, 2 Aug 2013 15:36:28 +0000 (17:36 +0200)]
Revert "Fix scope alignment"

This reverts commit e61ba5a72f46d5f67b361bd5f617f661f64e5fb7.

10 years agoCropping filter: fix description.
Jean-Baptiste Kempf [Fri, 2 Aug 2013 12:37:52 +0000 (14:37 +0200)]
Cropping filter: fix description.

Close #9117

10 years agovdpau: link against VDPAU_LIBS
Jean-Paul Saman [Fri, 2 Aug 2013 11:55:51 +0000 (13:55 +0200)]
vdpau: link against VDPAU_LIBS

10 years agoAdjust the DirectSound buffer size to support correctly 6 channels
Denis Charmet [Thu, 1 Aug 2013 21:03:46 +0000 (23:03 +0200)]
Adjust the DirectSound buffer size to support correctly 6 channels

The buffer size must be a multiple of the number of channels*sample_size or looping will break the channel layout.

Close #8483

10 years agoVersion all plugins when building for Windows
Rafaël Carré [Thu, 1 Aug 2013 15:02:52 +0000 (17:02 +0200)]
Version all plugins when building for Windows

Refs #9108

10 years agonet_Recv(): improve error handling (fixes #9081)
Rémi Denis-Courmont [Tue, 30 Jul 2013 15:07:14 +0000 (18:07 +0300)]
net_Recv(): improve error handling (fixes #9081)

10 years agoCorrect Winsock error #define's
Rémi Denis-Courmont [Tue, 30 Jul 2013 14:59:50 +0000 (17:59 +0300)]
Correct Winsock error #define's

10 years agobuild: detect variants of powerpc host_cpu (such as powerpc64)
Dominique Leuenberger [Mon, 29 Jul 2013 18:14:20 +0000 (20:14 +0200)]
build: detect variants of powerpc host_cpu (such as powerpc64)

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
10 years agoandroidsurface: Release the ANativeWindow before closing the dlopened library
Martin Storsjö [Mon, 29 Jul 2013 14:05:36 +0000 (17:05 +0300)]
androidsurface: Release the ANativeWindow before closing the dlopened library

This fixes ANativeWindow usage on 2.3 and 3.x. Therefore, revert
back to preferring ANativeWindow over the private symbols.

Signed-off-by: Martin Storsjö <martin@martin.st>
10 years agoandroidsurface: Only use ANativeWindow if the private symbols aren't found
Martin Storsjö [Mon, 29 Jul 2013 07:55:42 +0000 (10:55 +0300)]
androidsurface: Only use ANativeWindow if the private symbols aren't found

Even if ANativeWindow is a public API, the release call seems to
crash on a Nexus One (running 2.3) and on a Galaxy Tab (running 3.2).
The exact reason is not known or understood yet, but it might be
due to accessing and dealing with the Surface from both Java (via the
SurfaceHolder class) and via the ANativeWindow API.

Therefore, only use the ANativeWindow if the old methods aren't
found (that is, on 4.3).

The ANativeWindow output works fine on firmwares as early as 4.0
though.

Signed-off-by: Martin Storsjö <martin@martin.st>
10 years agocontrib: CCAS needs -c option
Rafaël Carré [Mon, 29 Jul 2013 05:29:03 +0000 (07:29 +0200)]
contrib: CCAS needs -c option

Fix ba8cad3aeec1

10 years agomacosx: fix compilation on 10.6 with clang 3.0
Felix Paul Kühne [Sun, 28 Jul 2013 22:35:13 +0000 (00:35 +0200)]
macosx: fix compilation on 10.6 with clang 3.0

10 years agoqtcapture: fix compilation on 10.6
Felix Paul Kühne [Sun, 28 Jul 2013 22:27:44 +0000 (00:27 +0200)]
qtcapture: fix compilation on 10.6

10 years agomacosx: fix compatibility header for compilation on 10.6
Felix Paul Kühne [Sun, 28 Jul 2013 22:23:06 +0000 (00:23 +0200)]
macosx: fix compatibility header for compilation on 10.6

10 years agoalways hide menu bar for fullscreen on osx redacted (fixes #9040)
David Fuhrmann [Sun, 28 Jul 2013 18:03:31 +0000 (20:03 +0200)]
always hide menu bar for fullscreen on osx redacted (fixes #9040)

10 years agoequalizer: fix inverted parameters in test case
Rémi Denis-Courmont [Sun, 28 Jul 2013 20:10:27 +0000 (23:10 +0300)]
equalizer: fix inverted parameters in test case

10 years agogettext: update to version 0.18.3
Rémi Denis-Courmont [Sun, 28 Jul 2013 20:02:10 +0000 (23:02 +0300)]
gettext: update to version 0.18.3

10 years agoconfigure: remove old hack, fix bootstrapping with newer libtool
Rémi Denis-Courmont [Sun, 28 Jul 2013 19:59:13 +0000 (22:59 +0300)]
configure: remove old hack, fix bootstrapping with newer libtool

10 years agoaudiounit_ios: Fill the remainder of the buffer with zeros
Martin Storsjö [Tue, 23 Jul 2013 13:58:36 +0000 (16:58 +0300)]
audiounit_ios: Fill the remainder of the buffer with zeros

If we didn't have enough data to fill the buffer, fill the rest
of it with zeros. This is better than playing back whatever happened
to be there from before.

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>
10 years agoRevert "audiounit_ios: try to handle data drops more efficiently"
Martin Storsjö [Tue, 23 Jul 2013 13:39:30 +0000 (16:39 +0300)]
Revert "audiounit_ios: try to handle data drops more efficiently"

This (more or less) reverts commit
2b7e016807507b632b8e0a7c874030bdffb2f604.

Stopping the aout from within the callback like this could
lead to deadlocks, where AudioOutputUnitStop in the callback thread
and AudioOutputUnitStart in the audio decoder thread blocked each
other (noticed at startup of playback on a 3GS with iOS 6.0).

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>
10 years agoaout: stick to single precision
Rémi Denis-Courmont [Sun, 28 Jul 2013 13:31:15 +0000 (16:31 +0300)]
aout: stick to single precision

10 years agoyoutube.lua: basic support for scrambled signatures
Pierre Ynard [Sat, 27 Jul 2013 19:52:00 +0000 (21:52 +0200)]
youtube.lua: basic support for scrambled signatures

Apparently this is the new fad and comes in several length-flavors

10 years agoavformat: leave SAR at 0/0 (undefined) if unknown
Rémi Denis-Courmont [Sat, 27 Jul 2013 19:36:41 +0000 (22:36 +0300)]
avformat: leave SAR at 0/0 (undefined) if unknown

10 years agoavformat: set pixel aspect ratio
Wills Wang [Sat, 27 Jul 2013 19:30:03 +0000 (22:30 +0300)]
avformat: set pixel aspect ratio

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
10 years agodoc: remove mad documentation
Rémi Denis-Courmont [Fri, 26 Jul 2013 19:52:51 +0000 (22:52 +0300)]
doc: remove mad documentation

This is the _VLC_ source code.

10 years agoavcodec: do not allocate unused output buffers
Rémi Denis-Courmont [Fri, 26 Jul 2013 19:18:33 +0000 (22:18 +0300)]
avcodec: do not allocate unused output buffers

The buffer from decoder_NewAudioBuffer() was only used if the audio
output format was planar. Now it is allocated only in that case.

This also remove leaks when decoder_NewAudioBuffer() failed.

10 years agoDo not compare arrays against NULL (always true)
Rémi Duraffort [Fri, 26 Jul 2013 17:56:51 +0000 (19:56 +0200)]
Do not compare arrays against NULL (always true)

Fix cid #1048854

10 years agoFix sizeof mismatch
Rémi Duraffort [Fri, 26 Jul 2013 17:29:07 +0000 (19:29 +0200)]
Fix sizeof mismatch

10 years agoFix many memory leaks (and potential ones)
Rémi Duraffort [Tue, 23 Jul 2013 19:55:22 +0000 (21:55 +0200)]
Fix many memory leaks (and potential ones)

In case of a wrongly formated file, the strings where never dealocated. Same if
some strings where empty.

10 years agoaraw: zero copy for native audio sample formats
Rémi Denis-Courmont [Fri, 26 Jul 2013 16:49:23 +0000 (19:49 +0300)]
araw: zero copy for native audio sample formats

10 years agodecoder: separate aout initialization from buffer allocation
Rémi Denis-Courmont [Fri, 26 Jul 2013 16:36:53 +0000 (19:36 +0300)]
decoder: separate aout initialization from buffer allocation

This has two benefits:
 - allow zero copy in audio decoders,
 - allow negotiation of the output formats (potentially for S/PDIF).

10 years agoaout: inline aout_DecDeleteBuffer()
Rémi Denis-Courmont [Fri, 26 Jul 2013 16:05:29 +0000 (19:05 +0300)]
aout: inline aout_DecDeleteBuffer()

10 years agoaout: inline aout_DecNewBuffer()
Rémi Denis-Courmont [Fri, 26 Jul 2013 16:04:14 +0000 (19:04 +0300)]
aout: inline aout_DecNewBuffer()

10 years agoandroidsurface: Use the ANativeWindow API where available
Martin Storsjö [Thu, 25 Jul 2013 13:05:36 +0000 (16:05 +0300)]
androidsurface: Use the ANativeWindow API where available

This makes the android vout work on Android 4.3.

This API is available since Android 2.3/Gingerbread (api level 9).
It's a public, official API, contrary to the earlier hacking by
accessing the internal C++ Surface class directly. In most cases,
the ANativeWindow API is a very thin wrapper doing pretty much
the same as the existing Surface class entry points we've been using.

In Android 4.3, the C++ Surface class lock function we've been using
was renamed again, but using the new corresponding function as
we've done before doesn't work any longer.

Therefore, use the public API where possible (loading it dynamically,
to not break compatibility with pre-2.3 devices).

The only non-public API used in the vout is now the YV12
pixel format, which doesn't work on all devices.

Signed-off-by: Martin Storsjö <martin@martin.st>
10 years agoequalizer: apply preamp twice on 2nd pass
Rafaël Carré [Thu, 25 Jul 2013 16:15:38 +0000 (18:15 +0200)]
equalizer: apply preamp twice on 2nd pass

Fixes #8760

10 years agoALSA: favor channel maps with fewer unused channels
Rémi Denis-Courmont [Thu, 25 Jul 2013 20:21:02 +0000 (23:21 +0300)]
ALSA: favor channel maps with fewer unused channels

10 years agoALSA: still use VLC configuration to determine available speakers
Rémi Denis-Courmont [Thu, 25 Jul 2013 20:16:34 +0000 (23:16 +0300)]
ALSA: still use VLC configuration to determine available speakers

The ALSA channels maps only indicate the order(s) of channels supported
by the audio adapter. They do not imply that all channels are actually
physically wired.

10 years agovdpau: improve decoder capability checks messages
Rémi Denis-Courmont [Thu, 25 Jul 2013 16:35:32 +0000 (19:35 +0300)]
vdpau: improve decoder capability checks messages

10 years agoavcodec: rationalize hwaccel copy/extract callback prototype
Rémi Denis-Courmont [Thu, 25 Jul 2013 16:05:20 +0000 (19:05 +0300)]
avcodec: rationalize hwaccel copy/extract callback prototype

10 years agoavcodec: rationalize hwaccel surface allocation callback
Rémi Denis-Courmont [Thu, 25 Jul 2013 15:57:18 +0000 (18:57 +0300)]
avcodec: rationalize hwaccel surface allocation callback

10 years agoavcodec: pass data pointer to hwaccel release callbacks and simplify
Rémi Denis-Courmont [Thu, 25 Jul 2013 15:45:24 +0000 (18:45 +0300)]
avcodec: pass data pointer to hwaccel release callbacks and simplify

10 years agoavcodec audio: update block pointer on realloc
Rafaël Carré [Thu, 25 Jul 2013 10:52:41 +0000 (12:52 +0200)]
avcodec audio: update block pointer on realloc

10 years agocontrib: gcrypt: upgrade to 1.5.3
Rafaël Carré [Thu, 25 Jul 2013 10:32:01 +0000 (12:32 +0200)]
contrib: gcrypt: upgrade to 1.5.3

10 years agoFix live555 breakage in last update
Jean-Baptiste Kempf [Wed, 24 Jul 2013 23:14:32 +0000 (01:14 +0200)]
Fix live555 breakage in last update

10 years agodirectsound: clear the ring buffer ahead of the write pointer (fix #8860)
Ludovic Fauvet [Wed, 24 Jul 2013 11:01:01 +0000 (13:01 +0200)]
directsound: clear the ring buffer ahead of the write pointer (fix #8860)

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
10 years agodirectsound: remove broken notification support
Ludovic Fauvet [Wed, 24 Jul 2013 11:01:00 +0000 (13:01 +0200)]
directsound: remove broken notification support

The original code was non-working because the notication was never
setup correctly and resulted in a DSERR_INVALIDCALL. A proper fix
would look like:

         if( sys->p_notify )
         {
+            IDirectSoundBuffer_Stop( aout->sys->p_dsbuffer );
             DSBPOSITIONNOTIFY notif = {.dwOffset = aout->sys->i_write, .hEventNotify = sys->hnotify_evt } ;
             if( IDirectSoundNotify_SetNotificationPositions( sys->p_notify, 1, &notif ) ==  DS_OK )
             {
+                IDirectSoundBuffer_Play( aout->sys->p_dsbuffer, 0, 0, DSBPLAY_LOOPING );
                 WaitForSingleObject( sys->hnotify_evt, INFINITE );
                 IDirectSoundBuffer_Stop( aout->sys->p_dsbuffer );
             }

But fixing it makes the problem worse since the notification is setup
too late and thus the buffer will replay a full loop of the back
buffer during a drain.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
10 years agodirectsound: remove the broken msleep fallback
Ludovic Fauvet [Wed, 24 Jul 2013 11:00:59 +0000 (13:00 +0200)]
directsound: remove the broken msleep fallback

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
10 years agodirectsound: clear the rest of the ring buffer too
Ludovic Fauvet [Wed, 24 Jul 2013 11:00:58 +0000 (13:00 +0200)]
directsound: clear the rest of the ring buffer too

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
10 years agolpcm: fix channel reordering for m2ts (close #8526)
Ludovic Fauvet [Wed, 24 Jul 2013 08:32:38 +0000 (10:32 +0200)]
lpcm: fix channel reordering for m2ts (close #8526)

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
10 years agovda: reset correctly the coded frame size.
Sebastien Zwickert [Wed, 24 Jul 2013 14:08:17 +0000 (16:08 +0200)]
vda: reset correctly the coded frame size.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
10 years agoavcodec: pass AVFrame.opaque to hwaccel release callback
Rémi Denis-Courmont [Wed, 24 Jul 2013 20:25:10 +0000 (23:25 +0300)]
avcodec: pass AVFrame.opaque to hwaccel release callback

With the reference counting stuff, the AVFrame structure is gone or
unreachable by the time the release function is called.

10 years agoavcodec: remove va parameter from hwaccel release callback
Rémi Denis-Courmont [Wed, 24 Jul 2013 20:32:27 +0000 (23:32 +0300)]
avcodec: remove va parameter from hwaccel release callback

10 years agohwdummy: remove no-op
Rémi Denis-Courmont [Wed, 24 Jul 2013 20:24:51 +0000 (23:24 +0300)]
hwdummy: remove no-op

10 years agovaapi: avoid usage of vlc_va_t * from the release callback
Rémi Denis-Courmont [Wed, 24 Jul 2013 20:29:21 +0000 (23:29 +0300)]
vaapi: avoid usage of vlc_va_t * from the release callback

This will simplify later commits.

10 years agovdpau: remove no-op
Rémi Denis-Courmont [Wed, 24 Jul 2013 20:25:01 +0000 (23:25 +0300)]
vdpau: remove no-op

10 years agovaapi: use AVFrame.opaque to store per-picture state directly
Rémi Denis-Courmont [Wed, 24 Jul 2013 20:21:45 +0000 (23:21 +0300)]
vaapi: use AVFrame.opaque to store per-picture state directly

10 years agodxva2: use AVFrame.opaque to store internal per-picture state directly
Rémi Denis-Courmont [Wed, 24 Jul 2013 20:21:00 +0000 (23:21 +0300)]
dxva2: use AVFrame.opaque to store internal per-picture state directly

10 years agovdpau: fix surface leaks if the picture is not extracted
Rémi Denis-Courmont [Wed, 24 Jul 2013 19:48:50 +0000 (22:48 +0300)]
vdpau: fix surface leaks if the picture is not extracted

This can happen in some corner cases such as flushing (luckily the leak
would automatically be resolved when the device instance is released).

10 years agovdpau: revector (no functional changes)
Rémi Denis-Courmont [Wed, 24 Jul 2013 19:39:44 +0000 (22:39 +0300)]
vdpau: revector (no functional changes)

10 years agoALSA: correct debug message
Rémi Denis-Courmont [Wed, 24 Jul 2013 19:07:27 +0000 (22:07 +0300)]
ALSA: correct debug message

10 years agoavcodec. remove redundant code
Rémi Denis-Courmont [Wed, 24 Jul 2013 17:34:47 +0000 (20:34 +0300)]
avcodec. remove redundant code

10 years agoXCB/XVideo: pad pictures to 32x16
Rémi Denis-Courmont [Wed, 24 Jul 2013 16:41:26 +0000 (19:41 +0300)]
XCB/XVideo: pad pictures to 32x16

This reenables direct rendering with libavcodec in most cases.

10 years agoavcodec: log why picture is unsuitable for direct rendering
Rémi Denis-Courmont [Wed, 24 Jul 2013 16:41:08 +0000 (19:41 +0300)]
avcodec: log why picture is unsuitable for direct rendering