vlc.git
23 months agovout: remove unused mouse visible/invisible event
Rémi Denis-Courmont [Sun, 6 Aug 2017 14:54:36 +0000 (17:54 +0300)]
vout: remove unused mouse visible/invisible event

The window/display already have events/controls for this, and nothing
else cared, cares and probably will ever care.

23 months agovout: remove core event thread
Rémi Denis-Courmont [Sat, 5 Aug 2017 20:49:13 +0000 (23:49 +0300)]
vout: remove core event thread

23 months agocaca: use new event thread helper
Rémi Denis-Courmont [Sat, 5 Aug 2017 20:48:14 +0000 (23:48 +0300)]
caca: use new event thread helper

23 months agoaa: use new event thread helper
Rémi Denis-Courmont [Sat, 5 Aug 2017 20:48:14 +0000 (23:48 +0300)]
aa: use new event thread helper

23 months agovout: add legacy event thread helper outside core
Rémi Denis-Courmont [Sat, 5 Aug 2017 20:47:32 +0000 (23:47 +0300)]
vout: add legacy event thread helper outside core

23 months agovout display: deprecate the "manage" callback (refs #17700)
Rémi Denis-Courmont [Sat, 5 Aug 2017 20:13:43 +0000 (23:13 +0300)]
vout display: deprecate the "manage" callback (refs #17700)

23 months agovout_display: remove/deprecate fullscreen state
Rémi Denis-Courmont [Sat, 5 Aug 2017 20:04:57 +0000 (23:04 +0300)]
vout_display: remove/deprecate fullscreen state

Live window state, this was moved from display to window eight years
ago (minus one week). When the fullscreen state changes, the window
plugin emits a resize event, which triggers any necessary updates in
the display. The display does not need to know about fullscreen state
as such. But it is retained on HWND platforms due to legacy code, just
like window state before.

23 months agosdl: remove video output
Rémi Denis-Courmont [Sat, 5 Aug 2017 18:08:45 +0000 (21:08 +0300)]
sdl: remove video output

The SDL video output has been forcefully disabled on MacOS X for 12
years. It has also been disabled in Windows builds for over 6 years.
Finally, it has been broken on X11 due to lack of windowing
integration, for several years, cannot be fixed without changes to
the SDL API, and is completely useless.

23 months agodirectfb: remove legacy video output
Rémi Denis-Courmont [Sat, 5 Aug 2017 18:00:06 +0000 (21:00 +0300)]
directfb: remove legacy video output

The DirectFB project has been dead for several years, and is ostensibly
not used. If you really need direct access to the Linux frame buffer,
there is still the raw LinuxFB module for the time being.

23 months agovout_display: mark CHANGE_WINDOW_STATE deprecated
Rémi Denis-Courmont [Sat, 5 Aug 2017 20:03:23 +0000 (23:03 +0300)]
vout_display: mark CHANGE_WINDOW_STATE deprecated

This was moved from display to window eight years ago.

23 months agodisplay: preserve actual height on A/R or crop change
Rémi Denis-Courmont [Sat, 5 Aug 2017 19:16:49 +0000 (22:16 +0300)]
display: preserve actual height on A/R or crop change

...rather than the last saved height. In practice, this does not make
much if any difference though, as:
- the saved height should equal the actual height in windowed mode,
- the code path has no effects in fullscreen mode.

23 months agommal vout: pictures reset is impossible
Rémi Denis-Courmont [Sat, 5 Aug 2017 18:23:01 +0000 (21:23 +0300)]
mmal vout: pictures reset is impossible

Assert that it does not happen, like with all or almost all other
non-resetting video outputs.

23 months agoold http: remove useless b_continuous member
Zhao Zhili [Fri, 4 Aug 2017 02:46:19 +0000 (10:46 +0800)]
old http: remove useless b_continuous member

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
23 months agodisplay: no need to pass the video_format_t for CROP/ASPECT_RATIO changes
Steve Lhomme [Wed, 2 Aug 2017 07:33:37 +0000 (09:33 +0200)]
display: no need to pass the video_format_t for CROP/ASPECT_RATIO changes

It's always found in vd->source.

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
23 months agodisplay: no need for a local video_format_t with VOUT_DISPLAY_CHANGE_SOURCE_CROP
Steve Lhomme [Wed, 2 Aug 2017 07:33:36 +0000 (09:33 +0200)]
display: no need for a local video_format_t with VOUT_DISPLAY_CHANGE_SOURCE_CROP

Receivers don't check the difference with the original value, they just apply
the new value.

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
23 months agodisplay: no need for a local video_format_t with VOUT_DISPLAY_CHANGE_SOURCE_ASPECT
Steve Lhomme [Wed, 2 Aug 2017 07:33:35 +0000 (09:33 +0200)]
display: no need for a local video_format_t with VOUT_DISPLAY_CHANGE_SOURCE_ASPECT

Receivers don't check the difference with the original value, they just apply
the new value.

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
23 months agoshare: Install precompiled lua scripts in /share instead of /lib
Hugo Beauzée-Luyssen [Thu, 3 Aug 2017 14:31:09 +0000 (16:31 +0200)]
share: Install precompiled lua scripts in /share instead of /lib

Fix #18639

23 months agolib: Automatically select the default teletext page on activation
Hugo Beauzée-Luyssen [Fri, 4 Aug 2017 13:05:02 +0000 (15:05 +0200)]
lib: Automatically select the default teletext page on activation

Fix #18638

23 months agodeinterlace:dxva: fix a bogus string test
Steve Lhomme [Fri, 4 Aug 2017 09:30:56 +0000 (11:30 +0200)]
deinterlace:dxva: fix a bogus string test

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: do not look for seekpoints for tracks that are not selected
Steve Lhomme [Fri, 4 Aug 2017 09:31:44 +0000 (11:31 +0200)]
demux:mkv: do not look for seekpoints for tracks that are not selected

That will reduce the potential scope of the data we will read before seeking.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: do not use invalid seekpoints for the final seek target
Steve Lhomme [Fri, 4 Aug 2017 09:31:43 +0000 (11:31 +0200)]
demux:mkv: do not use invalid seekpoints for the final seek target

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodxva2_deinterlace: use a function to fill the DXVA2_VideoProcessBltParams
Steve Lhomme [Tue, 1 Aug 2017 08:25:45 +0000 (10:25 +0200)]
dxva2_deinterlace: use a function to fill the DXVA2_VideoProcessBltParams

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodxva2_deinterlace: replace an assert by a debug warning
Steve Lhomme [Tue, 1 Aug 2017 08:25:44 +0000 (10:25 +0200)]
dxva2_deinterlace: replace an assert by a debug warning

Might fix #18599

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agod3d11va: fix Xbox limitation test
Steve Lhomme [Fri, 28 Jul 2017 11:37:33 +0000 (13:37 +0200)]
d3d11va: fix Xbox limitation test

The va_pool surface dimensions are only set if the surfaces are created
successfully. We must use the temporary video_format_t dimensions at this stage.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodxva2_deinterlace: fix potential double free when there are errors
Steve Lhomme [Tue, 1 Aug 2017 08:25:43 +0000 (10:25 +0200)]
dxva2_deinterlace: fix potential double free when there are errors

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodxva2_deinterlace: fix wrong copy/paste in log
Steve Lhomme [Tue, 1 Aug 2017 08:25:42 +0000 (10:25 +0200)]
dxva2_deinterlace: fix wrong copy/paste in log

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodeinterlace:dxva: log when the algorithm doesn't match the user choice exactly
Steve Lhomme [Wed, 2 Aug 2017 06:31:56 +0000 (08:31 +0200)]
deinterlace:dxva: log when the algorithm doesn't match the user choice exactly

Since available ones may vary from hardware to hardware.

23 months agodirect3d11: always place the picture based on the quad source dimensions
Steve Lhomme [Tue, 1 Aug 2017 13:30:03 +0000 (15:30 +0200)]
direct3d11: always place the picture based on the quad source dimensions

They may not always be equal to the vd->fmt.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodirect3d11: decide the main quad dimensions early
Steve Lhomme [Tue, 1 Aug 2017 13:01:22 +0000 (15:01 +0200)]
direct3d11: decide the main quad dimensions early

We decide once in Direct3D11Open() and keep using that.
The staging texture has the same dimensions.

Might fix #18578

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodirect3d11: fix possibly failing CopySubresourceRegion on odd dimensions
Steve Lhomme [Tue, 1 Aug 2017 13:01:21 +0000 (15:01 +0200)]
direct3d11: fix possibly failing CopySubresourceRegion on odd dimensions

The texture is created with extra size so we can use more for the copy.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodirect3d11: remove write-only i_x/y_offset in d3d_quad_t
Steve Lhomme [Tue, 1 Aug 2017 13:01:20 +0000 (15:01 +0200)]
direct3d11: remove write-only i_x/y_offset in d3d_quad_t

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux: hls: code cleaning
Zhao Zhili [Wed, 2 Aug 2017 12:59:09 +0000 (20:59 +0800)]
demux: hls: code cleaning

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: simplify the default Seekpoint constructor
Steve Lhomme [Thu, 3 Aug 2017 16:18:35 +0000 (18:18 +0200)]
demux:mkv: simplify the default Seekpoint constructor

Use the full constructor.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: make the default trust level of a Seekpoint to TRUSTED
Steve Lhomme [Thu, 3 Aug 2017 16:18:34 +0000 (18:18 +0200)]
demux:mkv: make the default trust level of a Seekpoint to TRUSTED

For a needle it doesn't matter the level, it's never used.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: pass a Seekpoint to add_seekpoint
Steve Lhomme [Thu, 3 Aug 2017 16:18:33 +0000 (18:18 +0200)]
demux:mkv: pass a Seekpoint to add_seekpoint

No need to pass all parameters, especially if the Seekpoint prototype grows.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: remove the unused duration field from Seekpoint
Steve Lhomme [Thu, 3 Aug 2017 16:18:32 +0000 (18:18 +0200)]
demux:mkv: remove the unused duration field from Seekpoint

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: simplify the code as we always disable the trust check with get_seekpoints...
Steve Lhomme [Thu, 3 Aug 2017 16:18:31 +0000 (18:18 +0200)]
demux:mkv: simplify the code as we always disable the trust check with get_seekpoints_around()

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: don't look for the upper seekpoint when we only want the lower one
Steve Lhomme [Thu, 3 Aug 2017 16:18:30 +0000 (18:18 +0200)]
demux:mkv: don't look for the upper seekpoint when we only want the lower one

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: remove useless initialization
Steve Lhomme [Thu, 3 Aug 2017 16:18:29 +0000 (18:18 +0200)]
demux:mkv: remove useless initialization

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: init an invalid Seekpoint with a proper invalid seek position
Steve Lhomme [Thu, 3 Aug 2017 16:18:28 +0000 (18:18 +0200)]
demux:mkv: init an invalid Seekpoint with a proper invalid seek position

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: use the TrustLevel type when looking for seekpoint boundaries
Steve Lhomme [Thu, 3 Aug 2017 16:18:27 +0000 (18:18 +0200)]
demux:mkv: use the TrustLevel type when looking for seekpoint boundaries

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: clean the type and default value of the seek position
Steve Lhomme [Thu, 3 Aug 2017 16:18:26 +0000 (18:18 +0200)]
demux:mkv: clean the type and default value of the seek position

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agohttp: Ensure mobile buttons are on one line
Mathieu Parent [Thu, 3 Aug 2017 16:27:04 +0000 (18:27 +0200)]
http: Ensure mobile buttons are on one line

Tested with 320x480 or wider.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agohttp: Allow to select media from playlist too
Mathieu Parent [Thu, 3 Aug 2017 16:27:03 +0000 (18:27 +0200)]
http: Allow to select media from playlist too

Signed-off-by: Mathieu Parent <math.parent@gmail.com>
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agohttp: Handle unparsable item name in /requests/playlist_jstree.xml
Mathieu Parent [Thu, 3 Aug 2017 16:27:02 +0000 (18:27 +0200)]
http: Handle unparsable item name in /requests/playlist_jstree.xml

Signed-off-by: Mathieu Parent <math.parent@gmail.com>
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agoconfigure: Explain how to disable avcodec or alsa
Mathieu Parent [Thu, 3 Aug 2017 16:27:01 +0000 (18:27 +0200)]
configure: Explain how to disable avcodec or alsa

Signed-off-by: Mathieu Parent <math.parent@gmail.com>
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agokwallet: Fix leak on error
Hugo Beauzée-Luyssen [Thu, 3 Aug 2017 13:55:40 +0000 (15:55 +0200)]
kwallet: Fix leak on error

CID #1372939

23 months agoupnp: Fix leak on error
Hugo Beauzée-Luyssen [Thu, 3 Aug 2017 13:52:26 +0000 (15:52 +0200)]
upnp: Fix leak on error

CID #1398416

23 months agolibavi: Fix expension unsafe helper
Hugo Beauzée-Luyssen [Thu, 3 Aug 2017 13:24:46 +0000 (15:24 +0200)]
libavi: Fix expension unsafe helper

23 months agogl: vaapi: use wayland CFLAGS
Olaf Hering [Mon, 31 Jul 2017 07:38:26 +0000 (09:38 +0200)]
gl: vaapi: use wayland CFLAGS

Use the CFLAGS from wayland-client.pc because the header files are not
in the standard directories. In openSUSE they are found in
/usr/include/wayland. Fixes commit 14ba7d2d7e ("gl: vaapi: add DRM
support")

Signed-off-by: Olaf Hering <olaf@aepfle.de>
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agosplitter: wall: fix aspect ratio
Victorien Le Couviour--Tuffet [Wed, 2 Aug 2017 14:37:13 +0000 (16:37 +0200)]
splitter: wall: fix aspect ratio

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
23 months agomacosx: remove spaces from blank line
Victorien Le Couviour--Tuffet [Wed, 2 Aug 2017 14:37:11 +0000 (16:37 +0200)]
macosx: remove spaces from blank line

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: get the track directly from FindTrackByBlock
Steve Lhomme [Wed, 2 Aug 2017 15:07:26 +0000 (17:07 +0200)]
demux:mkv: get the track directly from FindTrackByBlock

And use const is places where the track is not modified.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: don't rely on the iterator to get the track number
Steve Lhomme [Wed, 2 Aug 2017 15:07:25 +0000 (17:07 +0200)]
demux:mkv: don't rely on the iterator to get the track number

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: use an enum for the seekpoint trust level
Steve Lhomme [Wed, 2 Aug 2017 13:56:05 +0000 (15:56 +0200)]
demux:mkv: use an enum for the seekpoint trust level

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: do not try to use a seekpoint for a track that doesn't exist
Steve Lhomme [Wed, 2 Aug 2017 13:56:04 +0000 (15:56 +0200)]
demux:mkv: do not try to use a seekpoint for a track that doesn't exist

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agodemux:mkv: don't add seekpoints for tracks that don't exist
Steve Lhomme [Wed, 2 Aug 2017 15:06:57 +0000 (17:06 +0200)]
demux:mkv: don't add seekpoints for tracks that don't exist

Fixes #18596

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agoaccess_out_dummy: remove useless Close()
Rémi Denis-Courmont [Wed, 2 Aug 2017 15:14:07 +0000 (18:14 +0300)]
access_out_dummy: remove useless Close()

23 months agoaccess_output: allow pf_seek to be NULL, simplify
Rémi Denis-Courmont [Wed, 2 Aug 2017 15:11:40 +0000 (18:11 +0300)]
access_output: allow pf_seek to be NULL, simplify

23 months agoftp: avoid reserved identifier as function name
Rémi Denis-Courmont [Wed, 2 Aug 2017 15:11:15 +0000 (18:11 +0300)]
ftp: avoid reserved identifier as function name

23 months agosmem: Don't write to a read only value
Hugo Beauzée-Luyssen [Wed, 2 Aug 2017 15:05:56 +0000 (17:05 +0200)]
smem: Don't write to a read only value

23 months agodemux: subtitle: Fix leaks
Hugo Beauzée-Luyssen [Wed, 2 Aug 2017 14:36:08 +0000 (16:36 +0200)]
demux: subtitle: Fix leaks

23 months agomacosx: Improve DefaultValueSlider snapping logic
Marvin Scholz [Wed, 2 Aug 2017 14:11:13 +0000 (16:11 +0200)]
macosx: Improve DefaultValueSlider snapping logic

Snapping should work better now, regardless of the slider size.

23 months agomacosx: Correct drawing of Volume slider in RTL mode
Marvin Scholz [Wed, 2 Aug 2017 14:09:34 +0000 (16:09 +0200)]
macosx: Correct drawing of Volume slider in RTL mode

23 months agoaudio output: Allow i_physical_channels to be unset when using pass through
Hugo Beauzée-Luyssen [Tue, 1 Aug 2017 11:20:00 +0000 (13:20 +0200)]
audio output: Allow i_physical_channels to be unset when using pass through

Fix #18614

23 months agopacketizer: mpegvideo: use enum for values
Francois Cartegnie [Tue, 1 Aug 2017 14:10:36 +0000 (16:10 +0200)]
packetizer: mpegvideo: use enum for values

should clear things up

23 months agopacketizer: mpegvideo: simplify cc flags handling
Francois Cartegnie [Tue, 1 Aug 2017 13:56:18 +0000 (15:56 +0200)]
packetizer: mpegvideo: simplify cc flags handling

23 months agodemux: adaptive: signal buffering state change in both cases
Zhao Zhili [Fri, 28 Jul 2017 08:43:28 +0000 (16:43 +0800)]
demux: adaptive: signal buffering state change in both cases

Signed-off-by: Francois Cartegnie <fcvlcdev@free.fr>
23 months agodemux: adaptive: fix slow buffering
Zhao Zhili [Fri, 28 Jul 2017 11:54:50 +0000 (19:54 +0800)]
demux: adaptive: fix slow buffering

Signed-off-by: Francois Cartegnie <fcvlcdev@free.fr>
23 months agodemux: adaptive: fix debug log
Zhao Zhili [Mon, 31 Jul 2017 08:01:44 +0000 (16:01 +0800)]
demux: adaptive: fix debug log

Signed-off-by: Francois Cartegnie <fcvlcdev@free.fr>
23 months agodirect3d11: fix inverted logic for sides clamping issue
Steve Lhomme [Tue, 1 Aug 2017 09:22:49 +0000 (11:22 +0200)]
direct3d11: fix inverted logic for sides clamping issue

Fixes #18631

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agovout_subpicture: patch fully transparent menu palette
Francois Cartegnie [Tue, 1 Aug 2017 07:19:24 +0000 (09:19 +0200)]
vout_subpicture: patch fully transparent menu palette

23 months agodirect3d11: use all the RECT used for the picture placement to detect changes
Steve Lhomme [Mon, 31 Jul 2017 13:29:14 +0000 (15:29 +0200)]
direct3d11: use all the RECT used for the picture placement to detect changes

Otherwise there are some cases where resizing occurs but not updated on the
display.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agopacketizer: mpegvideo: don't output slice on sequence extension
Francois Cartegnie [Mon, 31 Jul 2017 20:26:09 +0000 (22:26 +0200)]
packetizer: mpegvideo: don't output slice on sequence extension

23 months agopanoramix: remove shadowing variable
Rémi Denis-Courmont [Mon, 31 Jul 2017 18:30:09 +0000 (21:30 +0300)]
panoramix: remove shadowing variable

23 months agoUnexport video_splitter_(New|Delete)
Rémi Denis-Courmont [Mon, 31 Jul 2017 18:30:08 +0000 (21:30 +0300)]
Unexport video_splitter_(New|Delete)

23 months agovout: rename one variable for clarity
Rémi Denis-Courmont [Mon, 31 Jul 2017 18:12:33 +0000 (21:12 +0300)]
vout: rename one variable for clarity

23 months agovout: shrink struct member
Rémi Denis-Courmont [Mon, 31 Jul 2017 18:07:18 +0000 (21:07 +0300)]
vout: shrink struct member

This can only take values -1, 0 and 1.

23 months agovout: remove write-only struct member
Rémi Denis-Courmont [Mon, 31 Jul 2017 18:06:39 +0000 (21:06 +0300)]
vout: remove write-only struct member

23 months agomacosx: Remove unnecessary progressbar images
Marvin Scholz [Mon, 31 Jul 2017 16:44:57 +0000 (18:44 +0200)]
macosx: Remove unnecessary progressbar images

23 months agomacosx: Remove unnecessary volume slider images
Marvin Scholz [Mon, 31 Jul 2017 16:40:33 +0000 (18:40 +0200)]
macosx: Remove unnecessary volume slider images

23 months agoadaptive: always call get_empty
Francois Cartegnie [Wed, 19 Jul 2017 16:42:58 +0000 (18:42 +0200)]
adaptive: always call get_empty

As it should no longer freeze with audio

23 months agoinput: decoder: don't wait for aout drain on non eof
Francois Cartegnie [Fri, 28 Jul 2017 13:37:10 +0000 (15:37 +0200)]
input: decoder: don't wait for aout drain on non eof

If there's no way to wait for aout drain without
draining (on EOF), don't wait indefinitively.

23 months agoes_out: ask for decoder drain on es_out_Del
Francois Cartegnie [Fri, 28 Jul 2017 13:04:48 +0000 (15:04 +0200)]
es_out: ask for decoder drain on es_out_Del

because when the es deletion is triggered by
demuxer and not on EOF, it enters a dead loop
waiting for input_DecoderIsEmpty (p_owner->drained).
non set p_owner->draining can't trigger p_owner->drained.

23 months agodemux: ogg: handle precise seek
Francois Cartegnie [Fri, 28 Jul 2017 08:47:50 +0000 (10:47 +0200)]
demux: ogg: handle precise seek

23 months agomacosx: Add VLCRendererMenuController to POTFILES
Marvin Scholz [Mon, 31 Jul 2017 16:16:22 +0000 (18:16 +0200)]
macosx: Add VLCRendererMenuController to POTFILES

23 months agomacosx: Use VLCVolumeSlider
Marvin Scholz [Mon, 31 Jul 2017 16:09:56 +0000 (18:09 +0200)]
macosx: Use VLCVolumeSlider

23 months agomacosx: Add VLCVolumeSlider(Cell) classes
Marvin Scholz [Mon, 31 Jul 2017 16:08:23 +0000 (18:08 +0200)]
macosx: Add VLCVolumeSlider(Cell) classes

23 months agomacosx: Fix initializing the Cells in Slider subclasses
Marvin Scholz [Thu, 27 Jul 2017 11:39:39 +0000 (13:39 +0200)]
macosx: Fix initializing the Cells in Slider subclasses

23 months agomacosx: Add missing super awakeFromNib call
Marvin Scholz [Thu, 27 Jul 2017 11:10:18 +0000 (13:10 +0200)]
macosx: Add missing super awakeFromNib call

The docs for awakeFromNib say that super has to be called when
overriding awakeFromNib, this was not done in the VLCControlsBarCommon.
It should not make any difference though in this case.

23 months agorawvid: drop unused include
Tristan Matthews [Wed, 26 Jul 2017 14:37:27 +0000 (10:37 -0400)]
rawvid: drop unused include

23 months agomkv: pass a const mkv_track_t to packetize_wavpack()
Steve Lhomme [Mon, 31 Jul 2017 13:03:00 +0000 (15:03 +0200)]
mkv: pass a const mkv_track_t to packetize_wavpack()

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agomkv: use a const iterator when updating the PCR
Steve Lhomme [Mon, 31 Jul 2017 13:02:59 +0000 (15:02 +0200)]
mkv: use a const iterator when updating the PCR

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agomkv: use a destructor to release all the resources allocated in an mkv_track_t
Steve Lhomme [Mon, 31 Jul 2017 13:02:58 +0000 (15:02 +0200)]
mkv: use a destructor to release all the resources allocated in an mkv_track_t

This way we can't forget any when releasing it from various places.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agomkv: store a unique_ptr version of the mkv_track_t in the track list
Steve Lhomme [Mon, 31 Jul 2017 13:02:57 +0000 (15:02 +0200)]
mkv: store a unique_ptr version of the mkv_track_t in the track list

The next step is to provide a destructor so the tracks_map_t can deal with the
lifecycle of the mkv_track_t.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agomkv:matroska_segment: handle track init failure
Steve Lhomme [Mon, 31 Jul 2017 13:02:56 +0000 (15:02 +0200)]
mkv:matroska_segment: handle track init failure

We should not use a track that failed to be initialized properly.
The logic of the return value is inverted to be more readable. Before
it would never fail anyway.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agomkv: don't let zlib_decompress_extra() destroy the track it's given
Steve Lhomme [Mon, 31 Jul 2017 13:02:55 +0000 (15:02 +0200)]
mkv: don't let zlib_decompress_extra() destroy the track it's given

It didn't clean the es_format_t that's fully filled.
Also emit an error telling we can't handle the compression.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agomkv: let the track init its es_format_t
Steve Lhomme [Mon, 31 Jul 2017 13:02:54 +0000 (15:02 +0200)]
mkv: let the track init its es_format_t

So it's not created in a funny state.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agomkv: remove dead code
Steve Lhomme [Mon, 31 Jul 2017 13:02:53 +0000 (15:02 +0200)]
mkv: remove dead code

There's never anything in the p_data_init buffer.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
23 months agoty: avoid aliasing
Rémi Denis-Courmont [Sun, 30 Jul 2017 19:13:53 +0000 (22:13 +0300)]
ty: avoid aliasing