]> git.sesse.net Git - vlc/log
vlc
11 years agoncurses: remove cargo cultivated vlc_object_alive()
Rémi Denis-Courmont [Sun, 24 Mar 2013 16:45:07 +0000 (18:45 +0200)]
ncurses: remove cargo cultivated vlc_object_alive()

The ncurses thread is stuck in getch(). Checking vlc_object_alive()
does not solve anything.

11 years agoyadif: use uintptr_t when casting pointer to integer
Rémi Denis-Courmont [Sun, 24 Mar 2013 16:38:17 +0000 (18:38 +0200)]
yadif: use uintptr_t when casting pointer to integer

11 years agoadjust: removed unused values
Rémi Denis-Courmont [Sun, 24 Mar 2013 16:36:31 +0000 (18:36 +0200)]
adjust: removed unused values

11 years agovisual: remove abusive VLC_USED qualifier
Rémi Denis-Courmont [Sun, 24 Mar 2013 16:33:03 +0000 (18:33 +0200)]
visual: remove abusive VLC_USED qualifier

Confusingly enough, aout_filter_RequestVout() is also used to release
the vout. In that case, ignoring the return value is perfectly fine.

11 years agolua: remove unused parameter
Rémi Denis-Courmont [Sun, 24 Mar 2013 16:31:20 +0000 (18:31 +0200)]
lua: remove unused parameter

11 years agolua: remove unused argument
Rémi Denis-Courmont [Sun, 24 Mar 2013 16:30:11 +0000 (18:30 +0200)]
lua: remove unused argument

11 years agolua: remove dead code
Rémi Denis-Courmont [Sun, 24 Mar 2013 16:15:31 +0000 (18:15 +0200)]
lua: remove dead code

11 years agomacosx: do registration of the log callback slightly later
David Fuhrmann [Sun, 24 Mar 2013 14:07:31 +0000 (15:07 +0100)]
macosx: do registration of the log callback slightly later

This fixes mysterious interface crashes during startup, because of wrong / corrupt
initialization of several objects.

11 years agoQt4: grammar
Rémi Denis-Courmont [Sun, 24 Mar 2013 13:22:38 +0000 (15:22 +0200)]
Qt4: grammar

11 years agoQt4: remove isFirstRun parameter
Rémi Denis-Courmont [Sun, 24 Mar 2013 12:35:47 +0000 (14:35 +0200)]
Qt4: remove isFirstRun parameter

This is totally redundant. qt-privacy-ask does the exact same thing.

11 years agoQt4: clarify PNAP dialog (fixes #8100)
Rémi Denis-Courmont [Sun, 24 Mar 2013 12:28:47 +0000 (14:28 +0200)]
Qt4: clarify PNAP dialog (fixes #8100)

Also remove #ifdef from translated string since gettext can obviously
not deal with that.

11 years agoavcodec: set default channel layout on encoding
Ilkka Ollakka [Sun, 24 Mar 2013 09:39:15 +0000 (11:39 +0200)]
avcodec: set default channel layout on encoding

11 years agoavcodec: on encoding, try to match vlc and avcodec formats
Ilkka Ollakka [Sun, 24 Mar 2013 11:42:32 +0000 (13:42 +0200)]
avcodec: on encoding, try to match vlc and avcodec formats

This can avoid unnecessary format conversion.

11 years agoaout: fix typo in comment
David Fuhrmann [Sat, 23 Mar 2013 16:53:22 +0000 (17:53 +0100)]
aout: fix typo in comment

11 years agoCorrect widespread misuse of the word 'subtitle'
Edward Wang [Thu, 14 Mar 2013 15:32:24 +0000 (15:32 +0000)]
Correct widespread misuse of the word 'subtitle'

Subtitle is the adjective describing what kind of text. So "subtitle text" is an adjective phrase describing what kind of encoding.

For example, we don't listen to "audios tracks", program "videos codecs", or open "playlists files". Therefore, use the correct form of the word subtitle for the context.

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>
11 years agoios: fix application crash due permitted opengl calls in background (fixes #8139)
Gleb Pinigin [Sat, 23 Mar 2013 10:31:39 +0000 (10:31 +0000)]
ios: fix application crash due permitted opengl calls in background (fixes #8139)

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>
11 years agoios: remove repetitive parameters
Gleb Pinigin [Sat, 23 Mar 2013 10:31:38 +0000 (10:31 +0000)]
ios: remove repetitive parameters

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>
11 years agoauhal: rewrite TimeGet function
David Fuhrmann [Sat, 23 Mar 2013 13:53:30 +0000 (14:53 +0100)]
auhal: rewrite TimeGet function

This fixes stuttering when resuming playback after pause.
Buffer size is increased to avoid packet drops.

Also fixes #8223

11 years agoauhal: fix memory corruption crash when device list changes (refs #8286)
David Fuhrmann [Sun, 10 Mar 2013 17:39:00 +0000 (18:39 +0100)]
auhal: fix memory corruption crash when device list changes (refs #8286)

11 years agoauhal: enable TimeGet for S/PDIF
David Fuhrmann [Sat, 9 Mar 2013 17:58:26 +0000 (18:58 +0100)]
auhal: enable TimeGet for S/PDIF

11 years agoNEWS: mention iTunes pausing support
David Fuhrmann [Sat, 9 Mar 2013 17:52:54 +0000 (18:52 +0100)]
NEWS: mention iTunes pausing support

11 years agomacosx: change shortcut for error panel
David Fuhrmann [Sat, 9 Mar 2013 17:40:19 +0000 (18:40 +0100)]
macosx: change shortcut for error panel

Cmd+^+M cannot be triggered, so change to Cmd+Alt+M.

11 years agosndio: fix typos
Alexandre Ratchov [Sat, 23 Mar 2013 11:41:32 +0000 (13:41 +0200)]
sndio: fix typos

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
11 years agoomxil_vout: Set pixel aspect ratio, for the broadcom renderer
Martin Storsjö [Fri, 22 Mar 2013 16:43:11 +0000 (18:43 +0200)]
omxil_vout: Set pixel aspect ratio, for the broadcom renderer

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil_vout: Use the actual pixel size as frame size
Martin Storsjö [Fri, 22 Mar 2013 10:06:52 +0000 (12:06 +0200)]
omxil_vout: Use the actual pixel size as frame size

Don't use the display size, which has taken the pixel aspect
ratio into account.

This still doesn't handle actually displaying anything with
non-square pixels correctly, though.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil: Clear the stride and slice height before setting port params on raspberry pi
Martin Storsjö [Fri, 22 Mar 2013 09:06:13 +0000 (11:06 +0200)]
omxil: Clear the stride and slice height before setting port params on raspberry pi

This fixes cases where the OMX decoder previously errored out,
when decoding videos with odd resolutions.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil_vout: Generalize and fix the stride/slice height calculations
Martin Storsjö [Fri, 22 Mar 2013 08:52:06 +0000 (10:52 +0200)]
omxil_vout: Generalize and fix the stride/slice height calculations

On the raspberry pi, the stride is filled in with the right value,
while the slice height is left untouched (and needs to be rounded
up to a multiple of 16).

This fixes rendering of more uncommon resolutions.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil_vout: Simplify code with a local variable
Martin Storsjö [Fri, 22 Mar 2013 08:50:27 +0000 (10:50 +0200)]
omxil_vout: Simplify code with a local variable

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoQt: ExtendedDialog: Add sync to config checkbox for audio filters
Francois Cartegnie [Thu, 21 Mar 2013 18:30:57 +0000 (19:30 +0100)]
Qt: ExtendedDialog: Add sync to config checkbox for audio filters
(fix #7924)

11 years agoQt: Equalizer: add presets icons
Francois Cartegnie [Thu, 21 Mar 2013 15:51:15 +0000 (16:51 +0100)]
Qt: Equalizer: add presets icons

11 years agoQt: Equalizer: rework (fix #7923)
Francois Cartegnie [Thu, 21 Mar 2013 14:21:54 +0000 (15:21 +0100)]
Qt: Equalizer: rework (fix #7923)

11 years agocore: gather message initialization code
Rémi Denis-Courmont [Thu, 21 Mar 2013 21:28:23 +0000 (23:28 +0200)]
core: gather message initialization code

Also remove libvlc_int_t.b_color, and (re)print greetings when starting
a new log.

11 years agocontrib: bootstrap: exit on error
Rémi Denis-Courmont [Thu, 21 Mar 2013 19:06:07 +0000 (21:06 +0200)]
contrib: bootstrap: exit on error

11 years agoequalizer: increase precision of frequency centers in debug output
Ronald Wright [Wed, 20 Mar 2013 23:10:01 +0000 (18:10 -0500)]
equalizer: increase precision of frequency centers in debug output

Currently, one of the debug lines in the equalizer module prints out center
frequency values truncated (not rounded) to zero decimal places, and when VLC
is set to use ISO center frequencies, this causes some level of disagreement
between the depiction of the truncated frequencies in the module debug output
and the depiction of the rounded frequencies in the Qt GUI labels.  The 62.5 Hz
ISO center frequency is one of these frequencies, where we have 62 Hz shown in
the debug output and 63 Hz shown in the Qt GUI.  This patch fixes the issue by
displaying the center frequencies in the debug output as real values rounded to
the nearest hundredth.

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
11 years agomacosx: make sure the fspanel's volume slider knob is entirely displayed and not...
Felix Paul Kühne [Thu, 21 Mar 2013 16:02:02 +0000 (17:02 +0100)]
macosx: make sure the fspanel's volume slider knob is entirely displayed and not cut off

11 years agomacosx: resize the retina bitmaps for the progressions knobs so AppKit doesn't do...
Felix Paul Kühne [Thu, 21 Mar 2013 15:59:36 +0000 (16:59 +0100)]
macosx: resize the retina bitmaps for the progressions knobs so AppKit doesn't do stupid things

This prevents the knob to be blurred in the bright interface style because the retina bitmap wasn't twice as large as the regular one

11 years agoaout: remove what should be an unneeded update of "mute"
Erwan Tulou [Wed, 20 Mar 2013 00:16:11 +0000 (01:16 +0100)]
aout: remove what should be an unneeded update of "mute"

if aout_MuteSet returns successfully, the "mute" variables at the aout
and playlist levels have already been updated via a call to aout_MuteReport.

This late update was a hack to make things work when a audio output driver
forgets to call aout_MuteReport. So this patch may bring regression for those
audio output drivers that still forget to call aout_MuteReport and yet
terminate successfully. This is likely to occur in corner cases, for instance
when no stream is available as was the case for pulse.c

11 years agopulse: ensure a report is sent if no stream is available
Erwan Tulou [Thu, 21 Mar 2013 14:31:20 +0000 (15:31 +0100)]
pulse: ensure a report is sent if no stream is available

11 years agolibvlc: pass log message directly
Rémi Denis-Courmont [Wed, 20 Mar 2013 17:20:56 +0000 (19:20 +0200)]
libvlc: pass log message directly

Let the callback format message context infos however it wants.

11 years agolibvlc: add functions to query log message "item" data
Rémi Denis-Courmont [Wed, 20 Mar 2013 17:18:59 +0000 (19:18 +0200)]
libvlc: add functions to query log message "item" data

Also make provisions for file name and line number in the future.

11 years agoQt4: replace awkward word 'keep-pressed'
Edward Wang [Fri, 15 Mar 2013 16:47:50 +0000 (12:47 -0400)]
Qt4: replace awkward word 'keep-pressed'

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
11 years agoWin32: use fputws instead of ToANSI+fputs
Jean-Baptiste Kempf [Tue, 19 Mar 2013 23:10:06 +0000 (00:10 +0100)]
Win32: use fputws instead of ToANSI+fputs

11 years agohttplive: handle better relative urls
Ilkka Ollakka [Wed, 20 Mar 2013 06:41:50 +0000 (08:41 +0200)]
httplive: handle better relative urls

if path starts with / it's related to root of server and not
current path.

Fixes: #8321
11 years agoskins2: use the "mute" variable
Erwan Tulou [Tue, 19 Mar 2013 22:23:27 +0000 (23:23 +0100)]
skins2: use the "mute" variable

11 years agoRemove unmaintained switcher module
Rafaël Carré [Tue, 19 Mar 2013 21:29:21 +0000 (22:29 +0100)]
Remove unmaintained switcher module

11 years agomacosx: switch fspanel icon set to use the brighter 'highlight' graphics in idle...
Felix Paul Kühne [Tue, 19 Mar 2013 21:08:30 +0000 (22:08 +0100)]
macosx: switch fspanel icon set to use the brighter 'highlight' graphics in idle mode, which is better to see.

just for testing purposes, for now

11 years agoWin32: High-Priority will not work in MetroMode
Jean-Baptiste Kempf [Tue, 19 Mar 2013 16:47:44 +0000 (17:47 +0100)]
Win32: High-Priority will not work in MetroMode

11 years agolivehttp: fix padding to work with iOS/Android
Ilkka Ollakka [Tue, 19 Mar 2013 13:34:01 +0000 (15:34 +0200)]
livehttp: fix padding to work with iOS/Android

Don't pad individual ts-packets, but pad just segments.

11 years agomacosx: fixed traffic light drawing in dark UI mode
Felix Paul Kühne [Tue, 19 Mar 2013 13:10:08 +0000 (14:10 +0100)]
macosx: fixed traffic light drawing in dark UI mode

11 years agoLook for inet_ntop/inet_pton in ws2_32
Rafaël Carré [Mon, 18 Mar 2013 16:49:19 +0000 (17:49 +0100)]
Look for inet_ntop/inet_pton in ws2_32

11 years agoopengl: Handle subpicture textures being aligned to power-of-two dimensions
Martin Storsjö [Thu, 14 Mar 2013 10:20:37 +0000 (12:20 +0200)]
opengl: Handle subpicture textures being aligned to power-of-two dimensions

This fixes subpicture textures on ES1.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoopengl: Split texture uploading to a separate function
Martin Storsjö [Tue, 12 Mar 2013 13:35:35 +0000 (15:35 +0200)]
opengl: Split texture uploading to a separate function

This allows to reuse the repacking logic for subpictures as well,
fixing display of subpictures on ES2, where GL_UNPACK_ROW_LENGTH
isn't available.

Also avoid mallocing/freeing the temporary buffer for each upload,
but keep one buffer in the context.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoopengl: Handle pixel pitch properly when repacking textures for upload
Martin Storsjö [Thu, 14 Mar 2013 09:40:26 +0000 (11:40 +0200)]
opengl: Handle pixel pitch properly when repacking textures for upload

This makes repacking of textures with multiple bytes per pixel (RGBA)
work properly.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agomacosx: fix fspanel pause button button alignment
Felix Paul Kühne [Mon, 18 Mar 2013 22:24:16 +0000 (23:24 +0100)]
macosx: fix fspanel pause button button alignment

11 years agomacosx/Resources: flip fs volume slider knob so it appears correctly on screen
Felix Paul Kühne [Mon, 18 Mar 2013 22:19:43 +0000 (23:19 +0100)]
macosx/Resources: flip fs volume slider knob so it appears correctly on screen

while a bit ugly, this is more efficient than doing so over and over again at runtime

11 years agomacosx: fix fullscreen volume slider knob alignment
Felix Paul Kühne [Mon, 18 Mar 2013 21:42:02 +0000 (22:42 +0100)]
macosx: fix fullscreen volume slider knob alignment

11 years agomacosx: fix time slider knob alignment in bright mode
Felix Paul Kühne [Mon, 18 Mar 2013 21:40:27 +0000 (22:40 +0100)]
macosx: fix time slider knob alignment in bright mode

11 years agolibvlc: simplify and clean up external logging API (fixes #8129)
Rémi Denis-Courmont [Mon, 18 Mar 2013 21:07:53 +0000 (23:07 +0200)]
libvlc: simplify and clean up external logging API (fixes #8129)

 - Scope the callback to its instance.
 - Avoid leaking the layout of an internal data structure.
 - Future-proof passing of extra informations through an opaque pointer.

11 years agoRename msg_item_* to vlc_log_* (cosmetic)
Rémi Denis-Courmont [Mon, 18 Mar 2013 20:38:05 +0000 (22:38 +0200)]
Rename msg_item_* to vlc_log_* (cosmetic)

(This is needed for the next change.)

11 years agoskins2: remove unused data
Rémi Denis-Courmont [Mon, 18 Mar 2013 20:28:36 +0000 (22:28 +0200)]
skins2: remove unused data

11 years agomacosx: use vlc_LogSet()
Rémi Denis-Courmont [Mon, 18 Mar 2013 20:27:23 +0000 (22:27 +0200)]
macosx: use vlc_LogSet()

11 years agoQt4: use vlc_LogSet()
Rémi Denis-Courmont [Mon, 18 Mar 2013 20:25:44 +0000 (22:25 +0200)]
Qt4: use vlc_LogSet()

11 years agoncurses: use vlc_LogSet()
Rémi Denis-Courmont [Mon, 18 Mar 2013 20:22:10 +0000 (22:22 +0200)]
ncurses: use vlc_LogSet()

11 years agologger: use vlc_LogSet()
Rémi Denis-Courmont [Mon, 18 Mar 2013 20:20:29 +0000 (22:20 +0200)]
logger: use vlc_LogSet()

11 years agolibvlc: simplify logging callback
Rémi Denis-Courmont [Mon, 18 Mar 2013 20:13:56 +0000 (22:13 +0200)]
libvlc: simplify logging callback

Remove static data, and keep each callback correctly scoped to its
LibVLC instance.

11 years agolibvlc: deinit actions in reverse order of init
Rémi Denis-Courmont [Mon, 18 Mar 2013 19:41:40 +0000 (21:41 +0200)]
libvlc: deinit actions in reverse order of init

11 years agoomxil: Generalize code for hardcoded component/role mappings
Martin Storsjö [Mon, 18 Mar 2013 20:45:55 +0000 (22:45 +0200)]
omxil: Generalize code for hardcoded component/role mappings

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoh264_nal: Support resuming bitstream format conversion across buffers
Martin Storsjö [Mon, 18 Mar 2013 19:39:20 +0000 (21:39 +0200)]
h264_nal: Support resuming bitstream format conversion across buffers

This is necessary in order to pass one single H264 packet in
multiple OMX buffers to the decoder.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil: Feed input data in multiple buffers if needed
Martin Storsjö [Mon, 18 Mar 2013 19:37:15 +0000 (21:37 +0200)]
omxil: Feed input data in multiple buffers if needed

This fixes decoding of higher resolution content on the raspberry
pi, where the decoder defaults to 80 KB input buffers and doesn't
react if trying to enlarge them in
ImplementationSpecificWorkarounds().

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoh264_nal: Fix a boundary check in the H264 bitstream conversion routine
Martin Storsjö [Mon, 18 Mar 2013 19:13:58 +0000 (21:13 +0200)]
h264_nal: Fix a boundary check in the H264 bitstream conversion routine

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agolibvlc: remove potential logging from within config_LoadCmdLine()
Rémi Denis-Courmont [Mon, 18 Mar 2013 17:29:24 +0000 (19:29 +0200)]
libvlc: remove potential logging from within config_LoadCmdLine()

11 years agolibvlc: remove redundant object name hack
Rémi Denis-Courmont [Mon, 18 Mar 2013 17:18:24 +0000 (19:18 +0200)]
libvlc: remove redundant object name hack

This was redundant since 1207a3698201c2835ea432f5cb98cac8e4c42f91.

11 years agolibvlc: initialize message settings early enough
Rémi Denis-Courmont [Mon, 18 Mar 2013 17:09:55 +0000 (19:09 +0200)]
libvlc: initialize message settings early enough

That is to say, before any message gets logged.

Unfortunately, since the modules bank uses logging, and provides the
configuration, the message settings cannot be stored in the
configuration file. All messages could be removed from the bank but
that would hurt debugging in some cases :-/

11 years agolibgcrypt 1.5.1 was released
Rafaël Carré [Mon, 18 Mar 2013 16:52:37 +0000 (17:52 +0100)]
libgcrypt 1.5.1 was released

11 years agoomxil: Add a configure option for enabling raspberry pi specific omx code
Martin Storsjö [Fri, 15 Mar 2013 15:35:49 +0000 (17:35 +0200)]
omxil: Add a configure option for enabling raspberry pi specific omx code

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil_vout: Set broadcom specific parameters
Martin Storsjö [Mon, 11 Mar 2013 14:52:28 +0000 (16:52 +0200)]
omxil_vout: Set broadcom specific parameters

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil: Add a broadcom/raspberry pi specific OMX header
Martin Storsjö [Mon, 11 Mar 2013 14:52:23 +0000 (16:52 +0200)]
omxil: Add a broadcom/raspberry pi specific OMX header

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil: Use the alternative 64 bit integer ABI on the raspberry pi
Martin Storsjö [Mon, 11 Mar 2013 14:52:21 +0000 (16:52 +0200)]
omxil: Use the alternative 64 bit integer ABI on the raspberry pi

This changes how some structs are laid out.

Most of OMX API implementation on the raspberry pi runs on the
VideoCore processor and not on the normal ARM thar runs the
userland. Some OMX structs are passed over to this processor as
whole blocks, so the OMX ABI used in the userland is tied to be the
same as the one used by the VideoCore firmware.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil: Signal OMX version 1.1.2 when targeting raspberry pi
Martin Storsjö [Mon, 11 Mar 2013 14:52:13 +0000 (16:52 +0200)]
omxil: Signal OMX version 1.1.2 when targeting raspberry pi

The broadcom OMX IL core requires this in order to respond to
requests.

In principle, this shouldn't probably break other OMX implementations,
unless they are picky about it in the same way, requiring version
1.1.1, but still keeping it behind ifdefs just in case.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil: Hardcode known roles for OMX broadcom components on raspberry pi
Martin Storsjö [Mon, 11 Mar 2013 14:52:11 +0000 (16:52 +0200)]
omxil: Hardcode known roles for OMX broadcom components on raspberry pi

This OMX core does not list any roles for the components.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil: Try to load the broadcom OMX IL core and host library on the raspberry pi
Martin Storsjö [Mon, 11 Mar 2013 14:51:47 +0000 (16:51 +0200)]
omxil: Try to load the broadcom OMX IL core and host library on the raspberry pi

The host library is intentionally not unloaded, since it starts
background threads when initialized, and the deinit function does
not stop the threads. Therefore, this library cannot be unloaded
once loaded and initialized.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil: Add an omxil based video output module
Martin Storsjö [Mon, 11 Mar 2013 14:52:20 +0000 (16:52 +0200)]
omxil: Add an omxil based video output module

This is kept under modules/codes/omxil due to sharing much code
with the omxil codec.

This is only tested with the OMX.broadcom.video_render component
on the Raspberry Pi so far, which in itself doesn't advertise any
roles at all, so this could possibly use the "iv_renderer.yuv.overlay"
role instead of plan "iv_renderer".

But it should be possible to generalize this further and adapt it to
any OMX IL video renderer component.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil: Factorize event queue handling to a separate struct
Martin Storsjö [Mon, 11 Mar 2013 14:52:18 +0000 (16:52 +0200)]
omxil: Factorize event queue handling to a separate struct

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil: Factorize event debug printing to a separate function
Martin Storsjö [Mon, 11 Mar 2013 14:52:16 +0000 (16:52 +0200)]
omxil: Factorize event debug printing to a separate function

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoomxil: Split omx core management to a separate file
Martin Storsjö [Mon, 11 Mar 2013 14:51:41 +0000 (16:51 +0200)]
omxil: Split omx core management to a separate file

The point of the omx core management is to try to load and init the
OMX core once, even if there's multiple e.g. decoder instances. The
reason for this is that there are OMX cores where the deinit function
will deinitialize the whole core on the first deinit call, even if
init has been called twice.

If this file is to be built as part of multiple separate shared
plugins, each of the plugins will initialize the OMX core separately,
which could cause issues with such OMX cores.

Since the second plugin that uses the OMX core, an OMX vout plugin,
isn't used automatically, and the only OMX implementation it currently
is tested to actually work with (the Broadcom OMX core on Raspberry
Pi) doesn't have issues with being uninitialized multiple times,
there's no need to jump through hoops to fix this theoretical issue
right now though.

For reference, if we really want to make sure the omx core is loaded
and initialized only once, the plugins need to be loaded with
RTLD_GLOBAL and all the variables in omxil_core.c need to be marked
with __attribute__ ((visibility ("default"))). But that's generally
not an option, and shouldn't be necessary right now at least.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agoaccess: add vnc/rfb: update LIST/NEWS
Francois Cartegnie [Mon, 18 Mar 2013 13:46:13 +0000 (14:46 +0100)]
access: add vnc/rfb: update LIST/NEWS

11 years agoaccess: add rfb/vnc viewonly module.
Francois Cartegnie [Sun, 17 Feb 2013 19:32:20 +0000 (20:32 +0100)]
access: add rfb/vnc viewonly module.

Uses libvnc to provide 8-32bpp viewer, compression control, rfb & x509
auth, desktop resize handling and vnc://host:port uri.

11 years agoi420->rv32 neon: improve scheduling & registers usage
Sébastien Toque [Tue, 12 Mar 2013 17:20:00 +0000 (18:20 +0100)]
i420->rv32 neon: improve scheduling & registers usage

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
11 years agoopengl: Set the color to full alpha in the non-shader draw path
Martin Storsjö [Mon, 18 Mar 2013 08:30:41 +0000 (10:30 +0200)]
opengl: Set the color to full alpha in the non-shader draw path

The color is set when drawing subpictures, but wasn't set when
drawing the main texture.

For some reason, the current color was reset between each
display call in some environments, making this issue not to be
visible.

Signed-off-by: Martin Storsjö <martin@martin.st>
11 years agolivehttp: Fix crash on when not using encryption
Ilkka Ollakka [Mon, 18 Mar 2013 07:05:35 +0000 (09:05 +0200)]
livehttp: Fix crash on when not using encryption

Call CryptKey only if we actually are encrypting stuff

Fixes: #8315
11 years agolivehttp: tell if playlist is vod
Ilkka Ollakka [Sun, 17 Mar 2013 15:45:46 +0000 (17:45 +0200)]
livehttp: tell if playlist is vod

Mark EXT-X-PLAYLSIT-TYPE:EVENT if it's still under generating and at the
end, mark it as VOD in cases where we keep all the segments and those
are in playlist.

Closes: #6919
11 years agox264: use x264 macro to get both B-frame types
Ilkka Ollakka [Fri, 15 Mar 2013 11:34:06 +0000 (13:34 +0200)]
x264: use x264 macro to get both B-frame types

11 years agoavcodec: flag blocks as corrupted if avpacket tells it is
Ilkka Ollakka [Fri, 15 Mar 2013 11:33:14 +0000 (13:33 +0200)]
avcodec: flag blocks as corrupted if avpacket tells it is

11 years agoavcodec: mark SP/SI/BI pictures
Ilkka Ollakka [Fri, 15 Mar 2013 10:32:11 +0000 (12:32 +0200)]
avcodec: mark SP/SI/BI pictures

11 years agoavformat: don't offset dts/pts to start from 0, avformat should nowadays handle that...
Ilkka Ollakka [Fri, 15 Mar 2013 10:07:33 +0000 (12:07 +0200)]
avformat: don't offset dts/pts to start from 0, avformat should nowadays handle that just fine

11 years agoavformat: don't mark buffer as need to be writable, as we don't write into it directly
Ilkka Ollakka [Fri, 15 Mar 2013 10:06:36 +0000 (12:06 +0200)]
avformat: don't mark buffer as need to be writable, as we don't write into it directly

11 years agoNEWS: mention livehttp encryption
Ilkka Ollakka [Sun, 17 Mar 2013 14:32:09 +0000 (16:32 +0200)]
NEWS: mention livehttp encryption

11 years agolivehttp: support static IV-use on encryption
Ilkka Ollakka [Sun, 17 Mar 2013 14:41:04 +0000 (16:41 +0200)]
livehttp: support static IV-use on encryption

generate-iv option will tell vlc to generate random IV to encrypt whole
stream instead of using segment number as IV.

11 years agolivehttp: stream encryption
Rafaël Carré [Sat, 16 Mar 2013 13:26:09 +0000 (15:26 +0200)]
livehttp: stream encryption

Uses segment number for IV and takes key-file where to read key for
encryption.

Modified-by: Ilkka Ollakka <ileoo@videolan.org>
11 years agohttplive: support absolute URI-locations for key-retrieval
Ilkka Ollakka [Sun, 17 Mar 2013 11:26:33 +0000 (13:26 +0200)]
httplive: support absolute URI-locations for key-retrieval