vlc.git
3 years agodemux: adaptive: fix signal/lock order
Francois Cartegnie [Thu, 26 May 2016 16:07:47 +0000 (18:07 +0200)]
demux: adaptive: fix signal/lock order

3 years agomediacodec: abort if ParseVideoExtraH264 fails
Thomas Guillem [Thu, 26 May 2016 17:48:41 +0000 (19:48 +0200)]
mediacodec: abort if ParseVideoExtraH264 fails

3 years agoadpcm: reject invalid QuickTime IMA files
Rafaël Carré [Wed, 25 May 2016 08:00:25 +0000 (10:00 +0200)]
adpcm: reject invalid QuickTime IMA files

DecodeAdpcmImaQT() can only decode up to stereo files.
Fix out of bound write.

Reported by: Patrick Coleman <blinken@gmail.com>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agocontribs: mad: Don't read past the input buffer
Hugo Beauzée-Luyssen [Wed, 25 May 2016 17:07:51 +0000 (19:07 +0200)]
contribs: mad: Don't read past the input buffer

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agodirectdraw: debug the reason a CreateSurface() fails
Steve Lhomme [Thu, 26 May 2016 16:16:01 +0000 (18:16 +0200)]
directdraw: debug the reason a CreateSurface() fails

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agodirectdraw: DDOVER_DDFX needs to be set when setting a DDOVERLAYFX
Steve Lhomme [Thu, 26 May 2016 16:16:00 +0000 (18:16 +0200)]
directdraw: DDOVER_DDFX needs to be set when setting a DDOVERLAYFX

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agodirectdraw: DDSCAPS_FRONTBUFFER is not a CAPS we can request on CreateSurface()
Steve Lhomme [Thu, 26 May 2016 16:15:59 +0000 (18:15 +0200)]
directdraw: DDSCAPS_FRONTBUFFER is not a CAPS we can request on CreateSurface()

effectively reverting 351cd5a85ac220b68ba7a1ef3c87ca81d77b7d26

Close #16905

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agodirectdraw: we also set the CAPS of the surface we create
Steve Lhomme [Thu, 26 May 2016 16:15:58 +0000 (18:15 +0200)]
directdraw: we also set the CAPS of the surface we create

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agoNEWS: missing xiphqt update
Francois Cartegnie [Thu, 26 May 2016 14:23:03 +0000 (16:23 +0200)]
NEWS: missing xiphqt update

3 years agoQt: fix the Win32/Qt5 tooltip focus/raising issue
Benoit du Payrat [Thu, 26 May 2016 12:49:12 +0000 (14:49 +0200)]
Qt: fix the Win32/Qt5 tooltip focus/raising issue

The time tooltip no longer puts the window in front on Windows, when
compiling VLC with Qt5

Close #12084 #8532

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agologger: android: display thread id
Thomas Guillem [Thu, 26 May 2016 11:58:59 +0000 (13:58 +0200)]
logger: android: display thread id

3 years agomediacodec: remove unused var
Thomas Guillem [Thu, 26 May 2016 11:57:14 +0000 (13:57 +0200)]
mediacodec: remove unused var

3 years agomkv: fix playback/seek for files where first block has PTS > 0
Filip Roséen [Thu, 26 May 2016 11:35:55 +0000 (13:35 +0200)]
mkv: fix playback/seek for files where first block has PTS > 0

Some files do not have their first cluster at PTS=0, meaning that we
might spend an eternity looking for such PTS (even though it is not at
all present in the file) when starting playback (seeking to 0).

This patch fixes the issue by always assigning a seekpoint for PTS=0 pointing
at the first cluster position. Instead of using the cluster-timecode, a
hardcoded value of 0 is used (so that every file will have a seekpoint at that
location).

Signed-off-by: Thomas Guillem <thomas@gllm.fr>
3 years agowin32:thread: keep INFINITE condition wait intact
Steve Lhomme [Thu, 26 May 2016 11:02:26 +0000 (13:02 +0200)]
win32:thread: keep INFINITE condition wait intact

Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
3 years agowin32:thread: SleepEx and WaitForMultipleObjectsEx have a millisecond precision
Steve Lhomme [Thu, 26 May 2016 11:02:25 +0000 (13:02 +0200)]
win32:thread: SleepEx and WaitForMultipleObjectsEx have a millisecond precision

Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
3 years agointerrupt: QueueUserAPC is always allowed on Win10
Hugo Beauzée-Luyssen [Thu, 26 May 2016 11:29:37 +0000 (13:29 +0200)]
interrupt: QueueUserAPC is always allowed on Win10

3 years agothread: fix android build
Thomas Guillem [Thu, 26 May 2016 11:37:16 +0000 (13:37 +0200)]
thread: fix android build

vlc_thread_id is already defined in src/linux/thread.c

3 years agomediacodec: refactor ParseVideoExtra
Thomas Guillem [Thu, 26 May 2016 09:47:41 +0000 (11:47 +0200)]
mediacodec: refactor ParseVideoExtra

3 years agomediacodec: fix MPEG-4 IFrame parsing
Thomas Guillem [Thu, 26 May 2016 09:42:41 +0000 (11:42 +0200)]
mediacodec: fix MPEG-4 IFrame parsing

3 years agomediacodec: refactor video block processing
Thomas Guillem [Thu, 26 May 2016 09:37:34 +0000 (11:37 +0200)]
mediacodec: refactor video block processing

Each codecs has its own callback.

3 years agomediacodec: p_block can change after a realloc
Thomas Guillem [Thu, 26 May 2016 09:03:38 +0000 (11:03 +0200)]
mediacodec: p_block can change after a realloc

This was not happening because blocks are allocated with a pre-padding of 32
bits. But this behavior could change in function of the packetizer or demuxer.

3 years agomediacodec: change pf_on_new_block argument
Thomas Guillem [Thu, 26 May 2016 07:58:57 +0000 (09:58 +0200)]
mediacodec: change pf_on_new_block argument

Because we may have to realloc blocks inside this callback (corner case).

3 years agoomxil: all video decoders need a valid size
Thomas Guillem [Thu, 26 May 2016 08:56:31 +0000 (10:56 +0200)]
omxil: all video decoders need a valid size

Therefore, this is not a quirk.

3 years agomediacodec: refactor StartMediaCodec
Thomas Guillem [Thu, 26 May 2016 08:55:47 +0000 (10:55 +0200)]
mediacodec: refactor StartMediaCodec

CSD and video size checks are now done in StartMediaCodec.

Call ParseVideoExtra from Open(). This was useless to call it from
StartMediaCodec since p_dec.fmt_in.p_extra couldn't change during the lifetime
of the module.

3 years agomessages: add thread ID in message meta-data
Rémi Denis-Courmont [Wed, 25 May 2016 20:00:50 +0000 (23:00 +0300)]
messages: add thread ID in message meta-data

3 years agothreads: add vlc_thread_self() and vlc_thread_id()
Rémi Denis-Courmont [Wed, 25 May 2016 19:51:04 +0000 (22:51 +0300)]
threads: add vlc_thread_self() and vlc_thread_id()

3 years agolinux: move specific files to dedicated directory
Rémi Denis-Courmont [Wed, 25 May 2016 18:42:10 +0000 (21:42 +0300)]
linux: move specific files to dedicated directory

3 years agoupnp: Use m_ prefix for class members instead of _ suffix
Hugo Beauzée-Luyssen [Wed, 25 May 2016 17:03:19 +0000 (19:03 +0200)]
upnp: Use m_ prefix for class members instead of _ suffix

3 years agoupnp: Work around some lack of limit understanding from a few servers
Hugo Beauzée-Luyssen [Wed, 25 May 2016 16:59:05 +0000 (18:59 +0200)]
upnp: Work around some lack of limit understanding from a few servers

3 years agoupnp: Always specify ContainerID instead of ObjectID
Hugo Beauzée-Luyssen [Wed, 25 May 2016 16:58:39 +0000 (18:58 +0200)]
upnp: Always specify ContainerID instead of ObjectID

Additionnaly, ensure we pass "0" instead of an empty string for the root
object

3 years agoupnp: Don't drop potential options from the control URL
Hugo Beauzée-Luyssen [Wed, 25 May 2016 16:54:06 +0000 (18:54 +0200)]
upnp: Don't drop potential options from the control URL

3 years agoupnp: Specify <StartingIndex> after <Filter>
Hugo Beauzée-Luyssen [Wed, 25 May 2016 16:16:53 +0000 (18:16 +0200)]
upnp: Specify <StartingIndex> after <Filter>

As not doing so prevents MS Media Sharing servers from understanding the
request.

3 years agointerrupt: fix typo
Thomas Guillem [Wed, 25 May 2016 14:02:31 +0000 (16:02 +0200)]
interrupt: fix typo

3 years agoaccess: dvb: scan fix illegal access (cid #1362073)
Francois Cartegnie [Wed, 25 May 2016 13:49:35 +0000 (15:49 +0200)]
access: dvb: scan fix illegal access (cid #1362073)

since defined values have changed

3 years agolibvlc: media: add missing break
Thomas Guillem [Wed, 25 May 2016 13:10:39 +0000 (15:10 +0200)]
libvlc: media: add missing break

3 years agommdevice: fix crash when unplugging a device
Thomas Guillem [Thu, 19 May 2016 12:28:19 +0000 (14:28 +0200)]
mmdevice: fix crash when unplugging a device

Restarting the audio output is not enough when a device is unplugged. Indeed,
the Start function will use the old (and invalid) device.
To fix this issue, re-select the default device in case of error.

3 years agommdevice: handle flush errors
Thomas Guillem [Thu, 19 May 2016 12:27:03 +0000 (14:27 +0200)]
mmdevice: handle flush errors

3 years agommdevice: support WASAPI device endpoint activation events
Thomas Guillem [Tue, 17 May 2016 16:36:40 +0000 (12:36 -0400)]
mmdevice: support WASAPI device endpoint activation events

Some devices expose more than one endpoint such as speakers and headphone
output. Disconnecting headphones does not remove the endpoint but instead marks
it as being in unplugged state.

Respond to the appropriate activation events in order to enable/disable
endpoints correctly when they are added/removed from set of "active" endpoints.

Fix #15966

Original patch by Sean McGovern <gseanmcg@gmail.com>

3 years agommdevice: be informative about device state changes
Thomas Guillem [Tue, 17 May 2016 16:36:39 +0000 (12:36 -0400)]
mmdevice: be informative about device state changes

Original patch by Sean McGovern <gseanmcg@gmail.com>

3 years agothreads: fix android build
Thomas Guillem [Wed, 25 May 2016 08:44:19 +0000 (10:44 +0200)]
threads: fix android build

3 years agoXiph metadata: fix bug causing redudant "extra" metadata
Filip Roséen [Tue, 17 May 2016 18:18:49 +0000 (20:18 +0200)]
Xiph metadata: fix bug causing redudant "extra" metadata

The IF_EXTRACT_FMT macro is to be used in a if-else-tree, the usage of
IF_EXTRACT prior to the following if-statement did however cause the
single if-else-tree to be split up into two (causing the fallback-else
at the end to be unconditionally hit, even if another branch had already
handled the data).

If "TITLE=" metadata was found, we would still end up inside the
fallback since the previous implementation of IF_EXTRACT_FMT would start
a new if-else-tree.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agoWin64: use high-entropy ASLR
Jean-Baptiste Kempf [Mon, 23 May 2016 08:16:06 +0000 (10:16 +0200)]
Win64: use high-entropy ASLR

And set the image base to >4GB

This should improve the application security

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agoinput decoder: clean-up of DecoderProcess
Filip Roséen [Mon, 23 May 2016 13:57:53 +0000 (15:57 +0200)]
input decoder: clean-up of DecoderProcess

The functionality remains the same, though the code should be easier to
follow and maintain.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agoinput decoder: removed misleading documentation
Filip Roséen [Mon, 23 May 2016 13:57:55 +0000 (15:57 +0200)]
input decoder: removed misleading documentation

The function does not return a value, as such the relevant comment is very misleading.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agocontrib: gcrypt: do not use the API via LoadLibrary on winstore builds
Steve Lhomme [Tue, 24 May 2016 12:11:35 +0000 (14:11 +0200)]
contrib: gcrypt: do not use the API via LoadLibrary on winstore builds

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agocontrib: gpg-error: fix UNICODE usage
Steve Lhomme [Tue, 24 May 2016 09:59:06 +0000 (11:59 +0200)]
contrib: gpg-error: fix UNICODE usage

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agowin32: reorder the code
Steve Lhomme [Mon, 23 May 2016 15:28:09 +0000 (17:28 +0200)]
win32: reorder the code

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agovideotoolbox: drop decoder restart code on fmt change
Francois Cartegnie [Tue, 24 May 2016 17:09:37 +0000 (19:09 +0200)]
videotoolbox: drop decoder restart code on fmt change

now done by core.
seems current annexb regression where sps block is dropped
after restart (packetizer changes fmt, dec is restarted by es,
then restarts by itself dropping current block which should be sps/pps)

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agoAdd CineForm fourcc
Emeric Grange [Tue, 24 May 2016 18:42:54 +0000 (20:42 +0200)]
Add CineForm fourcc

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agohttp: fix h2output test
Rémi Denis-Courmont [Tue, 24 May 2016 19:50:38 +0000 (22:50 +0300)]
http: fix h2output test

3 years agotest: robustify timer test
Rémi Denis-Courmont [Tue, 24 May 2016 19:09:34 +0000 (22:09 +0300)]
test: robustify timer test

3 years agothreads: add vlc_cond_timedwait_daytime()
Rémi Denis-Courmont [Mon, 23 May 2016 19:03:45 +0000 (22:03 +0300)]
threads: add vlc_cond_timedwait_daytime()

This separates waiting on UTC deadline from waiting on monotonic clock.
That way, storing the reference clock in vlc_cond_t is no longer
necessary. There was only one single condition variable using the UTC
clock (in VLM), so vlc_cond_t.clock was rather wasteful.

3 years agovlm: store dates as time_t / seconds
Rémi Denis-Courmont [Mon, 23 May 2016 18:28:39 +0000 (21:28 +0300)]
vlm: store dates as time_t / seconds

There was no point multiplying and dividing by CLOCK_FREQ all the time.
VLM never had sub-second precision.

Also use div() as appropriate.

3 years agolibvlc: media: remove flexible array member
Thomas Guillem [Fri, 20 May 2016 08:33:23 +0000 (10:33 +0200)]
libvlc: media: remove flexible array member

Because ISO C++ has no notation of flexible arrays and this a public header.

3 years agolibvlc: remove enum in union events
Thomas Guillem [Tue, 24 May 2016 09:19:48 +0000 (11:19 +0200)]
libvlc: remove enum in union events

Since it's not not universally binary compatible.

3 years agolibvlc: merge libvlc_MediaParsedStatus and libvlc_MediaParsedCharged events
Thomas Guillem [Tue, 24 May 2016 09:19:42 +0000 (11:19 +0200)]
libvlc: merge libvlc_MediaParsedStatus and libvlc_MediaParsedCharged events

Because having two differents events for the same objective is way too
confusing (libvlc_MediaParsedStatus was recently added by me).

libvlc_MediaParsedCharged is now always sent after a call to
libvlc_media_parse_*() (this was not the case if the pre-parsing was skipped).

The core vlc_InputItemPreparsedChanged event is now only attached from the
media player. This allows the libvlc_MediaParsedCharged event to be sent when a
media is parsed from a media player.

/!\ Behavior change in libvlc API /!\

The libvlc_event_t.u.media_parsed_changed.new_status value changed, this is now
a libvlc_media_parsed_status_t. Before this patch, this value was always set to
true, even in case of parse failure/skipped. Now this value can be skipped (1),
failed (2) or done (3). There should be no changes for users that were checking
if new_status was true, since this events was called for these 3 cases (that
are all > 0).

3 years agomacosx: fix packaging if lua is not being build
Felix Paul Kühne [Mon, 23 May 2016 20:48:34 +0000 (22:48 +0200)]
macosx: fix packaging if lua is not being build

3 years agoshare: disable vlc.desktop compilation on Darwin
Felix Paul Kühne [Mon, 23 May 2016 20:30:22 +0000 (22:30 +0200)]
share: disable vlc.desktop compilation on Darwin

3 years agoASF: add extrametadata
Jean-Baptiste Kempf [Mon, 23 May 2016 16:42:57 +0000 (18:42 +0200)]
ASF: add extrametadata

3 years agoupnp: Handle subtitle attributes
Hugo Beauzée-Luyssen [Mon, 23 May 2016 15:45:10 +0000 (17:45 +0200)]
upnp: Handle subtitle attributes

3 years agomkv: implemented (and enabled) matroska_segment_c::FastSeek
Filip Roséen [Sat, 21 May 2016 00:18:02 +0000 (02:18 +0200)]
mkv: implemented (and enabled) matroska_segment_c::FastSeek

The logic for precise and inprecise seeking is the same with the
addition that precise seeking should use the exact desired timestamp
to use with ES_OUT_SET_NEXT_DISPLAY_TIME, whereas inprecise seeking will
use the timestamp associated with the closest index frame.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agomkv: reimplemented matroska_segment_c::Seek
Filip Roséen [Sat, 21 May 2016 00:18:01 +0000 (02:18 +0200)]
mkv: reimplemented matroska_segment_c::Seek

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agomkv: reimplemented functionality provided by SegmentSeeker
Filip Roséen [Sat, 21 May 2016 00:18:00 +0000 (02:18 +0200)]
mkv: reimplemented functionality provided by SegmentSeeker

There is sadly a trade-off between having "perfect" seeking for all
tracks (ie. the preroll should start at the latest possible location for
all tracks within the mkv), and having seeking which is fast.

This patch changes the functionality in SegmentSeeker so that it
optimizes seeking for the most important tracks. If, during the scan for
seekpoints for these tracks, we stumble upon seekpoints for other tracks
- use them, if not; simply yield appropriate seekpoints for the priority
tracks.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agomkv: removed old seeking
Filip Roséen [Sat, 21 May 2016 00:17:59 +0000 (02:17 +0200)]
mkv: removed old seeking

Remove legacy logic in regards of seeking.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agomkv: fixed SegmentSeeker::get_search_areas
Filip Roséen [Sat, 21 May 2016 00:17:58 +0000 (02:17 +0200)]
mkv: fixed SegmentSeeker::get_search_areas

Fixed bug where a too big of an area would be returned from the
function, instead of just the area that is really not searched
(indexed).

This will, of course, speed up seeking since we will not rescan an area
where we already know everything there is to know.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agomkv: store priority tracks
Filip Roséen [Fri, 20 May 2016 21:41:31 +0000 (23:41 +0200)]
mkv: store priority tracks

When seeking, we should prioritize tracks depending on what type of ES
they are; this patch addes a data-member to matroska_segment_c that will
keep track of which tracks are more important than other.

It also adds logic to ComputePriorityTracks that will initialize this
data-member with the appropriate track-ids.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agomkv: add cue position regardless of partial support or not
Filip Roséen [Sat, 21 May 2016 00:17:56 +0000 (02:17 +0200)]
mkv: add cue position regardless of partial support or not

Even though there are currently some type of cues that we cannot handle,
we should add the aprox. location denoted by such cues as a seekpoint to
optimize seeking.

Adding the location with Seekpoint::DISABLED means that we will never
seek directly to the position, but it can be used as a hint when looking
for an appropriate keyframe during the preparation for seeking.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agomkv: add cluster position info from cues
Filip Roséen [Sat, 21 May 2016 00:17:55 +0000 (02:17 +0200)]
mkv: add cluster position info from cues

Every cue in the matroska format is based on the location of the cluster
in which it is part of, as such we should add this information (cluster
position) to the code responsible for seeking.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agoupnp: Fix potential free of uninitialized pointer
Hugo Beauzée-Luyssen [Mon, 23 May 2016 11:34:24 +0000 (13:34 +0200)]
upnp: Fix potential free of uninitialized pointer

3 years agomediacodec: add mpeg2 support
Duncan McNamara [Mon, 23 May 2016 07:34:44 +0000 (09:34 +0200)]
mediacodec: add mpeg2 support

Signed-off-by: Thomas Guillem <thomas@gllm.fr>
3 years agomediacodec: add vc1/wmv3 support
Duncan McNamara [Mon, 23 May 2016 07:34:43 +0000 (09:34 +0200)]
mediacodec: add vc1/wmv3 support

Signed-off-by: Thomas Guillem <thomas@gllm.fr>
3 years agoWMV strings are UTF-16 not latin1
Jean-Baptiste Kempf [Sun, 22 May 2016 18:52:03 +0000 (20:52 +0200)]
WMV strings are UTF-16 not latin1

3 years agoWMV: define 24 as the HEADER_SIZE define and use it
Jean-Baptiste Kempf [Sun, 22 May 2016 18:41:28 +0000 (20:41 +0200)]
WMV: define 24 as the HEADER_SIZE define and use it

3 years agoWMV: use defines instead of magic numbers
Jean-Baptiste Kempf [Sun, 22 May 2016 18:35:45 +0000 (20:35 +0200)]
WMV: use defines instead of magic numbers

3 years agoNSIS: improve the installation of folders
Daniel Amm [Fri, 22 Jan 2016 18:01:51 +0000 (19:01 +0100)]
NSIS: improve the installation of folders

Do not search for all matching folders/files, but only the specified folder.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agoNSIS: fix warnings
Daniel Amm [Fri, 22 Jan 2016 17:57:46 +0000 (18:57 +0100)]
NSIS: fix warnings

This fixes the following warnings (and resulting problems):

unknown variable/constant "{LANG_TRADITIONAL_CHINESE}" detected, ignoring (vlc.win32.nsi:552)
unknown variable/constant "" detected, ignoring (languages\romanian.nsh:51)
unknown variable/constant "{WEERGAWE}" detected, ignoring (languages\afrikaans.nsh:51)
unknown variable/constant "{LANG_TRADITIONAL_CHINESE}" detected, ignoring (vlc.win32.nsi:1097)
label "Line219-Remove" not used
label "Bengali" not used
label "Punjabi" not used

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agocontrib: mad: don't create dummy files, use foreign automake
Steve Lhomme [Fri, 20 May 2016 16:02:22 +0000 (18:02 +0200)]
contrib: mad: don't create dummy files, use foreign automake

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
3 years agodemux: mp4: more insane rtp cleanup
Francois Cartegnie [Sat, 21 May 2016 15:01:37 +0000 (17:01 +0200)]
demux: mp4: more insane rtp cleanup

clueless crap unchecked everywhere,
even demux returning function error codes

3 years agodemux: mp4: fix oob writes
Francois Cartegnie [Sat, 21 May 2016 14:32:42 +0000 (16:32 +0200)]
demux: mp4: fix oob writes

3 years agodemux: mp4: fix and simplify rtp stuff
Francois Cartegnie [Sat, 21 May 2016 14:04:29 +0000 (16:04 +0200)]
demux: mp4: fix and simplify rtp stuff

3 years agodemux: mp4: rtp setup: add checks and simplify tsro
Francois Cartegnie [Sat, 21 May 2016 13:26:55 +0000 (15:26 +0200)]
demux: mp4: rtp setup: add checks and simplify tsro

3 years agodemux: mp4: rtp setup: missing checks and simplify tssy
Francois Cartegnie [Sat, 21 May 2016 13:23:13 +0000 (15:23 +0200)]
demux: mp4: rtp setup: missing checks and simplify tssy

3 years agodemux: mp4: rtpsetup: add missing checks on tims
Francois Cartegnie [Sat, 21 May 2016 13:21:00 +0000 (15:21 +0200)]
demux: mp4: rtpsetup: add missing checks on tims

3 years agodemux: mp4: move also RTP setup stuff to static
Francois Cartegnie [Sat, 21 May 2016 13:17:06 +0000 (15:17 +0200)]
demux: mp4: move also RTP setup stuff to static

3 years agodemux: mp4: do not expose hint box readers
Francois Cartegnie [Sat, 21 May 2016 13:11:37 +0000 (15:11 +0200)]
demux: mp4: do not expose hint box readers

3 years agodemux: mp4: remove majors declared as atoms
Francois Cartegnie [Sat, 21 May 2016 13:03:39 +0000 (15:03 +0200)]
demux: mp4: remove majors declared as atoms

3 years agodemux: mp4: add support for XiphQt flac
Francois Cartegnie [Sat, 21 May 2016 12:37:15 +0000 (14:37 +0200)]
demux: mp4: add support for XiphQt flac

Untested.
Should be sufficient the way it worked
with vorbis (see xiphqt's decoder).

3 years agodemux: mp4: use defined atom instead of fourcc
Francois Cartegnie [Sat, 21 May 2016 12:36:36 +0000 (14:36 +0200)]
demux: mp4: use defined atom instead of fourcc

3 years agoRevert "access: dtv: try some default bandwidth values"
Francois Cartegnie [Fri, 20 May 2016 20:47:04 +0000 (22:47 +0200)]
Revert "access: dtv: try some default bandwidth values"

This reverts commit 7b1ac7887307d860372dbcab1c84a23517fe17b7.

3 years agoaccess: dvb: fix uninitialized members
Francois Cartegnie [Fri, 20 May 2016 15:37:49 +0000 (17:37 +0200)]
access: dvb: fix uninitialized members

3 years agoaccess: dtv: fix incorrect symbols
Francois Cartegnie [Fri, 20 May 2016 14:11:35 +0000 (16:11 +0200)]
access: dtv: fix incorrect symbols

refactoring error

3 years agodemux wmv: support metadata from the extended content description
Jean-Baptiste Kempf [Fri, 20 May 2016 16:53:07 +0000 (18:53 +0200)]
demux wmv: support metadata from the extended content description

3 years agosubtitles: fix potential invalid pointer use
Steve Lhomme [Fri, 20 May 2016 16:12:01 +0000 (18:12 +0200)]
subtitles: fix potential invalid pointer use

Signed-off-by: Thomas Guillem <thomas@gllm.fr>
3 years agocontribs: gnutls: Disable hardware acceleration on 64bits winrt builds
Hugo Beauzée-Luyssen [Fri, 20 May 2016 14:48:16 +0000 (16:48 +0200)]
contribs: gnutls: Disable hardware acceleration on 64bits winrt builds

It uses some forbidden functions

3 years agocontrib: shout: no thread for android
Thomas Guillem [Fri, 20 May 2016 09:26:46 +0000 (11:26 +0200)]
contrib: shout: no thread for android

3 years agocontrib: vpx: fix android build
Thomas Guillem [Fri, 20 May 2016 09:58:48 +0000 (11:58 +0200)]
contrib: vpx: fix android build

3 years agocontrib: vpx: disable optimizations when boostrapped that way
Steve Lhomme [Fri, 20 May 2016 09:58:34 +0000 (11:58 +0200)]
contrib: vpx: disable optimizations when boostrapped that way

Signed-off-by: Thomas Guillem <thomas@gllm.fr>
3 years agocontrib: vpx: don't build dependencies
Steve Lhomme [Fri, 20 May 2016 09:14:29 +0000 (11:14 +0200)]
contrib: vpx: don't build dependencies

Signed-off-by: Thomas Guillem <thomas@gllm.fr>
3 years agocontrib: vpx: fix ARM compilation
Steve Lhomme [Fri, 20 May 2016 09:14:28 +0000 (11:14 +0200)]
contrib: vpx: fix ARM compilation

Signed-off-by: Thomas Guillem <thomas@gllm.fr>
3 years agodvb: scan: improve scan vals and parsers
Francois Cartegnie [Tue, 26 Apr 2016 16:15:17 +0000 (18:15 +0200)]
dvb: scan: improve scan vals and parsers

parses missing descriptor and scanlist values.
outputs scan results to new format.