Steve Lhomme [Tue, 21 Mar 2017 10:27:59 +0000 (11:27 +0100)]
avcodec: remove unecessary check for thread safety on safe libavcodec builds
Technically there's not a version set after Anton's patches merge in FFmpeg yet
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Steve Lhomme [Tue, 21 Mar 2017 10:27:58 +0000 (11:27 +0100)]
contrib: ffmpeg: use a newer version of FFmpeg that is multithread safe with DXVA
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Marcel Schnirring [Tue, 21 Mar 2017 21:29:09 +0000 (22:29 +0100)]
fixed sorting playlist by album for multiple discs.
implicit title comparison when meta field is empty for both items had to be made explicit.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
KO Myung-Hun [Mon, 13 Mar 2017 04:56:31 +0000 (13:56 +0900)]
vlc_fixups: define macros for getaddrinfo() on OS/2
This fixes compilation breakage on OS/2 caused by commit
898b9dd532a73344267d65f5edff52697a7dc29c.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Filip Roséen [Thu, 23 Mar 2017 02:07:07 +0000 (03:07 +0100)]
demux/playlist: ProcessMRL: fallback to original if valid scheme
The changes introduced in
36bee77 had the unfortunate side-effect that
location-data that could not be parsed as a URL would be discarded,
for example an xspf playlist containing the line below would simply
ignore it (due to ProcessMRL rejecting it):
> dvb-t://frequency=
674000000:inversion=-1:bandwidth=8:code-rate-hp=2/3:code-rate-lp=1/2:modulation=64QAM:transmission=8:guard=1/32:
In order to fix the issue while still allowing uri-resolving for
URI-valid entities, the following changes make sure that we retain the
original string if it contains "://" and all prior characters are
valid in a scheme.
fixes: #18155
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Filip Roséen [Sat, 25 Mar 2017 03:21:35 +0000 (04:21 +0100)]
contrib/faad2: add patch to prevent crash on SCE followed by CPE
fixes: #18128
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Francois Cartegnie [Fri, 24 Mar 2017 13:47:57 +0000 (14:47 +0100)]
packetizer: h264: store mbaff flag
Francois Cartegnie [Thu, 23 Mar 2017 20:10:11 +0000 (21:10 +0100)]
demux: tta: avoid null deref on missing/failed decoder
Francois Cartegnie [Thu, 23 Mar 2017 20:09:46 +0000 (21:09 +0100)]
demux: tta: check stream_read return value
Francois Cartegnie [Thu, 23 Mar 2017 20:06:46 +0000 (21:06 +0100)]
demux: tta: check alloc for i_extra overflow
Francois Cartegnie [Thu, 23 Mar 2017 20:00:15 +0000 (21:00 +0100)]
demux: tta: don't set i_extra on failed alloc
Francois Cartegnie [Thu, 23 Mar 2017 19:53:42 +0000 (20:53 +0100)]
demux: tta: fix broken seek table
one more regression for using es_format_clean
Dennis Hamester [Wed, 22 Mar 2017 08:59:01 +0000 (09:59 +0100)]
nacl: Implement vlc_getaddr_info_i11e
Similar to other platforms, we simply defer the call to vlc_getaddrinfo;
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
Julian Scheel [Wed, 22 Mar 2017 08:58:53 +0000 (09:58 +0100)]
newlib/nacl: Provide flockfile implementation
newlib provides an flockfile implementation via sys/io.h, which is
marked as internal. As we do not have a better alternative make use of
it here, albeit being marked as internal.
Signed-off-by: Julian Scheel <julian@jusst.de>
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
Dennis Hamester [Wed, 22 Mar 2017 08:58:49 +0000 (09:58 +0100)]
android/thread: Remove monotonic clock checks
These checks are unnecessary since
6f3e18d534, because the calls to
pthread_condattr_setclock and pthread_cond_timedwait_monotonic_np have
been removed.
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
Steve Lhomme [Wed, 22 Mar 2017 15:50:51 +0000 (16:50 +0100)]
direct3d9: use point interpolation rather than none
We could use linear interpolation but in software decoding the decoder pool is
not initialized with black so we would interpolate with uninitialized (green)
lines/colons.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Wed, 22 Mar 2017 15:50:50 +0000 (16:50 +0100)]
direct3d9: fix the black background when the decoder texture is in YUV
fixes the bottom line issue in #17420
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Tue, 21 Mar 2017 02:02:20 +0000 (03:02 +0100)]
demux/mpeg: ps: fixup of
1901851
p_sys->i_first_scr is initialized to -1, denoting that no SCR has been
read from the underlying stream. This change is a simple fixup of
the changes introduced by
1901851 (as it lacks the correct comparision
in the relevant if-condition).
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Filip Roséen [Wed, 22 Mar 2017 10:04:27 +0000 (11:04 +0100)]
demux/playlist: m3u: fixup of
5cde0de
This lazy-mans method of debugging was added to identify that issue
with #18143, with that said; it has no value inside the codebase (and
was included in the patch by accident).
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Jean-Baptiste Kempf [Wed, 22 Mar 2017 19:00:54 +0000 (20:00 +0100)]
configure: fix inverted logic about Growl
Francois Cartegnie [Wed, 22 Mar 2017 15:48:35 +0000 (16:48 +0100)]
input: decoder: really fix aout_update_format check
since fmt_in.i_codec is stored once
a917dca7a1c6bfb2ce2f396f4015ce0e0398f101
Francois Cartegnie [Wed, 22 Mar 2017 15:35:29 +0000 (16:35 +0100)]
codec: lpcm: set both format and codec on update
since there's no guarantee aout_update_format is
pf_aout_format_update
Thomas Guillem [Wed, 22 Mar 2017 12:20:48 +0000 (13:20 +0100)]
audiounit_ios: add back the --spdif option
This option is for aout that can't negotiate S/PDIF. This is totally the case
for this aout.
Thomas Guillem [Wed, 22 Mar 2017 12:19:55 +0000 (13:19 +0100)]
audiounit_ios: fix cast warning
Francois Cartegnie [Wed, 22 Mar 2017 10:39:31 +0000 (11:39 +0100)]
packetizer: h264: set duration also using frame rate
Filip Roséen [Wed, 22 Mar 2017 03:40:08 +0000 (04:40 +0100)]
video_filter/blendbench: fix out-of-bound reads on ill-formed parameters
Prevent out-of-bounds access if --blendbench-base-chroma or
--blendbench-blend-chroma is ill-formed.
refs: #18116
fixes: #18146
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
Filip Roséen [Wed, 22 Mar 2017 03:40:07 +0000 (04:40 +0100)]
video_filter/blendbench: Create: add missing error-check
As blendbench.c:Filter relies on the p_sys->p_blend_image to refer to
a valid picture_t, we need to make sure that the relevant call is
successful in blendbench.c:Create (as we otherwise would have a
null-pointer dereference further down the road).
refs: #18116
fixes: #18145
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
Filip Roséen [Wed, 22 Mar 2017 03:38:35 +0000 (04:38 +0100)]
input/demux + misc/image: fix mime-type matching
These changes make sure that we query the source-stream's mime-type in
src/input/demux.c and src/misc/image.c, and not the full content-type.
The previous implementation would fail to identify the relevant
demuxer, and image chroma, if the stream's content-type contained
parameters following the mime-type.
fixes: #18144
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
Filip Roséen [Wed, 22 Mar 2017 03:38:34 +0000 (04:38 +0100)]
demux/playlist: fix mime-type matching
The previous implementation would inaccurately truncate the mime-type
check in case of parameters in the content-type string, which in turn
would lead to false-positive matches.
As the usage of CheckContentType is really meant to check the
mime-type of the content-type string, besides fixing the inaccurate
comparision, the function is also renamed to CheckMimeType.
fixes: #18143
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
Filip Roséen [Wed, 22 Mar 2017 03:38:33 +0000 (04:38 +0100)]
vlc_stream: add stream_MimeType
As there is quite common for stream-consumers to be interested in the
associated mime-type, if any, of a stream; this helper function allows
access to that directly (instead of consumers having to manually parse
the string returned by stream_ContentType).
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
Filip Roséen [Wed, 22 Mar 2017 03:38:22 +0000 (04:38 +0100)]
access/http: Control: remove redundant casts
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
Filip Roséen [Wed, 22 Mar 2017 03:38:21 +0000 (04:38 +0100)]
access/http: Control: fix STREAM_GET_SIZE
When a stream-control receives STREAM_GET_SIZE it is supposed to write
the size to an uint64_t. The previous implementation would treat the
passed pointer as int64_t, effectivelly causing undefined-behavior.
These changes fixes that.
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
Dennis Hamester [Mon, 20 Mar 2017 16:57:18 +0000 (17:57 +0100)]
Include sys/uio.h where necessary
All touched files use struct iovec without making sure sys/uio.h is
included.
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>
Rémi Denis-Courmont [Tue, 21 Mar 2017 19:41:10 +0000 (21:41 +0200)]
Qt: fix build w/o X11
Rémi Denis-Courmont [Tue, 21 Mar 2017 19:37:23 +0000 (21:37 +0200)]
oldrc: simplify quit
Hugo Beauzée-Luyssen [Tue, 21 Mar 2017 18:59:06 +0000 (19:59 +0100)]
contrib: microdns: Update sha512sum
Hugo Beauzée-Luyssen [Tue, 21 Mar 2017 17:36:08 +0000 (18:36 +0100)]
sd: Fix microdns build with 0.0.4
Steve Lhomme [Fri, 17 Mar 2017 10:39:22 +0000 (11:39 +0100)]
direct3d11: adjust the max luminance based on the source and output transfers
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Fri, 17 Mar 2017 10:20:23 +0000 (11:20 +0100)]
direct3d11: rework the full/studio range adjustement
Since we can have a display in studio range with (SPU) in full range. And since
the YUV->RGB add a studio to full range conversion we need to undo it when
displaying on a studio range swapchain.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Wed, 15 Mar 2017 15:08:24 +0000 (16:08 +0100)]
direct3d11: pick the best swapchain colorspace for the source video
The transfer manipulation is only done if the source doesn't match the display.
On systems older than Windows 10, RGB BT.709 Gamma 2.2 is always picked.
Use the ST2084 target for HLG sources as well.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Wed, 15 Mar 2017 14:53:31 +0000 (15:53 +0100)]
direct3d11: list the colorspaces supported by the swapchain
It will only give a list on Windows 10.
The DXGI_COLOR_SPACE_TYPE field is generated based on the values in the
structure to make sure the values match the DXGI value.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Thu, 16 Mar 2017 11:35:56 +0000 (12:35 +0100)]
direct3d11: only saturate before delivering the RGBA
The saturate should be done after the pixel is ready for EOTF.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Thu, 16 Mar 2017 10:46:14 +0000 (11:46 +0100)]
direct3d11: fix the HLG to Linear luminance
The standard mentions the 12.0 coefficient. The rendering problem is elsewhere.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Wed, 15 Mar 2017 17:00:21 +0000 (18:00 +0100)]
direct3d11: do the tone mapping for HLG sources too
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Wed, 15 Mar 2017 16:55:27 +0000 (17:55 +0100)]
direct3d11: add support for transfer conversion for BT.709 and sRGB sources
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Wed, 15 Mar 2017 13:13:52 +0000 (14:13 +0100)]
direct3d11: fix typos in ST2084 inverse EOTF
This was untested so far but now we have proper PQ output handled by Windows.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Mon, 20 Mar 2017 16:03:58 +0000 (17:03 +0100)]
direct3d11: log the Pixel Shader stages in debug builds
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Mon, 20 Mar 2017 16:01:43 +0000 (17:01 +0100)]
direct3d11: add a function to determine if a shader will be for RGB or YUV
In other words, if a YUV->RGB matrix conversion will be used
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Mon, 20 Mar 2017 16:05:47 +0000 (17:05 +0100)]
direct3d11: rename COLORSPACE_RGB_FULL as it's used for other things
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Steve Lhomme [Mon, 20 Mar 2017 15:44:49 +0000 (16:44 +0100)]
direct3d11: factorize the same string usage
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Hugo Beauzée-Luyssen [Tue, 21 Mar 2017 15:36:34 +0000 (16:36 +0100)]
contrib: use libmicrodns 0.0.4
Fix #17666
Francois Cartegnie [Tue, 21 Mar 2017 12:42:02 +0000 (13:42 +0100)]
packetizer: mpegvideo: align height on 32 for field pictures
Francois Cartegnie [Tue, 21 Mar 2017 12:33:13 +0000 (13:33 +0100)]
packetizer: mpegvideo: fix visible/macroblocks size
Francois Cartegnie [Tue, 21 Mar 2017 10:06:39 +0000 (11:06 +0100)]
Revert "lpcm: validate block_nb_frames in DVD/LPCM cases"
This reverts commit
8447e3d87b65c36d57a2c55d4afb717bb1945321.
Francois Cartegnie [Tue, 21 Mar 2017 10:07:54 +0000 (11:07 +0100)]
input: decoder: compare audio codec in aout_update_format
Source of undersized buffers on format change.
Since audio format is derived from fmt.i_codec, the update
path is inconsistent, not updating the aout format in case
fmt.i_codec changes (only comparing the derived audio.i_format)
Francois Cartegnie [Mon, 20 Mar 2017 17:49:00 +0000 (18:49 +0100)]
demux: h26x: set demux fps to packetizer
otherwise computed dts/pts will always be incorrect if no vui
Thomas Guillem [Mon, 20 Mar 2017 17:33:12 +0000 (18:33 +0100)]
opengl: alloc 1 extra pic when direct rendering
Since pictures are held while being rendered by the GPU.
Ref #18116
Thomas Guillem [Mon, 20 Mar 2017 17:31:00 +0000 (18:31 +0100)]
vout: splitter: fix double picture release in case of error
Ref #18116
Francois Cartegnie [Mon, 20 Mar 2017 16:18:28 +0000 (17:18 +0100)]
packetizer: h264: fix regression with recovery frames handling
No non iframe output with single NAL blocks regression.
Francois Cartegnie [Mon, 20 Mar 2017 15:24:01 +0000 (16:24 +0100)]
demux: ps: use existing video or audio as scr fallback
did not work with audio less track or misconfigured
track like PMF, and then no pcr was issued
Francois Cartegnie [Fri, 17 Mar 2017 13:21:28 +0000 (14:21 +0100)]
demux: ps: save first scr for time computation
Francois Cartegnie [Mon, 20 Mar 2017 15:13:40 +0000 (16:13 +0100)]
demux: ps: renamed time_track_id/index
avoid confusion and errors.
tracks id are different.
Thomas Guillem [Thu, 16 Mar 2017 12:09:13 +0000 (13:09 +0100)]
decoder: don't set IsEmpty while draining
If the decoder is draining, more pictures or blocks can be outputted, so wait
the drain to finish before querying outputs.
Filip Roséen [Mon, 20 Mar 2017 13:42:45 +0000 (14:42 +0100)]
demux/ty: Demux: fix comment related to "closed captions/XDS"
As a result of
dbae0e3, a comment was incorrectly split up so that it
implies that each split applies to one particular value (which is not
correct).
These changes fix that by putting both entities into a single comment,
hopefully signalling that either "0x01" or "0x02" means "closed
captions/XDS" (and not each for one).
--
Thanks for Francois Cartegnie for pointing out the mistake:
- https://mailman.videolan.org/pipermail/vlc-devel/2017-March/112291.html
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 13:26:50 +0000 (14:26 +0100)]
demux/vobsub: Control: remove redundant casts
The type yield by va_arg( args, T ) is already guaranteed to be T;
there is no need for the explicit cast.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 13:26:57 +0000 (14:26 +0100)]
demux/ty: Demux: replace if-else with switch
A switch should be easier to maintain, while also improving
readability.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 13:26:55 +0000 (14:26 +0100)]
demux/ty: Control: remove redundant casts
The type yield by va_arg( args, T ) is already guaranteed to be T;
there is no need for the explicit cast.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 13:26:52 +0000 (14:26 +0100)]
demux/mpc: remove redundant casts
The type yield by "va_arg( args, T )" is already guaranteed to be T,
there's no need for the explicit extra cast.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 13:26:46 +0000 (14:26 +0100)]
codec/subsusf: remove unused argument for CreateTextRegion
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 13:13:14 +0000 (14:13 +0100)]
codec/svcdsub: use NULL instead of integer literal in pointer assignment
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Francois Cartegnie [Mon, 20 Mar 2017 13:14:33 +0000 (14:14 +0100)]
packetizer: h264: use bottom_field flag
Francois Cartegnie [Mon, 20 Mar 2017 12:52:49 +0000 (13:52 +0100)]
packetizer: h264: store original slice type
And do not tag blocks with switching slices
Filip Roséen [Fri, 17 Mar 2017 02:36:02 +0000 (03:36 +0100)]
video_filter/deinterlace: algo_x: remove XDeint8x8Set (unused)
This function has been unused since 2011.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Fri, 17 Mar 2017 02:36:00 +0000 (03:36 +0100)]
video_filter/deinterlace: algo_x: remove median (unused)
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Fri, 17 Mar 2017 02:35:57 +0000 (03:35 +0100)]
lua/libs: variables: remove luaL_checklightuserdata (unused)
The last usage of this function was removed by
fe5df3d in March 2012.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Fri, 17 Mar 2017 02:35:53 +0000 (03:35 +0100)]
lua/intf: remove luaL_register_submodule (unused)
Last usage was removed in June 2008 as part of the changed instroduced
by
95addd57388b78fdfb6a58c90437085de8bbbc33.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Fri, 17 Mar 2017 02:26:15 +0000 (03:26 +0100)]
demux/mpeg: es: lower diagnostic during probe
A failed vlc_stream_Peek is not an error during Open as there might be
other modules that can handle the relevant input, as such the
diagnostic is now lowered from msg_Err to msg_Dbg.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 12:41:58 +0000 (13:41 +0100)]
gui/qt: remove unnecessary forward declarations in global scope
Given that these identifiers are only necessary for the data-member
declarations within struct intf_sys_t, it is preferred to simply use
elaborated type specifiers (the semantics are equivalent).
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 12:41:57 +0000 (13:41 +0100)]
gui/qt: AddonItemDelegate: use mem-initializer
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 12:41:56 +0000 (13:41 +0100)]
gui/qt: dialogs/plugins: prevent deprecated usage of QStyleOptionViewItemV4 if Qt >= 5
QStyleOptionViewItemV4 is deprecated starting from Qt5, and the
warning diagnostic when compiling with Qt5 has been bugging me for
quite some time.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 12:41:55 +0000 (13:41 +0100)]
gui/qt: actions_manager: const qualification and mem-initializer
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 12:41:54 +0000 (13:41 +0100)]
gui/qt: qt.hpp: remove unused macro VISIBLE
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 12:41:53 +0000 (13:41 +0100)]
gui/qt: qt.hpp: remove macro TOGGLEV
The macro is only used at one location, and as such there is no need
for the define to exist in qt.hpp.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 12:41:52 +0000 (13:41 +0100)]
gui/qt: main_interface: remove legacy code
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Mon, 20 Mar 2017 12:41:51 +0000 (13:41 +0100)]
gui/qt: interface_widgets: remove legacy code
The affected sections were disabled (wrapped in "#if 0") by
098dab901c1 back in July 2008 and has been untouched ever since.
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Filip Roséen [Sat, 18 Mar 2017 17:00:33 +0000 (18:00 +0100)]
control/globalhotkeys: xcb: fix p_sys->p_map population
As described in #18136, certain configurations end up causing a
double-free in xcb.c:Close due to the same value returned from
xcb_key_symbols_get_keycode appearing multiple times in p_sys->p_map.
The code responsible for clean-up assumes that every value refers to a
separate allocated resource, whereas Mapping potentially populates
p_sys->p_map with the same value several times.
These changes make sure that initialization vs clean-up is in harmony.
fixes: #18136
Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>
Hugo Beauzée-Luyssen [Mon, 20 Mar 2017 08:45:58 +0000 (09:45 +0100)]
upnp: Remove leftover debug
Francois Cartegnie [Fri, 17 Mar 2017 17:57:35 +0000 (18:57 +0100)]
packetizer: h264: use POC for computing missing pts
Francois Cartegnie [Tue, 14 Mar 2017 11:20:41 +0000 (12:20 +0100)]
packetizer: h264: decode and compute POC
Francois Cartegnie [Fri, 17 Mar 2017 17:29:29 +0000 (18:29 +0100)]
packetizer: h264: split slice parsing
Francois Cartegnie [Tue, 14 Mar 2017 11:55:52 +0000 (12:55 +0100)]
packetizer: h264_nal: parse required elements for POC
Francois Cartegnie [Fri, 17 Mar 2017 16:35:40 +0000 (17:35 +0100)]
packetizer: h264: don't pass values to parse slice
Filip Roséen [Fri, 17 Mar 2017 02:35:50 +0000 (03:35 +0100)]
demux/adaptive: FakeESOut: silence warning (unused expression)
Most compilers recognize static_cast<void>( ... ) as a mechanism to
silence warnings related to unused expressions, the previous
implementation however casted the result to void* which at all does
not have the same semantics.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Filip Roséen [Fri, 17 Mar 2017 02:35:58 +0000 (03:35 +0100)]
misc/gnutls: gnutls_ClientHandshake fix typo in dialog contents
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Filip Roséen [Fri, 17 Mar 2017 10:26:30 +0000 (11:26 +0100)]
se/archive: prevent potential call to archive_{entry, read}_free with NULL
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Filip Roséen [Fri, 17 Mar 2017 10:25:18 +0000 (11:25 +0100)]
se/archive: issue diagnostic on reset failure
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Filip Roséen [Fri, 17 Mar 2017 02:22:14 +0000 (03:22 +0100)]
se/archive: protect paths relying on p_sys->p_entry != NULL
The following paths will stumble into undefined-behavior land if
p_sys->p_entry is NULL.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Filip Roséen [Fri, 17 Mar 2017 02:22:13 +0000 (03:22 +0100)]
se/archive: add missing error-check in jump-callback
If we are unable to exit the current source-stream (be that for
whatever reason), we should not ignore the error as it /may/ have
fatal consequences down the road.
Better safe than sorry.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Filip Roséen [Fri, 17 Mar 2017 11:28:33 +0000 (12:28 +0100)]
se/archive: ExtractorOpen: add missing error-check in open
The same check is correctly implemented in archive.c:DirectoryOpen,
but is missing from ExtractorOpen - resulting in a null-pointer
dereference if a libarchive handle fails to be created.
fixes: #18133
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
Filip Roséen [Fri, 17 Mar 2017 02:22:12 +0000 (03:22 +0100)]
se/archive: do not reset source position if dead
If the libarchive_exit_cb is called due to clean-up (because we are
dead), there is no point in seeking back to zero. It might even fail
if our source module reacts to vlc_killed (such as
stream_filter/cache_read).
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>