vlc.git
4 years agosecuretransport: Fix compilation
Marvin Scholz [Sun, 26 Feb 2017 12:23:44 +0000 (13:23 +0100)]
securetransport: Fix compilation

4 years agotls: remove tls_Recv() and tls_Send()
Rémi Denis-Courmont [Sun, 26 Feb 2017 11:09:38 +0000 (13:09 +0200)]
tls: remove tls_Recv() and tls_Send()

4 years agochromecast: use vlc_tls_Read() and simplify
Rémi Denis-Courmont [Sun, 26 Feb 2017 11:08:58 +0000 (13:08 +0200)]
chromecast: use vlc_tls_Read() and simplify

No (dys)functional changes.

4 years agotls: reorganize and update documentation
Rémi Denis-Courmont [Sun, 26 Feb 2017 10:41:07 +0000 (12:41 +0200)]
tls: reorganize and update documentation

4 years agowin32: do not redefine SHUT_* macros, fix warnings
Rémi Denis-Courmont [Sun, 26 Feb 2017 09:58:27 +0000 (11:58 +0200)]
win32: do not redefine SHUT_* macros, fix warnings

4 years agonet: remove old redefines of getaddrinfo() macros
Rémi Denis-Courmont [Sun, 26 Feb 2017 09:56:37 +0000 (11:56 +0200)]
net: remove old redefines of getaddrinfo() macros

4 years agost: partial update
Rémi Denis-Courmont [Sun, 26 Feb 2017 08:38:41 +0000 (10:38 +0200)]
st: partial update

4 years agohttp: use vlc_tls_t for HTTP 2 connection test
Rémi Denis-Courmont [Sat, 25 Feb 2017 21:29:26 +0000 (23:29 +0200)]
http: use vlc_tls_t for HTTP 2 connection test

4 years agohttp: use vlc_tls_t for HTTP 1 connection test
Rémi Denis-Courmont [Sat, 25 Feb 2017 21:29:26 +0000 (23:29 +0200)]
http: use vlc_tls_t for HTTP 1 connection test

4 years agotls: introduce vlc_tls_SocketPair()
Rémi Denis-Courmont [Sat, 25 Feb 2017 21:28:42 +0000 (23:28 +0200)]
tls: introduce vlc_tls_SocketPair()

This creates a pair of mutually connected stream vlc_tls_t.

4 years agotls: add support for TCP Fast Open (refs #16067)
Rémi Denis-Courmont [Sat, 25 Feb 2017 21:07:16 +0000 (23:07 +0200)]
tls: add support for TCP Fast Open (refs #16067)

This only works underneath client-side TLS so far.

4 years agotls: revector
Rémi Denis-Courmont [Sat, 25 Feb 2017 20:09:05 +0000 (22:09 +0200)]
tls: revector

4 years agohttp: use vlc_tls_SocketOpenTLS()
Rémi Denis-Courmont [Sat, 25 Feb 2017 19:44:21 +0000 (21:44 +0200)]
http: use vlc_tls_SocketOpenTLS()

4 years agotls: add vlc_tls_SocketOpenTLS()
Rémi Denis-Courmont [Sat, 25 Feb 2017 19:40:45 +0000 (21:40 +0200)]
tls: add vlc_tls_SocketOpenTLS()

4 years agotls: drop unused vlc_tls_SocketOpen() parameter
Rémi Denis-Courmont [Sat, 25 Feb 2017 17:38:34 +0000 (19:38 +0200)]
tls: drop unused vlc_tls_SocketOpen() parameter

4 years agotls: drop obj and sys from vlc_tls_t
Rémi Denis-Courmont [Sat, 25 Feb 2017 17:36:24 +0000 (19:36 +0200)]
tls: drop obj and sys from vlc_tls_t

4 years agohttp: add opaque data pointer to connection structures
Rémi Denis-Courmont [Sat, 25 Feb 2017 17:07:20 +0000 (19:07 +0200)]
http: add opaque data pointer to connection structures

This is to pass a VLC object (so far for logging only).

4 years agodemux/mkv: fix playback of files with invalid A_AAC/MPEG{2, 3}/ specification
Filip Roséen [Sat, 25 Feb 2017 19:44:22 +0000 (20:44 +0100)]
demux/mkv: fix playback of files with invalid A_AAC/MPEG{2, 3}/ specification

Tracks matching the two introduced statements are invalid, but we still played
them fine prior to the refactoring (perhaps intentionally).

In order to stay somewhat back-compatible with earlier versions the two cases
are now added back (since there are actual files in the wild with the invalid
spec).

refs #4250

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agodemux/mkv: use appropriate helper for A_ALAC
Filip Roséen [Sat, 25 Feb 2017 19:21:55 +0000 (20:21 +0100)]
demux/mkv: use appropriate helper for A_ALAC

The correct helper-function is named fill_extra_data_alac, and not
fill_extra_data. The errournous usage seems to have been introduced during the
big refactoring of the demuxer.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agodemux/mkv: removed unused forward-declaration
Filip Roséen [Sat, 25 Feb 2017 19:21:54 +0000 (20:21 +0100)]
demux/mkv: removed unused forward-declaration

There is not a single reference to struct mkv_index_t within the
demuxer.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agodemux/mkv: silence warning related to struct vs class
Filip Roséen [Sat, 25 Feb 2017 19:21:53 +0000 (20:21 +0100)]
demux/mkv: silence warning related to struct vs class

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agoconfigure: use AM_CONDITIONAL outside of if statement
Jean-Baptiste Kempf [Sat, 25 Feb 2017 19:14:27 +0000 (20:14 +0100)]
configure: use AM_CONDITIONAL outside of if statement

4 years agomacOS: Rework dmg creation
Marvin Scholz [Sat, 25 Feb 2017 18:30:23 +0000 (19:30 +0100)]
macOS: Rework dmg creation

This adds the possibility to build fancy DMGs with background, dmg icon
and special icon positions using the dmgbuild python tool.

If dmgbuild is not available, it will fallback and make a normal DMG.

4 years agoconfigure: Check for dmgbuild on macOS
Marvin Scholz [Sat, 25 Feb 2017 18:22:16 +0000 (19:22 +0100)]
configure: Check for dmgbuild on macOS

4 years agomacOS: Add DMG icon and background
Doney den Ouden [Sat, 25 Feb 2017 18:17:34 +0000 (19:17 +0100)]
macOS: Add DMG icon and background

4 years agorecord: protect against arbitrary file overwrite
Rémi Denis-Courmont [Sat, 25 Feb 2017 14:38:44 +0000 (16:38 +0200)]
record: protect against arbitrary file overwrite

Setting the no-overwrite flag of the file access output ensures that
record will not indirectly trigger overwriting of an unintended file.

There are two occurences to this problem:

- While probing the muxer format, VLC uses the notoriously insecure and
  obsolescent (in POSIX.2008) tempnam() function. This leads to an
  arbitrary file overwrite vulnerability via symbolic links.
  However, the record plugin really should not need to create temporary
  files to probe a muxer format.

- While actually recording content to a permanent file, overwriting an
  existing file is not a good idea. This is presumably not a
  vulnerability insofar as the output directory belongs to the user.
  Regardless, the record plugin should ensure that the output filename
  does not already exists (e.g. by creating the file).

So basically, this is a stopgap measure.

4 years agogui/qt: info_panels: remove things related to QSpinBox
Filip Roséen [Sat, 25 Feb 2017 06:56:14 +0000 (07:56 +0100)]
gui/qt: info_panels: remove things related to QSpinBox

These are no longer required as there is nothing that uses it anymore.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agocodec/kate: remove TigerConfigurationCallback + OnConfigurationChanged (unused)
Filip Roséen [Sat, 25 Feb 2017 06:39:02 +0000 (07:39 +0100)]
codec/kate: remove TigerConfigurationCallback + OnConfigurationChanged (unused)

The last usage of these functions, TigerConfigurationCallback is the
only one that ever called OnConfigurationChanged, was removed as part
of 231b08e in May 2011.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agodemux/adaptive: Downloader: remove unused private data-member
Filip Roséen [Sat, 25 Feb 2017 06:39:01 +0000 (07:39 +0100)]
demux/adaptive: Downloader: remove unused private data-member

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agodemux/adaptive: SubSegment: remove unused private data-member
Filip Roséen [Sat, 25 Feb 2017 06:39:00 +0000 (07:39 +0100)]
demux/adaptive: SubSegment: remove unused private data-member

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agogui/qt: info_panels: remove comments containing legacy code
Filip Roséen [Sat, 25 Feb 2017 06:38:59 +0000 (07:38 +0100)]
gui/qt: info_panels: remove comments containing legacy code

These two lines were commented out by 9794062c2b1 and has remained
untouched since aug 2007.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agogui/qt: info_panels: remove comments containing legacy code
Filip Roséen [Sat, 25 Feb 2017 06:38:58 +0000 (07:38 +0100)]
gui/qt: info_panels: remove comments containing legacy code

These sections were commented out as part of 86a25b2c1b9 and
fe3beadd6b0 (2007-09), and has been unchanged ever since.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agoaccess/live555: remove unused function strempty
Filip Roséen [Sat, 25 Feb 2017 06:38:57 +0000 (07:38 +0100)]
access/live555: remove unused function strempty

The last usage of this function was removed by d8314d6.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agogui/qt: info_panels: remove function setSpinBounds (unused)
Filip Roséen [Sat, 25 Feb 2017 06:38:56 +0000 (07:38 +0100)]
gui/qt: info_panels: remove function setSpinBounds (unused)

The usage of setSpinBounds was removed in September 2007 (by
e6e2ad0572dce549b74d15ebb23727714b3590b2),

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agogui/qt: update look and feel related to extra metadata
Filip Roséen [Sat, 25 Feb 2017 06:36:17 +0000 (07:36 +0100)]
gui/qt: update look and feel related to extra metadata

These changes allow for several cosmetic/functional advantages, such
as:

 - automatically expand right-most column to fill entire width
 - automatically expand rows to the contents within them
 - scroll the table by pixel instead of by row
 - remove edit triggers (as these do not have any meaning currently)

fixes #17656

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agogui/qt: use QTableWidget instead of QTreeWidget for extra-metadata
Filip Roséen [Sat, 25 Feb 2017 06:36:16 +0000 (07:36 +0100)]
gui/qt: use QTableWidget instead of QTreeWidget for extra-metadata

In terms of applicability, it makes more sense to store the extra
metadata in a table than in a tree, especially given that we don't
support nested metadata in either case.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agogui/qt: rename ExtraMetaPanel::extraMetaTree to extraMeta
Filip Roséen [Sat, 25 Feb 2017 06:36:15 +0000 (07:36 +0100)]
gui/qt: rename ExtraMetaPanel::extraMetaTree to extraMeta

If the name of a data-member's type is present in its name,
changing the type of said data-member makes it so that the patch is
hard to follow.

These changes simply removes the unnecessary type-indication from the
data-members name, mostly to make future refactoring easier to follow.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agogui/qt: extended_panels: ChangeVFiltersString: NULL check
Filip Roséen [Sat, 25 Feb 2017 06:31:45 +0000 (07:31 +0100)]
gui/qt: extended_panels: ChangeVFiltersString: NULL check

GetVFilterType returns NULL on error, meaning that it is not
guaranteed that psz_filter_type contains what we require: this change
make sure that we abort ChangeVFiltersString if the type is unknown.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agogui/qt: extended_panels: refactor OptionFromWidgetName
Filip Roséen [Sat, 25 Feb 2017 06:31:44 +0000 (07:31 +0100)]
gui/qt: extended_panels: refactor OptionFromWidgetName

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agogui/qt: extended_panels: change return-type of ChangeFiltersString
Filip Roséen [Sat, 25 Feb 2017 06:31:43 +0000 (07:31 +0100)]
gui/qt: extended_panels: change return-type of ChangeFiltersString

By returning a QString instead of a heap-allocated c-style string we
reduce complexity, and hopefully increase correctness as there is less
manual memory-management to worry about.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agogui/qt: extended_panels: refactor ChangeFiltersString
Filip Roséen [Sat, 25 Feb 2017 06:31:42 +0000 (07:31 +0100)]
gui/qt: extended_panels: refactor ChangeFiltersString

The previous implementation was unnecessary complex, and by using the
utilities available for string manipulation in Qt we can drastically
reduce the lines of code.

These changes also fixes an issue where a user would be unable to
remove a video-filter if the command-line specified the same
video-filter twice.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agogui/qt: extended_panels: declare helper-functions as static
Filip Roséen [Sat, 25 Feb 2017 06:31:41 +0000 (07:31 +0100)]
gui/qt: extended_panels: declare helper-functions as static

These functions are only used within the translation-unit in question,
and as such it does not make sense for them to have external linkage.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agosdp: fix compilation after 84ec386
Jean-Baptiste Kempf [Sat, 25 Feb 2017 14:18:05 +0000 (15:18 +0100)]
sdp: fix compilation after 84ec386

4 years agogui/qt: simple_preferences: combine ppsz_language and ppsz_language_map
Filip Roséen [Sat, 25 Feb 2017 06:30:39 +0000 (07:30 +0100)]
gui/qt: simple_preferences: combine ppsz_language and ppsz_language_map

Having things as two separate arrays is very hard to maintain as one
entry in ppsz_language must correctly correspond to one element in
ppsz_language_text.

These changes refactors the implementation so that a single array is
used, where each element has two members so that it is easier to see
what corresponds to what.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agoaccess/dcp: AESKey::decryptyRSA: fix narrowing-conversion in case-label
Filip Roséen [Sat, 25 Feb 2017 06:28:21 +0000 (07:28 +0100)]
access/dcp: AESKey::decryptyRSA: fix narrowing-conversion in case-label

narrowing-conversions are not allowed in case-statements as of C++11,
meaning that the former implementation caused standard-compliant
compilers to emit a diagnostic on the line in question.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agoOptimize vlc_memstream_puts() with constant string
Rémi Denis-Courmont [Sat, 25 Feb 2017 13:55:39 +0000 (15:55 +0200)]
Optimize vlc_memstream_puts() with constant string

If the string given to vlc_memstream_puts() is constant, or at least
its length is constant, use vlc_memstream_write() instead and let GCC
compute the length at build-time.

4 years agomemstream: fix vlc_memstream_write() error return
Rémi Denis-Courmont [Sat, 25 Feb 2017 13:45:45 +0000 (15:45 +0200)]
memstream: fix vlc_memstream_write() error return

4 years agoRemove ifdefs for GCC version below 4.4
Rémi Denis-Courmont [Sat, 25 Feb 2017 13:37:38 +0000 (15:37 +0200)]
Remove ifdefs for GCC version below 4.4

According to INSTALL, we require GCC 4.8 already.

4 years agoxa: handle header read error, fix struct size
Rémi Denis-Courmont [Sat, 25 Feb 2017 12:08:23 +0000 (14:08 +0200)]
xa: handle header read error, fix struct size

This fixes the vlc_stream_Read() compiler warning.

As pointed out by Filip, some archs could pad the structure to a
boundary larger than 8 bytes, defeating the sizeof() value.

4 years agoxa: do not set callbacks on error
Rémi Denis-Courmont [Sat, 25 Feb 2017 12:00:04 +0000 (14:00 +0200)]
xa: do not set callbacks on error

4 years agosdp: use vlc_memstream
Rémi Denis-Courmont [Sat, 25 Feb 2017 11:14:01 +0000 (13:14 +0200)]
sdp: use vlc_memstream

4 years agomtp: use a temporary file descriptor
Rémi Denis-Courmont [Sat, 25 Feb 2017 07:55:26 +0000 (09:55 +0200)]
mtp: use a temporary file descriptor

This fixes insecure use of tempnam(), fixes a potential file leak onto
the filesystem and simplifies the code.

4 years agovlc_xml_encode: fix inverted logic
Rémi Denis-Courmont [Sat, 25 Feb 2017 08:04:36 +0000 (10:04 +0200)]
vlc_xml_encode: fix inverted logic

4 years agopacketizer: mpeg4audio: frame size depends of object type
Francois Cartegnie [Fri, 24 Feb 2017 21:34:56 +0000 (22:34 +0100)]
packetizer: mpeg4audio: frame size depends of object type

4 years agopacketizer: mpeg4audio: avoid flapping between load and adts
Francois Cartegnie [Fri, 24 Feb 2017 20:13:28 +0000 (21:13 +0100)]
packetizer: mpeg4audio: avoid flapping between load and adts

4 years agopacketizer: mpeg4audio: avoid divbyzero in date increment
Francois Cartegnie [Fri, 24 Feb 2017 20:07:41 +0000 (21:07 +0100)]
packetizer: mpeg4audio: avoid divbyzero in date increment

4 years agopacketizer: mpeg4audio: missing parsing of extension channels
Francois Cartegnie [Fri, 24 Feb 2017 17:02:08 +0000 (18:02 +0100)]
packetizer: mpeg4audio: missing parsing of extension channels

4 years agopacketizer: mpeg4audio: fail on unknown parsing
Francois Cartegnie [Fri, 24 Feb 2017 19:23:36 +0000 (20:23 +0100)]
packetizer: mpeg4audio: fail on unknown parsing

4 years agopacketizer: mpeg4audio: use refactored code
Francois Cartegnie [Fri, 24 Feb 2017 17:49:30 +0000 (18:49 +0100)]
packetizer: mpeg4audio: use refactored code

4 years agopacketizer: mpeg4audio: refactor reading audiospecificconfig
Francois Cartegnie [Fri, 24 Feb 2017 17:01:32 +0000 (18:01 +0100)]
packetizer: mpeg4audio: refactor reading audiospecificconfig

4 years agopacketizer: mpeg4audio: fix LatmGetValue
Francois Cartegnie [Fri, 24 Feb 2017 16:42:57 +0000 (17:42 +0100)]
packetizer: mpeg4audio: fix LatmGetValue

Never worked

4 years agostrings: use vlc_memstream in vlc_xml_encode()
Rémi Denis-Courmont [Fri, 24 Feb 2017 20:01:38 +0000 (22:01 +0200)]
strings: use vlc_memstream in vlc_xml_encode()

4 years agovlmshell: use vlc_memstream instead of sprintf()
Rémi Denis-Courmont [Fri, 24 Feb 2017 19:56:51 +0000 (21:56 +0200)]
vlmshell: use vlc_memstream instead of sprintf()

4 years agohttp: use vlc_memstream instead of sprintf()
Rémi Denis-Courmont [Fri, 24 Feb 2017 19:44:02 +0000 (21:44 +0200)]
http: use vlc_memstream instead of sprintf()

4 years agocpu: use vlc_memstream instead of sprintf()
Rémi Denis-Courmont [Fri, 24 Feb 2017 19:38:55 +0000 (21:38 +0200)]
cpu: use vlc_memstream instead of sprintf()

4 years agoes: use vlc_memstream instead of unchecked sprintf()
Rémi Denis-Courmont [Fri, 24 Feb 2017 19:32:29 +0000 (21:32 +0200)]
es: use vlc_memstream instead of unchecked sprintf()

4 years agoaudiounit_ios: fix mute state
Thomas Guillem [Fri, 24 Feb 2017 09:47:43 +0000 (10:47 +0100)]
audiounit_ios: fix mute state

4 years agoaudiounit_ios: factor AVAudioSession playback state
Thomas Guillem [Fri, 24 Feb 2017 09:17:07 +0000 (10:17 +0100)]
audiounit_ios: factor AVAudioSession playback state

This refactor adds more code since it adds a lot more of error checking.

4 years agoauhal/audiounit_ios: factor common code
Thomas Guillem [Thu, 23 Feb 2017 16:09:55 +0000 (17:09 +0100)]
auhal/audiounit_ios: factor common code

coreaudio_common.c/.h contains the common code for auhal.c and audiounit_ios.m.

TODO: factor StartAnalog.

4 years agoaudiounit_ios: pause can't be called not started
Thomas Guillem [Fri, 24 Feb 2017 09:03:30 +0000 (10:03 +0100)]
audiounit_ios: pause can't be called not started

4 years agoaudiounit_ios: remove unnecessary casts
Thomas Guillem [Fri, 24 Feb 2017 08:52:15 +0000 (09:52 +0100)]
audiounit_ios: remove unnecessary casts

4 years agoaudiounit_ios: fix StartAnalog error handling
Thomas Guillem [Fri, 24 Feb 2017 08:51:07 +0000 (09:51 +0100)]
audiounit_ios: fix StartAnalog error handling

Also fix a AudioUnit leak in case of error.

4 years agoaudiounit_ios: fix int/bool return mismatch
Thomas Guillem [Fri, 24 Feb 2017 08:49:26 +0000 (09:49 +0100)]
audiounit_ios: fix int/bool return mismatch

4 years agoaudiounit_ios: remove forward declaration
Thomas Guillem [Fri, 24 Feb 2017 08:47:23 +0000 (09:47 +0100)]
audiounit_ios: remove forward declaration

4 years agoaudiounit_ios: au_component doesn't need to be saved
Thomas Guillem [Thu, 23 Feb 2017 17:24:12 +0000 (18:24 +0100)]
audiounit_ios: au_component doesn't need to be saved

4 years agoauhal: au_component doesn't need to be saved
Thomas Guillem [Thu, 23 Feb 2017 15:44:13 +0000 (16:44 +0100)]
auhal: au_component doesn't need to be saved

4 years agodemux: avformat: workaround AAC in FLV incorrect timestamps (fix #8195)
Francois Cartegnie [Fri, 24 Feb 2017 10:15:13 +0000 (11:15 +0100)]
demux: avformat: workaround AAC in FLV incorrect timestamps (fix #8195)

4 years agodemux: avformat: invert flv check order
Francois Cartegnie [Fri, 24 Feb 2017 10:09:19 +0000 (11:09 +0100)]
demux: avformat: invert flv check order

we might want to override timestamps, not just
deal with video pts

4 years agognutls: require GnuTLS 3.5.0+ for gnutls_session_get_flags()
Thomas Guillem [Fri, 24 Feb 2017 08:27:47 +0000 (09:27 +0100)]
gnutls: require GnuTLS 3.5.0+ for gnutls_session_get_flags()

4 years agostream_extractor/archive: prevent null-pointer dereference
Filip Roséen [Thu, 23 Feb 2017 13:54:27 +0000 (14:54 +0100)]
stream_extractor/archive: prevent null-pointer dereference

Given that p_sys->p_obj is yet to be initialized with the value of the
obj, the implementation should use the latter (given that the former
will be zero-initialized, but nothing more).

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agom4: fix c++ check for -f flags
Filip Roséen [Thu, 23 Feb 2017 13:54:17 +0000 (14:54 +0100)]
m4: fix c++ check for -f flags

Both gcc and clang generate warnings for unsupported -f$FLAG by
default, meaning that the previous implementation would consider
unsupported flags as supported (as a warning is not an error that
fails compilation).

The addition of -Werror treats warnings as errors, and will prevent
false-positives in terms of -f$FLAG support.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
4 years agognutls: require GnuTLS 3.3.0+ for thread safety
Rémi Denis-Courmont [Thu, 23 Feb 2017 20:44:27 +0000 (22:44 +0200)]
gnutls: require GnuTLS 3.3.0+ for thread safety

4 years agognutls: enable False Start (fixes #18042)
Rémi Denis-Courmont [Thu, 23 Feb 2017 20:39:38 +0000 (22:39 +0200)]
gnutls: enable False Start (fixes #18042)

4 years agognutls: show session flags in debug
Rémi Denis-Courmont [Thu, 23 Feb 2017 20:39:21 +0000 (22:39 +0200)]
gnutls: show session flags in debug

4 years agocodec: faad: signal discontinuity to aout
Francois Cartegnie [Thu, 23 Feb 2017 18:14:42 +0000 (19:14 +0100)]
codec: faad: signal discontinuity to aout

4 years agocodec: faad: reset date on error
Francois Cartegnie [Thu, 23 Feb 2017 18:12:16 +0000 (19:12 +0100)]
codec: faad: reset date on error

4 years agocodec: faad: flush buffer on decoder flush
Francois Cartegnie [Thu, 23 Feb 2017 17:32:15 +0000 (18:32 +0100)]
codec: faad: flush buffer on decoder flush

4 years agolib: remove redundant test
Rémi Denis-Courmont [Thu, 23 Feb 2017 19:10:17 +0000 (21:10 +0200)]
lib: remove redundant test

If index < 0, then (size_t) index > SSIZE_MAX.

4 years agoftp: factor network I/O code using vlc_tls_t
Rémi Denis-Courmont [Thu, 23 Feb 2017 18:39:09 +0000 (20:39 +0200)]
ftp: factor network I/O code using vlc_tls_t

4 years agognutls: add "missing" non-blocking flag
Rémi Denis-Courmont [Wed, 22 Feb 2017 21:04:01 +0000 (23:04 +0200)]
gnutls: add "missing" non-blocking flag

Since version 3.0.0, applications are supposed to set this when
operating in non-blocking mode. At this point, this seems to only
really affect DTLS though.

4 years agoposix: assert that we do not close a corrupt FD
Rémi Denis-Courmont [Wed, 22 Feb 2017 21:32:33 +0000 (23:32 +0200)]
posix: assert that we do not close a corrupt FD

This helps detect double-close or other bugs.

4 years agotls: fix/rationalize closing the session vs closing the socket
Rémi Denis-Courmont [Thu, 23 Feb 2017 18:54:30 +0000 (20:54 +0200)]
tls: fix/rationalize closing the session vs closing the socket

4 years agotest/epg: fix sign warnings
Thomas Guillem [Thu, 23 Feb 2017 13:46:03 +0000 (14:46 +0100)]
test/epg: fix sign warnings

4 years agotest/mrl_helper: fix sign warning
Thomas Guillem [Thu, 23 Feb 2017 13:45:49 +0000 (14:45 +0100)]
test/mrl_helper: fix sign warning

4 years agodemux: fix vlc_array_count sign
Thomas Guillem [Thu, 23 Feb 2017 13:38:38 +0000 (14:38 +0100)]
demux: fix vlc_array_count sign

4 years agoinput: fix vlc_array_count sign
Thomas Guillem [Thu, 23 Feb 2017 13:38:31 +0000 (14:38 +0100)]
input: fix vlc_array_count sign

4 years agosout: fix vlc_array_count sign
Thomas Guillem [Thu, 23 Feb 2017 13:37:35 +0000 (14:37 +0100)]
sout: fix vlc_array_count sign

4 years agovout: fix sign warning
Thomas Guillem [Thu, 23 Feb 2017 13:37:25 +0000 (14:37 +0100)]
vout: fix sign warning

4 years agolib/media_list: fix sign warnings
Thomas Guillem [Thu, 23 Feb 2017 13:33:36 +0000 (14:33 +0100)]
lib/media_list: fix sign warnings

4 years agodialog: use vlc_array_index_of_item
Thomas Guillem [Thu, 23 Feb 2017 13:29:25 +0000 (14:29 +0100)]
dialog: use vlc_array_index_of_item