Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

This can often catch bugs in the plugin descriptors and modules core.

audioqueue: don't use the main CFRunLoop for the (empty) data callback, but one of AudioQueue's internal threads

Otherwise, we will encounter delays of up to 10 secs when stopping the AudioQueue

This is mainly intended to be used as a slave input, e.g.:
 # vlc video.mkv --input-slave timecode:// --timecode-fps 30/1001

ConfigLoadString allocates memory for 'size' bytes and copys a NULL terminated
string to this buffer.
ConfigSaveString instead expects the buffer to be NULL terminated but:
 - the 'size' saved to the cache is the char count
 - the string is written without terminating NULL.

So obviously ConfigLoadString does not match the behaviour of ConfigSaveString.
It seems to me that the cached module configuration is not working at all due
to comparing the module file names within CacheFind - this should never have
found a match.

I think this would need some more testing.

Regards,
Mario
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

Split it in several functions

After OpenDemux reads data using stream_Read(), it writes a '\0' to
the buffer after the newly-read data, but if the stream returned exactly
i_read_max bytes, this '\0' will end up just past the end of the allocated
psz_sdp array (see the call to realloc at the beginning of the loop).
Adjust the realloc call to allocate this one extra byte.
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

merge declaration and initial assignement

Replace macro with static inline and use bounds checking

While useful, this is not permitted by the standard specification.
This patch will waste a little bit of memory in affected allocations...

This reverts commit 5086a9ed.

When we moved the code to use the legacy method *or* the new method,
we lost the functionality that actually initialized the memory allocated
for the entire subpicture region (including the part where captions
weren't being rendered.

This was most obvious when regression testing on OSX, where it
resulted in old captions continuing to appear on screen even
after being cleared, or garbage in the render buffer.

Add a line that effectively zeros out the buffer prior to use.
Signed-off-by: Rafaël Carré <funman@videolan.org>