Fontconfig is great, but we have too many issue, notably the fact that
the fontcache isn't save correctly in some locations, notably on the
network. It also rebuilds the entire cache, each time you add ONE font.

Therefore, this commits uses Win32 EnumFontFamilies, and then gets the
font path with the Registry + concatenation of CLSID_FONTS and then back
to freetype.

However, this commit is broken in at least 3 ways:
 - it does not use the index to select the font,
 - it will break on localized fonts settings (Eastern),
 - I don't check the Script yet,

 I will try and fix in the future.

Ref and close: #4470, #4401, #4137, #3950, #3208

Some channels are sending epg events from start of the day

dtv is build automaticly and we anyway assume new enough libdvbpsi that has
pkg-config enabled. Also ts demux wasn't build automaticly previously (as
dvb was disabled by default some time ago).

 - preserve sample format from VLC, if possible,
 - preserve physical channels mapping,
 - remove a useless callback,
 - fix small race conditions with pa_*_wait() on startup,
 - cosmetic changes.

adds pl_forceplay,pl_forcepause,pl_resume
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

On every data update, I was cleaning the overlapped entries, and on the
whole list. Apologies for the O(n*n).

The telnet interface is removed. In its place, a small wrapper is used
to provide backward compatibility with --lua-config configuration.

Signed-off-by: Laurent Aimar <fenrir@videolan.org>

As for the previous commit, this seems more harmful than helpful.

Due to the UTF-16 subtitle hack present in the stream layer, this
commit will break in-memory UTF-16 subtitles character encoding if
UTF-8 autodetection was disabled and unless UTF-8 is the selected
encoding. On the other hand, this commit fixes subtitles encoding
if there is more than one subtitles stream in the whole input.

Four access plugins need access to the input at the moment. In all
cases, it would misbehave if the input (item) of the main access was
used.

attachment: failing seems like the correct thing. I don't see a use for
attachment from stream_UrlNew().

CDDA uses the input item to append child nodes for individual tracks.
This is wrong in case of stream_UrlNew(), as the input items would be
wrongly appended to the main access item. Like directory, FTP, etc,
CCDA should probably return a playlist in this particular case.

MMSH uses the input item for redirection. This would not work properly
for stream_UrlNew(): it would rewrite the main access URL instead of
the child stream URL. This failing seems actually safer. Handling the
redirection within the MMS plugin would probably be better.

VCDX uses the input item to fill disc-wide meta data from the GET_TITLE
control request. It should use GET_META instead. In any case, it should
not pollute the meta data of the main access.

Bugs have been filed for the last 3 cases. Fortunately, I doubt any of
these access plugins was ever used through stream_UrlNew().

Ideally, I guess a text SPU should be emitted instead.

It does not seem to be needed in this case.

It is not used anywhere (it was meant for use with Corba).
We have better ways to take snapshots nowadays.

The stream output is not a decendent of the input in VLC 1.2. So this
plugin does currently not work at all. This commit only removes
non-working code; it does not break or fix anything as such.

In previous versions, this code could cause crashes when #gather was
used, as the input object would change behind the back.

The sout instance is _not_ a child of the input, so this cannot work.
In practice, finding the input would always fail.