. removed the bug that prevents from ./configuring with libdvb-0.5.x

  (PID 5 first bytes were not set in some cases)

. update version number and ChangeLog

2 bug fixes in the stable branch ! Ready for a 0.5.4 release ?

Description of the bug:

The PAT gives the list of the available programs in the current stream and
the PID of their PMT (called PMT_PID).

Usually, each program has a different PMT_PID.
When streaming, we just need to put a filter on this PID and send the packets
just as other packets. PAT is the only table we need to change.

But the 13818-1 standard don't tell anything about that. Some multiplexers
use the same PMT_PID for more than one program ... The pervious filter is
not enough !
For each PMT with the right PID, we need to check to which program number
it refers and send the packet only if it's the right one.
Moreover, we can't just send it like we received it ... we need to regenerate
it in order to keep the continuity counter.


By the way, the libdvbpsi should not complain about receiving a PMT with a
program number that doesn't match. It should only ignore quietly that's the
reason why I removed the message in my last commit (it uses lost of CPU to
display it when streaming a lot of channels).


With last libdvbpsi,this commit should fix all that. Please try it and
tell me if it works well.

   The 4 unused bytes are not any more duplicated
   TODO - Just the same for SPU.

. Update version number

vls.cfg : added an example of chapter & title selection
          removed the "programcount" parameter, which should now be useless

* configure.in: fix a bug in DVB configuration when driver headers are not
  make installed
* dvbinput.cpp: fix a bug when streaming more than one DVB channel

preparing the release:
* update Changelog
* update version number to 0.5.2

. Added the loop on bad files. (after a Read Error)
. Updated  version numbers & ChangeLog for 0.5.1 release

  . Changed Runtime Libraries to 'Multi-Threaded DLL' instead of
  'single-threaded'.

  This is very likely to correct the famous Windows Bug ! I found
  successful result doing that on vls 0.3.3 and vls-0.4.0.

  I could not test on vls 0.5.0, since it has some other problems
  (unrelated).

  Thank you gibalou !

No more access to the driver (suppress all the ioctl)
You can find the lib at:
  http://www.metzlerbros.org/dvb/

--
nitrox
<private> "The school should fix its --ttl option too ;-)" </private>

'm_pDir' member depends on 'HAVE_OPENDIR' and not 'WIN32'

  . Lazy correction of a bug when using command line parameters.

  . Fix to my previous commit

  . Some compilation fixes
  . Removed the BUGGY_VLC flag (see discussion in vls-devel)

  . Backported fixes in the stable branch.

dvbinput: Fixes for multiple dvb card using driver 1.0.0 (http://linuxtv.org/)
configure.in: Change library location (not any more in DVB/libdvb)
              use --with-libdvb if you did not make install the libdvb

. dvdreader now compiles
. still a little problem with getopt compilation

. dvbinput: multiple streams support courtesy of Andrew de Quincey
. dvbready: timeout workaround courtesy of Andrew de Quincey

I have created 0.5.0 branch, I would like to release vls-0.5.0 by the end
of the week... ;-)

. ProgramCount is not needed anymore for local input

. You can define local input just with:
    BEGIN "xxx"  where xxx is a number
     ...
    END
  The 'local1' input is implicitly declared

. Compilation warnings fixed

Description in RPM and Debian packages updated.

vls.spec : preparing the next release.

. specific compilation flags for modules (especially for dvbinput)

. corrected the SendMethod kludge (default value is 0)

. Updated version number for the release

. Updated ChangeLog

. Changed default debug mode to disable
   (run ./configure --enable-debug to enable it)

. Fixed DVB-T and multiples streams thanks to Peter Martin and
   Andrew de Quincey

for your three lines of code :p).
* Couldn't find a way to have C++ plug-ins on Darwin, though.