#14 Wrong cell_time accumulation in multiangle features

[elupus]

the time calculated for cell_start and pg_start in a
cell change event is wrong for any PGC that contains
multiple angles. all cells are counted when time is
added up, not only one of the angles.

same issue exist in the seeking code.

adding code like this to the loops that accumulate time
from cells fixes this for most dvd's.

  /* only count the first angle cell */
  if(  state->pgc->cell_playback[i].block_type ==

BLOCK_TYPE_ANGLE_BLOCK
&& state->pgc->cell_playback[i].block_mode !=
BLOCK_MODE_FIRST_CELL )
continue;