the toc doesn't notice the real pregap audio data, so marks the HTOA
as silence, and puts all following audio data in one file.
Update .cue file to reflect this; ie, create a cue file that misses the
HTOA
Fix a subtle bug in our CDDB disc id calculation.
The length of the audio should be calculated as the delta
between leadout and start already converted (and truncated)
to seconds.
* morituri/test/bloc.cue:
* morituri/test/test_image_toc.py:
Fix up tests for this.