Re: Issue 34 in lilypond: Grace synchronization

[Keith OHara]

David Kastrup <dak <at> gnu.org> writes:

> Rational grace_part_.set_infinite (-1)
> 
> should do.  When make-moment is called omitting the grace part of the
> argument list, this should become the default.
> 

Another interesting question, I think, is whether get_length() should
return 1/2 G-inf or 1/2 G0 for a two-beat-long segment of music.

Looking through the current code, I keep changing my mind on which 
would be least surprising.  The class Rational would need a bit of 
work to support math on infinities.

So far the proposed encoding for grace timing looks promising. 
There are many comparisons for ==  equality between Moments, though,
so I imagine inserting lots of temporary tests for equality between
the main_part_s as scaffolding during the change.