Re: Repeat volta puzzle

[David Kastrup]

Jan-Peter Voigt <address@hidden> writes:

> Am 01.02.14 15:07, schrieb David Kastrup:
>> Try upgrading LilyPond.  Current versions get this right.  You'll also
>> need to remove the bar checks at the end of the alternatives as they are
>> _not_  placed on a proper bar boundary.
> do they?
> If I compile this example with 2.18 (after using convert-ly), the last
> measure is still too long, as long as I do not include
>
> \set Timing.measureLength = #(ly:make-moment 5/8)
>
> in the last measure of the first alternative and
> \set Timing.measureLength = #(ly:make-moment 6/8)
> at the beginning of the second alternative.
>
> (This can be shortened with commands or variables)
> Should it catch the timing and do this automatically in recent versions?

2.19.0 captures the timing at the start of the first alternative and
restores it at the end of each alternative (issue 355 in the tracker).
Obviously, that's not a good idea at the end of the last alternative.
2.19.1 fixes that (issue 3792 in the tracker).

Neither of those changes are present in 2.18.0, so probably "current
versions get this right" was a bit too optimistic: the current
_unstable_ releases get this right, and there might be some incentive to
pull both of those changes into 2.18.1.  It's not totally clear to me
that this is the best option regarding people's expectations.

-- 
David Kastrup