On 2016/12/30 16:59:55, Dan Eble wrote:
On 2016/12/29 17:47:46, david.nalesnik wrote: > Please review. Thanks!
I haven't reviewed the code, but the description makes me wonder, is
the current
behavior not a bug? Gould says "In tradition engraving," which
appears to be
her term for the style using church rests, "a rest bar takes the space
the
graphic symbol needs ..." (Behind Bars, p.565). Although she goes on
to
recommend against "tradition engraving," it seems that the bars should
be sized
to the rests, not the other way around.
Is it important to preserve the old algorithm at all? If not, call it
a bug and
throw it out. If it is important to preserve, could it be handled
with a
conversion rule so that your new algorithm can be the default?
I've had time to think about this, and I think my comments below will be a more direct response to your concerns! Bars *are* sized to fit the rest, but this issue is complicated by available space. When ragged-right is #t, bars fit the rest, though an effort is made to give extra space to rests proportional to their duration (the property space-increment). Gould is in favor of this practice. Setting spacing-increment to 0 gives you measures with only enough space to contain their rests: { \override MultiMeasureRest.expand-limit = 101 \override MultiMeasureRest.space-increment = 0 \compressFullBarRests R1*3 R1*5 R1*7 R1*9 R1*101 } But when the contents of the measure aren't the only factors in deciding measure size, what happens to the church rest? The engraving examples I cited allow flexibility in the spacing of church rest symbols. When the rest is too diffuse, however, it becomes harder to take in. So I believe that the bug here is that LilyPond does not limit the stretchability of the rest. The third patch set addresses that. https://codereview.appspot.com/319910043/ _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel