Comment #1 on issue 2683 by [email protected]: Beamlets with incorrect
orientation and uneven spacing in beamlet-test.ly
http://code.google.com/p/lilypond/issues/detail?id=2683
The test beamlet-test.ly changed with the fix to issue 2386, and apparently
the change is smaller than the threshold in the regression tester.
The fix to issue 2386 seems to intentionally include tuplets when setting
the priority to beam respecting the (time-signature) beat. Maybe we want
to include only tuplets divisible by 2 ?
I would have thought, contrary to Colin and Arnold, that the way 2.15.41
and 2.14.2 handle this is correct. If the beam were not forced across the
5-tuplets, these "simply wrong" beamlet directions would be the only
possible directions.
If there is doubt about how people will really want them to be printed in
real music, I suggest removing the 5-tuplet cases from the test suite.
Attachments:
A.png 9.2 KB
a.ly 259 bytes
