#4684 Kneed beam group should keep ‘main beams’

[Simon Albrecht]

Reported by Gilberto Agostinho.

If the number of beams in a kneed group changes, the result may become very difficult to read: Lily doesn’t keep one main beam and place the side beams sensibly, but instead decides from distance to the note head on each new note, with odd results:

%% staircase.ly %%%%%%%
\version "2.19.32"
\autochange {
   c'''64[ c c32
   \repeat unfold 6 {c'''64 c c32}
   c'''64 c c32]
}

See the other two attachments for other examples of current output vs. better arrangements, as well as the mailing list threads (user and bug).
Elaine Gould’s esteemed opinion is referenced in http://lilypond.1069038.n5.nabble.com/Kneed-beams-and-changes-of-staff-tp184339p184472.html.

[Gilberto Agostinho]

Simon, the link for the examples of Elaine Gould is broken, I think you left a period after the ".html" ending.

Also, as I wrote in the mailing lists, I am offering a bounty of 30 EUR for anyone who fixes this, but unfortunately the bounty is valid only until the end of February 2016 (which is the deadline for my project).

[Simon Albrecht]

That with the link was a markdown issue – fixed now.
And the bounty is properly indicated :-)

[Gilberto Agostinho]

If anyone is interested: I managed to extend the deadline of my master's composition a bit further, so in case anyone is up for tackling this issue, the bounty will be active until the 15th of April.