Re: Fixes segfault in beam quanting. (issue4339047)

[mtsolo]

There is no unquanted_x - the x span is calculated earlier upstream, and
the quanting acts purely on y values.

My guess is that Lily does not know how much room to put between stems
and just kinda gives up.  If you look at:


{
\override Voice . NoteHead #'stencil = ##f
\autoBeamOff
\repeat unfold 4 { d16 }
}

, which never touches the beaming code, you'll see the same ugly
solution of everything getting pushed to the right of the page.

I think the best solution may be to use this patch and put a note in the
docs warning users to create a tiny stencil (ie (make-filled-box '(0 .
0.001) '(0 . 001))) and then make it transparent if they want headless
stems whose noteheads don't shift stems depending on their direction.
My workaround for the current piece I'm working on has been to use
ghosted mensural notation, which places the stem in the middle of the
notehead.  So, when the stem changes direction, it does not move to the
left or right.

http://codereview.appspot.com/4339047/