issue #1493: Problem with horizontal beams

issue #1493: Problem with horizontal beams

file: lily/beam-quanting.cc

In spite of infinite damping, due to numeric precision problems
(rounding/adding) the resulting beams arbitrarily still had a
remaining non-zero slope, sometimes even in the "wrong" direction.

In good TeX manner and in line with concaveness (already implemented),
we'll use a threshold value of 10000 for forcing horizontal beams.

[lemzwerg, 2018-02-10 12:40:30]

LGTM!

[Dan Eble, 2018-02-10 13:26:35]

LGTM, though I have a couple of observations.

https://codereview.appspot.com/337560043/diff/1/lily/beam-quanting.cc
File lily/beam-quanting.cc (right):

https://codereview.appspot.com/337560043/diff/1/lily/beam-quanting.cc#newcode717
lily/beam-quanting.cc:717: if ((concaveness >= 10000) || (damping >= 10000))
If I interpret this correctly, you're trying to avoid a later division when the
divisor is large.  That divisor is (concaveness + damping), so perhaps this
should  be (concaveness + damping) >= threshold.

https://codereview.appspot.com/337560043/diff/1/lily/beam-quanting.cc#newcode730
lily/beam-quanting.cc:730: slope = 0.6 * tanh (slope) / (damping + concaveness);
Pre-existing issue: division by zero if damping = -concaveness.  (Not saying you
have to fix it, but you might want to if you are making other changes.)

[Be-3, 2018-02-10 13:37:40]

https://codereview.appspot.com/337560043/diff/1/lily/beam-quanting.cc
File lily/beam-quanting.cc (right):

https://codereview.appspot.com/337560043/diff/1/lily/beam-quanting.cc#newcode717
lily/beam-quanting.cc:717: if ((concaveness >= 10000) || (damping >= 10000))
On 2018/02/10 13:26:34, Dan Eble wrote:
> If I interpret this correctly, you're trying to avoid a later division when
the
> divisor is large.  That divisor is (concaveness + damping), so perhaps this
> should  be (concaveness + damping) >= threshold.

Hi Dan,

Thanks for the hint.
But while concaveness and damping play the same role in the damping formula,
they are still two separate things and therefore, I have kept them separate on
purpose.

If both concaveness and damping stay below 10000, the beams should not be forced
to be horizontal, even if the *sum* of concaveness and damping exceeds 10000.

Torsten

[Be-3, 2018-02-10 14:38:39]

off-topic check for division by 0 added