Re: Sketch for broken beams with consistent slopes (issue 4961041)

[mtsolo]

Hey all,

Sorry for hijacking this issue to turn it into another one.  I'll take
care of the variable name stuff in a second patch.

My summer of lily comes to a close with this monstrosity.  Some
gotchyas:

1)  Do not compile the regtests against this: they will break, as
several parts of Lily and several regtests use callbacks that I've
deleted (I don't have time to fix that before I leave).  Instead, copy
and paste any regtest with the word beam in it into a directory and
compile those to get a sense of what this patch does.

2)  This patch currently breaks cross-staff beam nice slope finding
(it'll probably take me an hour-ish to figure out why and fix it).
Other than that, I believe that it leaves other slopes in tact.

3)  You'll see that, by approaching beam slope this way, a lot of code
duplication that arose from chaining functions together is avoided.

4)  It'd be nice to have more callbacks to fine tune how
Beam_scoring_problem does its thing as it is now gimungous.

5) Most importantly, apply this patch and compile:

\version "2.15.10"

\relative c' {
\override Beam #'breakable = ##t
a8 [ b c d e f g \bar "" \break f e d c b a ]
}

\relative c' {
\override Beam #'breakable = ##t
\override Beam #'consistent-broken-slope = ##t
a8 [ b c d e f g \bar "" \break f e d c b a ]
}

\relative c' {
\override Beam #'breakable = ##t
a8 [ b c d e f \bar "" \break a c e g b ]
}

\relative c' {
\override Beam #'breakable = ##t
\override Beam #'consistent-broken-slope = ##t
a8 [ b c d e f \bar "" \break a c e g b ]
}

And there ya have it!

Cheers,
MS

P.S. Sorry for the generally kludgy nature of some aspects of this
patch, which may or may not include wholesale deletion of certain things
that are actually important.  This will be addressed in later patchsets
- this was thrown together in a rush :)

http://codereview.appspot.com/4961041/