Re: Make sure slurs actually avoid stafflines. (issue 15400049)

[janek . lilypond]

New patch set uploaded.
I expect that it may not be very good - i'm tired and cannot think
clearly.

I have a rough idea of how this problem should be fixed ideally, but
implementing such solution would require much more work.



https://codereview.appspot.com/15400049/diff/1/lily/slur-configuration.cc
File lily/slur-configuration.cc (left):

https://codereview.appspot.com/15400049/diff/1/lily/slur-configuration.cc#oldcode62
lily/slur-configuration.cc:62:
On 2013/10/24 03:33:43, Keith wrote:
> On 2013/10/23 04:22:23, Keith wrote:
> > > but the function was broken: it only moved slurs in one direction

> We concluded the same thing about the old code.  The old code intended
to move
> slurs in one direction only.

> I was only confused by your comment. I interpreted 'broken' to mean
"failing to
> to what was intended".

Ah, ok.  For me the intent was to avoid the collision, and in this
regard old code failed.

https://codereview.appspot.com/15400049/diff/1/lily/slur-configuration.cc
File lily/slur-configuration.cc (right):

https://codereview.appspot.com/15400049/diff/1/lily/slur-configuration.cc#newcode69
lily/slur-configuration.cc:69: // Get the point(s) where the curve is
horizontal (i.e. the belly):
On 2013/10/23 04:22:23, Keith wrote:
> For slurs on a slope, the spot where the slur is horizontal might not
look like
> its belly.  I would just call it "the horizontal part"

Done.

https://codereview.appspot.com/15400049/diff/1/lily/slur-configuration.cc#newcode84
lily/slur-configuration.cc:84: Real const slur_th = state.thickness_ *
staff_th * 10;
On 2013/10/24 03:33:43, Keith wrote:
> On 2013/10/23 21:41:24, janek wrote:
> > On 2013/10/23 04:22:23, Keith wrote:
> > > Why 10?  Is that the thickness of the fattest part of the slur?
> >
> > No, it's because staff_th is measured in staffspaces, while
state.thickness_
> > appears to be measured in 1/10ths of staffline-thicknesses

> state.thickness_ already has the \layout{line-thickness} as a factor
in its
> initialization in Slur_score_state::fill().

> Staff_symbol_referencer::line_thickness() returns something that
scales with
> \layout{line-thickness}

> So, it looks like you have two factors of \layout{line-thickness} and
the *10
> roughly cancels one of them.

Indeed.
So, should i change the definition of thickness_ in
Slur_score_state::fill so that it includes staff thickness?  Because
right now it seems that info is inaccurate.

https://codereview.appspot.com/15400049/