issue 5319: Make skylines reflect grob rotation

issue 5319: Make skylines reflect grob rotation

modified:   ../lily/include/stencil.hh
Add parameter to skylines_from_stencil for rot (in analogy to pad)

modified:   ../lily/stencil-integral.cc

Stencil::skylines_from_stencil
If new rotation parameter rot is specified, use a rotated copy of
the original stencil to build the skyline transformation matrix.
As the rotation is rarely used, the general (non-rotated) case will
still use the original stencil as before.

Grob::vertical_sklyines_from_grob and Grob::horizontal_skylines_from_grob
Read "rotation" property an pass it (SCM) to skylines_from_stencil.

modified:   ../lily/accidental.cc

Accidental_interface::horizontal_skylines
Read "rotation" property and pass it (SCM) to skylines_from_stencil.

Announced in ../Documentation/changes.tely

Documentation example slightly adapted (20° to 15°) to new skyline
behaviour (padding will now be respected). In all affected languages:
modified:   ../Documentation/de/notation/changing-defaults.itely
modified:   ../Documentation/es/notation/changing-defaults.itely
modified:   ../Documentation/fr/notation/changing-defaults.itely
modified:   ../Documentation/it/notation/changing-defaults.itely
modified:   ../Documentation/ja/notation/changing-defaults.itely
modified:   ../Documentation/notation/changing-defaults.itely

modified:   ../input/regression/skyline-color-rotation.ly
(20° to 15°, cf. documentation)

new:   ../input/regression/skyline-grob-rotation.ly
New regression test.

[Be-3, 2018-05-03 12:51:11]

Please review...

Inadvertently, file
../lily/key-signature-interface.cc
slipped into the patch (from issue 5312 currently in staging).

So ignore is (or shall I upload a new, cleansed patch for reviewing?)

Thanks,
Torsten

[Be-3, 2018-05-03 17:41:33]

New upload due to inconsistency

[lemzwerg, 2018-05-03 17:42:22]

LGTM

https://codereview.appspot.com/342060043/diff/1/lily/key-signature-interface.cc
File lily/key-signature-interface.cc (right):

https://codereview.appspot.com/342060043/diff/1/lily/key-signature-interface....
lily/key-signature-interface.cc:114: padding += (intersection (ht_right,
last_ht_left).length ()
The outermost parentheses on the right side are redundant.

[Be-3, 2018-05-03 17:58:00]

On 2018/05/03 17:42:22, lemzwerg wrote:
> LGTM
> 
>
https://codereview.appspot.com/342060043/diff/1/lily/key-signature-interface.cc
> File lily/key-signature-interface.cc (right):
> 
>
https://codereview.appspot.com/342060043/diff/1/lily/key-signature-interface....
> lily/key-signature-interface.cc:114: padding += (intersection (ht_right,
> last_ht_left).length ()
> The outermost parentheses on the right side are redundant.

Hi Werner,

the key-signature-interface.cc slipped in only accidentally and does not belong
to this issue. It has disappeared from 2nd patch set. There was also a mismatch
in changes.tely and James aked me to make a new patch against current master.
Apart from that, yes, the parentheses are redundant, but they were also there in
the original coding and so I kept them...

All the best,
Torsten

[dak, 2018-06-09 20:43:22]

https://codereview.appspot.com/342060043/diff/20001/lily/stencil-integral.cc
File lily/stencil-integral.cc (right):

https://codereview.appspot.com/342060043/diff/20001/lily/stencil-integral.cc#...
lily/stencil-integral.cc:1132: Stencil *q = unsmob<Stencil> (s->smobbed_copy
());
Once you create a smobbed copy, it must be protected from garbage collection. 
You don't do this here at all.  What you need to do is to retain the SCM value
on the stack, not its unsmobbification.  If you don't want to keep unsmobbing
it, you can keep both and do an scm_remember_upto_here (scm_q);
at the point up to which it must not be collected.  However, it seems that uses
here are few enough that you can just do
SCM q = s->smobbed_copy ();
and use unsmob<Stencil> (q) where you now use q.