Issue 3254: align unassociated lyrics using NoteColumn extent.

Issue 3254: align unassociated lyrics using NoteColumn extent.

This makes unassociated lyrics behave consistently with
associated lyrics.  Until now, "standalone" lyrics
were left-aligned (more precisely: their X-offset was 0);
changing self-alignment-X didn't have any effect on them.
Now it's possible to specify their alignment, like with
associated lyrics.

Also, this changes how issue 104 was resolved (see 0b14e8b2e122d) - 
alignment of syllables that are associated to a context without
noteheads is no longer aborted - and solves issue 247.

Later on, these changes should allow using aligned_on_parent for
other grobs (such as DynamicTexts).

Expected changes in output: all lyrics should be centered by default.

[janek, 2014-06-21 13:44:32]

Hi all!

Seems i'm back to LilyPond business (hopefully for long :)!
Please review - more patches (depending on this one) will follow.

best,
Janek

[lemzwerg, 2014-06-21 16:52:45]

LGTM.

https://codereview.appspot.com/108110044/diff/1/input/regression/unassociated...
File input/regression/unassociated-lyrics-alignment.ly (right):

https://codereview.appspot.com/108110044/diff/1/input/regression/unassociated...
input/regression/unassociated-lyrics-alignment.ly:5: texidoc = "Lyrics without
an associatedVoice should align properly.
Please format this text with @code{...} and the like.

[janek, 2014-06-21 21:35:53]

update documentation, fix texidoc formattic (acc. to Werner comment)

[janek, 2014-06-21 21:39:33]

Updated formatting, and updated documentation.

https://codereview.appspot.com/108110044/diff/1/input/regression/unassociated...
File input/regression/unassociated-lyrics-alignment.ly (right):

https://codereview.appspot.com/108110044/diff/1/input/regression/unassociated...
input/regression/unassociated-lyrics-alignment.ly:5: texidoc = "Lyrics without
an associatedVoice should align properly.
On 2014/06/21 16:52:44, lemzwerg wrote:
> Please format this text with @code{...} and the like.

Done.

[lemzwerg, 2014-06-22 07:17:38]

LGTM, thanks!

[janek, 2014-06-22 07:26:51]

You're welcome!  This is very encouraging :)

2014-06-22 9:17 GMT+02:00  <lemzwerg@googlemail.com>:
> LGTM, thanks!
>
> https://codereview.appspot.com/108110044/

[janek, 2014-06-22 08:28:17]

also fix issue 247

[janek, 2014-06-22 08:54:32]

remove comment from documentation that is no longer valid

[lemzwerg, 2014-06-22 09:47:33]

https://codereview.appspot.com/108110044/diff/60001/lily/lyric-engraver.cc
File lily/lyric-engraver.cc (right):

https://codereview.appspot.com/108110044/diff/60001/lily/lyric-engraver.cc#ne...
lily/lyric-engraver.cc:187: " Syllable will be attached to a PaperColumn
instead."));
I suggest that you change this last sentence to

  Syllables will be attached to PaperColumn grobs instead.

so that everything is in plural form.

[janek, 2014-06-22 10:27:31]

On 2014/06/22 09:47:33, lemzwerg wrote:
> https://codereview.appspot.com/108110044/diff/60001/lily/lyric-engraver.cc
> File lily/lyric-engraver.cc (right):
> 
>
https://codereview.appspot.com/108110044/diff/60001/lily/lyric-engraver.cc#ne...
> lily/lyric-engraver.cc:187: " Syllable will be attached to a PaperColumn
> instead."));
> I suggest that you change this last sentence to
> 
>   Syllables will be attached to PaperColumn grobs instead.
> 
> so that everything is in plural form.

My intention was as follows: the warning is emitted for every such syllable, so
it should speak about each syllable in singular.  However, the cause of the
warning is that an associatedVoice was set for the whole lyrics context, hence
the plural in first sentence.  I'd rather change first sentence to singular as
well, but i couldn't find a way to do this without making the warning confusing.

Maybe there is a convenient way of displaying such warinings only once? In that
case plural form would fit.

[dak, 2014-06-22 11:15:35]

https://codereview.appspot.com/108110044/diff/60001/Documentation/notation/vo...
File Documentation/notation/vocal.itely (right):

https://codereview.appspot.com/108110044/diff/60001/Documentation/notation/vo...
Documentation/notation/vocal.itely:520: play1 a4 game4
Why change the lyrics here?  "play a game" is not idiomatic.

[janek, 2014-06-22 12:45:47]

https://codereview.appspot.com/108110044/diff/60001/Documentation/notation/vo...
File Documentation/notation/vocal.itely (right):

https://codereview.appspot.com/108110044/diff/60001/Documentation/notation/vo...
Documentation/notation/vocal.itely:520: play1 a4 game4
On 2014/06/22 11:15:35, dak wrote:
> Why change the lyrics here?  "play a game" is not idiomatic.

I wanted to make lyrics shorter so that the alignment
not-necessarily-to-the-notes-but-to-other-moments will be obvious.  With longer
syllables, lyrics affects note spacing and it's not clear what is aligned to
what.

What about a completely artifictial example:

<<
  \new Voice = "melody" {
    \time 3/4
    c2 e4 g2 f
  }
  \new Lyrics \lyricmode {
    a1 b4 c4
  }
>>
?

[dak, 2014-06-22 12:57:44]

On 2014/06/22 12:45:47, janek wrote:
>
https://codereview.appspot.com/108110044/diff/60001/Documentation/notation/vo...
> File Documentation/notation/vocal.itely (right):
> 
>
https://codereview.appspot.com/108110044/diff/60001/Documentation/notation/vo...
> Documentation/notation/vocal.itely:520: play1 a4 game4
> On 2014/06/22 11:15:35, dak wrote:
> > Why change the lyrics here?  "play a game" is not idiomatic.
> 
> I wanted to make lyrics shorter so that the alignment
> not-necessarily-to-the-notes-but-to-other-moments will be obvious.  With
longer
> syllables, lyrics affects note spacing and it's not clear what is aligned to
> what.
> 
> What about a completely artifictial example:
> 
> <<
>   \new Voice = "melody" {
>     \time 3/4
>     c2 e4 g2 f
>   }
>   \new Lyrics \lyricmode {
>     a1 b4 c4
>   }
> >>
> ?

Start of the coffee canon? <URL:https://www.youtube.com/watch?v=8PN23MIqoLY> 
Though somewhat boring rhythmically.

[janek, 2014-06-22 19:05:54]

change example to caffee

[Keith, 2014-06-24 05:41:53]

Looks very good.
I looked for unwanted side-effects, and found no problems.
The old code did a very bad thing, ignoring the "self-alignment-X/Y" request if
the parent was a paper-column.

https://codereview.appspot.com/108110044/diff/80001/lily/paper-column.cc
File lily/paper-column.cc (right):

https://codereview.appspot.com/108110044/diff/80001/lily/paper-column.cc#newc...
lily/paper-column.cc:226: Paper_column::get_generic_interface_extent (Grob
*column, SCM interface, Axis a)
I don't understand what 'generic' means over just get_interface_extent(c, i, a)
but I leave it to you to name your function.

https://codereview.appspot.com/108110044/diff/80001/lily/self-alignment-inter...
File lily/self-alignment-interface.cc (right):

https://codereview.appspot.com/108110044/diff/80001/lily/self-alignment-inter...
lily/self-alignment-interface.cc:82: MAKE_SCHEME_CALLBACK
(Self_alignment_interface, centered_on_note_columns, 1);
Looks like somebody had a similar idea.  This function seems to be used only for
dynamics in a Dynamics context.

[janek, 2014-06-24 06:56:43]

Thanks for review!

https://codereview.appspot.com/108110044/diff/80001/lily/paper-column.cc
File lily/paper-column.cc (right):

https://codereview.appspot.com/108110044/diff/80001/lily/paper-column.cc#newc...
lily/paper-column.cc:226: Paper_column::get_generic_interface_extent (Grob
*column, SCM interface, Axis a)
On 2014/06/24 05:41:52, Keith wrote:
> I don't understand what 'generic' means over just get_interface_extent(c, i,
a)
> but I leave it to you to name your function.

Hmm.  Maybe i'll indeed change it to get_interface_extent.

https://codereview.appspot.com/108110044/diff/80001/lily/self-alignment-inter...
File lily/self-alignment-interface.cc (right):

https://codereview.appspot.com/108110044/diff/80001/lily/self-alignment-inter...
lily/self-alignment-interface.cc:82: MAKE_SCHEME_CALLBACK
(Self_alignment_interface, centered_on_note_columns, 1);
On 2014/06/24 05:41:53, Keith wrote:
> Looks like somebody had a similar idea.  This function seems to be used only
for
> dynamics in a Dynamics context.

Good catch!  Indeed, in the next patch I'm cleaning up dynamics alignment so
that it uses the newly improved aligned_on_parent everywhere instead of having
multiple alignment callbacks (some of them hardcoded in C++).  Stay tuned! :)

[janek, 2014-06-26 08:01:11]

pushed as c73b41b3e7be6d7280c6336cf03610cd7aed3000 - thanks for review!  I'll
post a follow-up patch later today.