Re: [frogs] New frog in an empty pond?

[Pavel Roskin <proski@xxxxxxx>]

Quoting "mike@xxxxxxxxxxxxxxxx" <mike@xxxxxxxxxxxxxxxx>:

> Hey Pavel,
>
> Thanks for your patience - I'm a bit short on time these days so I   
> can only give feedback in chunks, but I'll try to make it as   
> comprehensive as I can.
>
> The ideal UI for something like this is to be able to do:
>
> a8 [ <\change Staff = "foo" a \change Staff = "bar" a \change Staff   
> = "other" a > a8 ]

Yes, that the best solution, as it would make Lilypond know that the  
notes belong to one chord.

> There are five main issues that will present themselves:
>
> 1) Implementing the iterator that will handle the change events.
> 2) Finding out how to get information to the correct staff contexts.
> 3) Either expanding NoteColumn to be able to cross staves or create   
> NoteColumnColumn that works like SpanBar in collecting NoteColumns   
> that belong to the same chord.  The former is undoubtedly more   
> difficult but likely a better approach.
> 4) Expand the algorithms in note-collision.cc to work across staves.
> 5) Allow a beam to fall in the middle of a cross-staff chord.
>
> I'd work on them in this order.  1-4 seem necessary in order to   
> integrate this into LilyPond whereas 5 can be done in a subsequent   
> patch.
>
> Some hints for how to proceed:
>
> 1)  Copy logic from change-iterator.cc into event-chord-iterator.cc.
> 2)  Look at auto-beam-engraver,cc and span-bar-stub-engraver.cc to   
> see how information can be passed around different contexts.
> 3)  Look at span-bar.cc to see how grobs can group together objects   
> that fall on different staves and see how this can be applied to   
> note-column.cc.
>
> 4) and 5) will come after 1,2, and 3.  You'll already have your work  
>  cut out for you with 1, 2, & 3.  E-mail me with any questions for   
> what things mean in these files!

I think I'll try it soon.

However, I would prefer to spend some time on the current approach  
(extending the stems).  If I succeed, I'll be able to provide a  
stopgap solution for everybody that would work with Lilypond 2.16 and  
maybe even 2.14.  Even if I don't succeed, I'll expand my knowledge of  
Scheme and Lilypond internals.

As I understand, I need to make a grob that is typeset after stems.   
It could be a separate grob, such as an arpeggio.  Alternatively, I'll  
try to make one stem a parent of the other one.

Let me play with it a little bit.  Then I'll be more confident with  
the C++ code.

--
Regards,
Pavel Roskin

--
----
Join the Frogs!