Re: converting music stream back to ly code

[David Kastrup]

Janek Warchoł <address@hidden> writes:

> David, Jan-Peter,
>
> 2013/11/15 David Kastrup <address@hidden>:
>> Janek Warchoł <address@hidden> writes:
>>
>>> a quick question: given some lilypond code
>>>
>>> { c2 d e f }
>>
>> in what form?  Text?  Music expression?
>
> I meant the textual form.
>
>>> is it/would it be possible to get its music stream representation,
>>> manipulate that music stream and convert it back to lilypond code?
>>
>> What do you call "its music stream representation"?  LilyPond does not
>> have something like a music stream.  It has music expressions, and it
>> can iterate them resulting in stream events in a certain order (and
>> there is some event recorder for the part combiner and music quotes that
>> can save those events).
>
> ok, i got confused by Erik Sandberg's thesis as i don't know Lily
> internals in this area.

I think it plausible that during Erik Sandberg's thesis "stream events"
as opposed to music expressions were invented.  But there is no concrete
stream: you can listen to selected subsets of events at different
context levels, and every context has a different view of events passing
it.  It's more like a waterway network than a single stream, really.

> Sorry about that.
>
> Thanks for the information; it seems that it's not so easy to do what
> i was thinking of.

No idea.  But at least the buzzphrases don't map seamlessly to
LilyPond's operation.  It's usually pretty straightforward to solve any
_particular_ task with the engraver system.  It's just that generic
tasks are not as easily mapped to something workable.

The music quote and part combiner stuff are somewhat more generic, and
they tend to break down quite badly as soon as you use more than a
single Voice.

-- 
David Kastrup