Re: Issue 1772 in lilypond: Completion_rest_engraver fails with fractional rests

[address@hidden]

On Jul 22, 2011, at 12:24 PM, address@hidden wrote:

> 
> Comment #1 on issue 1772 by address@hidden: Completion_rest_engraver fails 
> with fractional rests
> http://code.google.com/p/lilypond/issues/detail?id=1772
> 
> As Neil Puttock reported Completion_heads_engraver behaves the same:
> http://lists.gnu.org/archive/html/bug-lilypond/2011-07/msg00606.html
> 
> I adapted his example a little bit:
> % ---------------8<------------------
> \version "2.15.5"
> <<
> % bad
> \new Voice
> \with { \remove "Note_heads_engraver" \consists "Completion_heads_engraver"  }
> { c''1*6/4 c''2 }
> 
> % good
> \new Voice
> \with { \remove "Note_heads_engraver" \consists "Completion_heads_engraver"  }
> { c''1. c''2 }
> 
> % ---------------8<------------------
> 
> Here one can see that both voices are correctly aligned horizontally, 
> although the durations of the "bad" example don't sum up correctly (-> the 
> value isn't only split up in wrong parts but completely wrong)
> While this may be totally clear to everybody I thought it should be 
> explicitely stated here in the tracker.
> 
> HTH
> Urs
> 

Just a thought: it may be easier to fix this by scrapping the completion 
engravers entirely and having music functions that operate on music streams.  
For example:

a4. a1 a2 a8

in the completion heads engraver becomes

a4. a2 ~ a8 ~ a4. a2 a8

This would be a lot nicer as

a4. a8 ~ a2 ~ a4. a2 a8

and even nicer as

a4. a8 ~ a2 ~ a4. a8 a2 (in addition to completing heads, rearranging based on 
beat conventions à la autobeam engraver).

A music function can glance ahead in time, see what the appropriate way to 
split would be, and then execute it.

The autobeam engraver can be rewritten the same way.

Cheers,
MS