Re: Chapter "Collision resolution"

[Trevor Daniels]

David Kastrup wrote Thursday, January 13, 2011 12:47 PM


> I've decided that this is not a restriction, but a bug.  Here is 
> the
> reason: when merge-differently-headed is set and three or more 
> notes are
> in the note column, heads _are_ getting merged but get assigned 
> the
> wrong notehead.

No, they're not.  See what happens when \shiftOnn is
commented out in the example I send previously.  I've
removed the first bar so you can see it more clearly:

\relative c'' {
<<
 {
   \mergeDifferentlyHeadedOn
%    \shiftOnn
   g'2 <fis d>
 } \\ {
   e,2 r
 } \\ {
   e8 a b c s2
 }
>>
}

> So this bug needs to be resolved either
>
> a) completely by letting Lilypond pick the right notehead in this 
> case
> b) by turning it into a restriction, namely not letting Lilypond 
> attempt
>   merging heads at all when there are too many of them

(b) is exactly what Lily does at present.  See the example
above.  As I explained earlier, the 'merging' you thought
you saw was simply the effect of bad source code causing
a collision of two noteheads.

> I am taking a look at the code to figure out what happens and what 
> one
> might want to do about this.

:)

Trevor