Re: tie problem

[Kieren MacMillan]

Hi, Rainer:

> I've following code:
> <c d g>1\arpeggio ~ | << <c d g>1\fermata \\ {r2 r4 b'32 (c d e f g  
> a b)} >>
>
> According to the documentation the two "<c d g>" should be tied.
> Unfortunately this is not the case. Could you help me please?

Well, you've run into a subtlety of Lilypond that could, perhaps, be  
better documented...  ;-)

Consider the following code snippet:

%%% BEGIN SNIPPET %%%

\version "2.9.17"

\paper
{
        indent = 0\in
        line-width = 3\in
}

theMusic = \relative c'
{
        <c d g>1 ~ | << <c d g>1\fermata \\ { s2 c4 c } >>  \break
        <c d g>1 ~ | << { <c d g>1\fermata } \new Voice { s2 c4 c } >>
}

\score
{
        \theMusic
}

%%% END SNIPPET %%%

Notice that the tie does not work "as expected" in the first example,  
but does in the second.

This is because, in the first example, the << \\ >> construct  
explicitly instantiates TWO voices, BOTH of which are in addition to  
the one which contains the <c d g> that starts the tie -- as a  
result, the tie doesn't know where to end, because its Voice doesn't  
continue on into the <<>> block.

In the second example, the \\ is replaced by an explicit (manual)  
instantiation using \new Voice -- this ensures that anything before  
the \new Voice command is considered part of the Voice that existed  
before the <<>> block began, and so the tie knows where to terminate.

Does that make sense?
Or, at the very least, does it explain why you're seeing what you're  
seeing?  =)

Best regards,
Kieren.