Re: TupletNumber regression

[address@hidden]

On 22 mai 2012, at 19:17, JD Duncan wrote:

>> I'm not top posting.
> 
> The following snippet shows a measure where the TupletNumber has good 
> placement 
> in 2.14.2 but strange placement in 2.15.38.
> 
> JD
> 
> 
> \language english
> 
> %%%%  Score begins here
> \parallelMusic  #'(pianoRH dynamics  pianoLH) {
> 
> % Measure 9
> <<  { \stemDown df''2\arpeggio s2 } \\
>       { \stemUp \times 2/3 { g''8[ af'' c'''] }  df'''8 f''' 
>       \ottava #+1 
>       \times 2/3 { <g'' g'''>8 \( <af'' af'''> <g'' g'''> \) }
>       \times 2/3 { <f'' f'''>  \( <df'' df'''> <f'' f'''> \) }
>       \ottava #0 } >> |
> s1\f | 
> \clef bass <df af f'>2\arpeggio \clef treble <af' f''>2   |
> 
> 
> % End of Music
> }
> 
> \score {
>       \new PianoStaff
>       <<
>         \set PianoStaff.connectArpeggios = ##t
> 
>         \new Staff = "RH" \with { \consists "Page_turn_engraver" } \pianoRH
>         \new Dynamics \dynamics
>         \new Staff = "LH"   \pianoLH
>       >>
> 
> \layout { }
> }
> 

Dear JD,

In LilyPond 2.15.38, TupletNumber placement automatically follows the voice, 
not the stem direction.  You've inverted the traditional stem direction of 
voices (usually the top voice points up), and the TupletNumber does not 
automatically follow this.  To change this, you can add an \override 
TupletBracket #'direction = #UP in your upward-pointing voice.

If you feel that this is not good default behavior, let us know and we'll write 
up a bug report about changing the default back to 2.14.2.  Either way, it 
should be noted in the change log.

Cheers,
MS