On Fri, 2008-02-29 at 15:55 +0100, Rune Zedeler wrote:
> The outlines used to detect collisions are way too rough in the case of 
> dynamics.
> 
> \version "2.11.0"
> \layout { ragged-last = ##t }
> {
>    \dynamicUp
>    d''2  e''8 \< c'' c''4\!
>    d''2( e''8)\< c'' c''4\!
>    d''2( e''8)\> c'' c''4\!
>    d''2( e''8)\< c'' c''4\!\f
> }
> 
> In this example we notice the following.
> The first measure looks fine.
> In the second measure the cresc is for too high. The only difference to 
> first measure is the added slur, so it must be the slur, that pushes the 
> dynamic too high up.

Currently, all of the outside-staff placement is done using bounding
boxes. In this situation, I agree that the output is bad but the changes
needed to use more accurate outlines are probably too far-reaching to be
made so late in the 2.11 cycle. Anyway, I have added this as #584.

Joe



_______________________________________________
bug-lilypond mailing list
bug-lilypond@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-lilypond

Reply via email to