Re: accidental problem

[Rune Zedeler]

(taking discussion back on list)

Citat Laurent Martelli <address@hidden>:

>   >> I'm using lily 1.6.6 (I also tried 1.6.7 and it has the same
>   >> behaviour), and it looks like in key where there's a flat (aes
>   >> for instance), if you have music with that note and a sharp, it
>   >> prints a natural in front of the sharp. This small test case
>   >> shows it:
> 
>   Rune> Yes this is true.  But is it a bug?  
> 
> I've just noticed this on some score I typeset some times ago, so I
> thought the behaviour changed. And it seems so weird to put a natural
> just in front of a sharp! I mean, it's redundant, isn't it ?

Well, think of this example:

mel = \notes { \key c \major fisis2 gis }
\score {
  \notes { \mel \transpose bes \mel }
}

Even though the two measures are the same just in different keys, then, with 
your approach, the first one would get two different accidentals (a doublesharp 
and a sharp) whereas the second one would get two identical accidentals (two 
sharps). I consider this very counterintuitive.
Currently the second one also gets two different accidentals (a naturalsharp 
and a sharp).

>   Rune> \property Score.extraNatural = ##f
> 
> Thanks for the tip.
> 
> I suggest that it be the default behaviour. I have no reference to
> back my preference, but maybe we could just ask lily users.

The default behaviour should not be like this - because lilypond defaults to 
classical notation, and in classical notation you (for sure) get a natural 
before a sharp if the sharp follows a doublesharp. I.e. in cisis cis the sharp 
in front of the cis for sure gets a natural. The question is whether to put a 
natural in front of the cis in ces cis.

-Rune