Re: Issue 1741 in lilypond: Tablature Bug when pitch is higher then beh in a quarter-tone string

[lilypond]

Updates:
        Labels: Patch-new

Comment #10 on issue 1741 by address@hidden: Tablature Bug when pitch is  
higher then beh in a quarter-tone string
http://code.google.com/p/lilypond/issues/detail?id=1741

The tabulature code uses ly:pitch-semitones, and ly:pitch-semitones  
maddeningly does not return a rational number, but rather an integer.  The  
routine responsible for this is called Pitch::rounded_semitone_pitch in  
pitch.cc.  In striking contrast to its name, it does not round, but rather  
it truncates.  To add insult to injury, it truncates towards zero, namely  
middle-C.  Since it makes no sense to change rounding direction when  
passing middle-C, and since rounding might be what is expected usually, I  
propose the following patch.

Personally, I would not round at all but rather return a rational when the  
result is not an integer.  However, I don't know what assumptions other  
code might make.  While I don't consider it likely that code getting an  
integer for an in-between note will do something sensible, I have not tried  
out what happens when staying with a rational in case of doubt here.

Attachments:
        pitch.diff  674 bytes