Re: Microtonal key signatures fix

[Keith OHara]

Benkő Pál writes:
> Graham Breed writes:
> > If the accidentals aren't halves, the MIDI production fails.
> Please explain in more detail, I never had such problems with MIDI.

The following (derived from an example in the manual) crashes:

\include "makam.ly"
mode =  #`((6 . ,(- KOMA)) (3 . ,BAKIYE))
\score {
  \relative c' {
    \set Staff.keyAlterationOrder = #mode
    \key c \mode
    c4 cc db fk
    gbm4 gfc gfb efk
    fk4 db cc c
  }
  \layout{}
  \midi{}
}

If we adapt this score to Felipé's integer representation, his patch avoids the crash but 
outputs "6-sharps" as the key signature in MIDI.  So regardless of the internal 
representation of pitch, the computation producing the midi key signature needs Graham's 
improvement.

The same problem comes up in European music with tunings like 19-ET, popular in 
the 1600s, where C-sharp is 2/5 of a whole tone above C.

> Graham Breed writes:
> I've improved my patch.  This line gives correct MIDI key signatures
> for all plausibly authentic scenarios:
>
>   (apply + (map (lambda (p) (round (* (cdr p) 2))) pitch-list)) )
[...]
> It would be nice to get a version number for
> it being suppressed that I can submit a snippet against.

GrahamB,
  There is already a report for this bug, at 
http://code.google.com/p/lilypond/issues/detail?id=748

  Actually, issue 748 reports two problems, so I combined your patch with a fix 
for the other problem
soon link a patch to fix both.  It will be linked to issue 748 as soon as the 
self-test script finishes.
--
Keith