2016-10-01 11:35 GMT+02:00 Thomas Morley <[email protected]>:
> With the following code LilyPond creates 4 midi-files, but two of them
> are named identical, overriding each other. So finally only three
> midi-files are present.
>
> \version "2.19.48"
>
> \book {
>   \score { { c'4 }  \midi {} }
>   \score { { cis'4 } \midi {} }
> }
>
> \book {
>   \score { { d'4 } \midi {} }
>   \score { { dis'4 } \midi {} }
> }
>
> Output from terminal:
>
> GNU LilyPond 2.19.49
> Processing `missing-midi.ly'
> Parsing...
> Interpreting music...
> Interpreting music...
> MIDI output to `missing-midi.midi'...
> MIDI output to `missing-midi-1.midi'...
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> Interpreting music...
> Interpreting music...
> MIDI output to `missing-midi-1.midi'...
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> MIDI output to `missing-midi-1-1.midi'...
> Success: compilation successfully completed
>
> The problem is located in `write-performances-midis' from midi.scm.
> The naming-procedure is insufficient for the above use-case.

It looks like the `rest'-variable in `write-performances-midis'
increases with each bookpart in a book.
But not with a second book in same file.
So bookparts are registered and taken into account, but not entire books.

_If_ this analysis is correct, some deeper fix (in paper-book.cc ?)
has to be done, fixing `write-performances-midis' will not be enough,
imho.
And we need to decide first how the midi-numbering _should_ be done.

Cheers,
  Harm

_______________________________________________
bug-lilypond mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/bug-lilypond

Reply via email to