Updates:
Status: Invalid
Comment #5 on issue 2985 by [email protected]: lilypond: Umlauts and other
special characters incorrectly exported to PDF meta data
http://code.google.com/p/lilypond/issues/detail?id=2985
Actually, it isn't. Metadata is encoded either in PDFDocEncoding
(basically Latin-1), or in UTF16BE with BOM. LilyPond does this using
ly:encode-string-for-pdf, the result is correct, and Evince is wrong.
Neither xpdf nor pdfinfo have a problem with the encoded material which is
_not_ UTF-8. That may have been a stupid design decision for PDF formats
by Adobe, but Evince is not free to invent its own PDF standard.
So the original report was wrong because the file was encoded badly, and
the modified report was wrong because the PDF is right but Evince does not
interpret it correctly. This should be turned into a bug report for Evince.
