Unicode confusion

[Kurt Starsinic]

Hi,

I've been using lilypond quite happily for about a year, and recently
I've been delving into multilingual issues.  I'm trying to set some
German language text, and I'm having a hard time entering it.

I've looked at the example typography-demo.ly file, which produces
beautiful output.  The beginning lyrics there display as:

S<LATIN SMALL LETTER U WITH DIAERESIS> - <LATIN SMALL LETTER SHARP S>es

The actual bytes in the sample score that correspond to the LATIN
SMALL LETTER U WITH DIARESIS (U+00FC) are:  C3 83 C2 BC.

The *proper* UTF-8 encoding of U+00FC requires two bytes:  C3 BC.

The two-byte encoding I suggest looks correct in every Unicode viewer
I've tried (e.g., gvim, yudit, Firefox).  The four-byte encoding looks
like line noise.  The four-byte encoding is actually the UTF-8 for
<LATIN CAPITAL LETTER A WITH TILDE><VULGAR FRACTION ONE QUARTER>. 
Correct UTF-8 encodings of Unicode characters do *not* display
properly in my lilypond scores.

As near as I can figure, lilypond is expecting UTF-8 to be
twice-encoded--U+00FC is encoded to C3 BC, which is then *re*-encoded
to C3 83 C2 BC.

Am I missing something here?  Is anybody creating lilypond scores such
that they can read the lyrics *both* in .ly files and in the .pdf
output?  If not, how in the world do you manage to input your lyrics?

Thanks in advance for any advice,
Kurt