Potential discrepancy between tests and DerivedBidiClass.txt

[asibahi]

Hello

This section in the bmp tests seems to be in contradiction with the values in DerivedBidiClass.txt.

        // noncharacters
        assert_eq!(bidi_class('\u{FDD0}'), L);
        assert_eq!(bidi_class('\u{FDD1}'), L);
        assert_eq!(bidi_class('\u{FDEE}'), L);
        assert_eq!(bidi_class('\u{FDEF}'), L);
        assert_eq!(bidi_class('\u{FFFE}'), L);
        assert_eq!(bidi_class('\u{FFFF}'), L);

While in there:

FDD0..FDEF    ; BN # Cn  [32] <noncharacter-FDD0>..<noncharacter-FDEF>
# ...
FFFE..FFFF    ; BN # Cn   [2] <noncharacter-FFFE>..<noncharacter-FFFF>

I noticed the discrepancy when the tests in my zig port failed when I switched the data source to uucode. The discrepancy does not seem to have any impact at all as the conformance tests in BidiTest.txt and BidiCharacterTest.txt pass fine.

[Manishearth]

Not really sure where this comes from! DOn't have time to investigate but would probably be okay with a PR that changes that behavior.