is_leading_utf8_byte returns wrong answer for 0xf8..=0xff (utf-8 invalid bytes are not leading bytes)

[thomcc]

is_leading_utf8_byte seems to assume it's input byte came from a stream of utf-8 encoded text

bstr/src/utf8.rs

Lines 797 to 802 in 91edb3f

	fn is_leading_utf8_byte(b: u8) -> bool {
	// In the ASCII case, the most significant bit is never set. The leading
	// byte of a 2/3/4-byte sequence always has the top two most significant
	// bigs set.
	(b & 0b1100_0000) != 0b1000_0000
	}

Specifically, it returns true for bytes in the range 0xf8..=0xff, which is wrong. These bytes are not leading bytes, nor are they trailing bytes. They're bytes which can only appear in sequences of invalid utf-8 -- there's no case where they appear in a stream of valid utf8-encoded text, but obviously bstr doesn't want to rely on this.

This is an internal function so it may not matter, but I have a hard time understanding much of the code in this module (I am not someone who looks at a DFA state table and sees meaning in the numbers), so I am unsure whether or not it can cause problems... It's also totally possible this is handled in some other way, I was just finally getting around to doing #44 and noticed this.

(I did try some trivial tests to try and get weird things to happen without avail, but figured it was better to report it -- when I look at those tables my eyes glaze over, but hopefully the code's author understands it!)