ZnCharacterReadStream (and friends) should publish position in terms of Character position

[dalehenrich]

The current implementation wraps a Byte stream and the position of the ZnCharaterReadStream is returned using the byte position not character position, so the following test fails returning 4 and not 3 as expected for character position:

	| string bytes stream res |
	string := 'eißendeße'.
	bytes := ZnUTF8Encoder new encodeString: string.
	stream := (ZnCharacterReadStream on: bytes readStreamPortable).
	res := stream next; next; next.
	self assert: res equals: $ß.
	self assert: stream position equals: 3.

This is from the test ZnCharacterStreamTests >> testUtf8EncodingStreamPosition.

[kurtkilpela]

I took a look at Pharo's behavior around ZnCharacterReadSteam>>#position. Using a similar test case, Pharo's implementation also returns the byte position rather than the character position. Is this a semantic change we want to make? Or do we want to maintain semantics w/ Pharo?

| string bytes stream res |
string := 'eißßßßßßßßßßßßendeße'.
bytes := ZnUTF8Encoder new encodeString: string.
stream := (ZnCharacterReadStream on: bytes readStream).
res := stream next; next; next; next; next.
stream position. "-> 8 rather than 5"

[dalehenrich]

I am in favor of having the code work correctly over Pharo compatibility and skipping 1/2 characters does not seem correct to me ...

[kurtkilpela]

The tests associated with this issue have been temporarily removed from the test set in ecf4900. They will be restored when this issue is addressed.