This class takes a text stream and produces a sequence of HTML tokens.

It requires its source stream to support #peek.

	private-initialization
	initialize:

	private-tokenizing
	nextAttributeValue nextComment nextName nextSpaces nextTag nextTagOrComment nextText skipSpaces

	stream protocol
	atEnd

	tokenizing
	next

	instance creation
	on:

	private	top
	nextChar
	peekChar

	private-initialization	top
	initialize:

	private-tokenizing	top
	nextAttributeValue return the next sequence of alphanumeric characters; used to read in the value part of a tag's attribute, ie <tagname attribname=attribvalue>
	nextComment we've seen < and the next is a !. read until the whole comment is done
	nextName return the next sequence of alphanumeric characters
	nextSpaces read in as many consecutive space characters as possible
	nextTag we've seen a < and peek-ed something other than a !. Parse and return a tag
	nextTagOrComment next character is a $<. So read either a tag or a token
	nextText returns the next textual segment
	skipSpaces skip as many consecutive space characters as possible

	stream protocol	top
	atEnd are there any more tokens? This is equivalent to whether there is any more input

	tokenizing	top
	next return the next HtmlToken, or nil if there are no more

	initialization	top
	initialize HtmlTokenizer initialize

	instance creation	top
	on: