Update the non-ascii definition.

Author: tabatkins

README.md

@@ -2,7 +2,7 @@ Standards-Based CSS Parser
 ==========================
 
 This project implements a standards-based CSS Parser.
-I'm writing the CSS Syntax spec <http://dev.w3.org/csswg/css-syntax/>,
+I'm the editor of the CSS Syntax spec <http://drafts.csswg.org/css-syntax/>,
 and need an implementation of it for testing purposes.
 
 This parser is *not* designed to be fast,
@@ -18,7 +18,7 @@ to make the same change in the parser.
 It is intended to fully and completely match browser behavior
 (at least, as much as the final spec does).
 
-There's a [dingus](https://rawgit.com/tabatkins/parse-css/master/example.html) for testing it out,
+There's a [dingus](https://tabatkins.github.io/parse-css/example.html) for testing it out,
 or just quickly checking what some CSS parses into.
 
 Using the Library
@@ -43,13 +43,7 @@ and thus this parser,
 is *extremely generic*.
 It doesn't have any specific knowledge of CSS rules,
 just the core syntax,
-so it won't throw out invalid or unknown things,
-and it can't even actually parse the contents of blocks
-(because it doesn't know if they'll contain rules or declarations,
-and those are ambiguous without any context).
-I plan to add some functions that add more CSS knowledge
-(in an extensible way, so it'll retain anything custom that you want to handle yourself),
-but for now you have to do all the verification and additional parsing yourself.
+so it won't throw out invalid or unknown things.
 
 Parsing Functions
 -----------------

parse-css.js

@@ -17,7 +17,25 @@ function hexdigit(code) { return digit(code) || between(code, 0x41,0x46) || betw
 function uppercaseletter(code) { return between(code, 0x41,0x5a); }
 function lowercaseletter(code) { return between(code, 0x61,0x7a); }
 function letter(code) { return uppercaseletter(code) || lowercaseletter(code); }
-function nonascii(code) { return code >= 0x80; }
+function nonascii(code) {
+	return (
+		code == 0xb7 ||
+		between(code, 0xc0, 0xd6) ||
+		between(code, 0xd8, 0xf6) ||
+		between(code, 0xf8, 0x37d) ||
+		between(code, 0x37f, 0x1fff) ||
+		code == 0x200c ||
+		code == 0x200d ||
+		code == 0x203f ||
+		code == 0x2040 ||
+		between(code, 0x2070, 0x218f) ||
+		between(code, 0x2c00, 0x2fef) ||
+		between(code, 0x3001, 0xd7ff) ||
+		between(code, 0xf900, 0xfdcf) ||
+		between(code, 0xfdf0, 0xfffd) ||
+		code >= 0x10000
+	);
+}
 function namestartchar(code) { return letter(code) || nonascii(code) || code == 0x5f; }
 function namechar(code) { return namestartchar(code) || digit(code) || code == 0x2d; }
 function nonprintable(code) { return between(code, 0,8) || code == 0xb || between(code, 0xe,0x1f) || code == 0x7f; }
@@ -1373,6 +1391,5 @@ exports.parseADeclaration = parseADeclaration;
 exports.parseAComponentValue = parseAComponentValue;
 exports.parseAListOfComponentValues = parseAListOfComponentValues;
 exports.parseACommaSeparatedListOfComponentValues = parseACommaSeparatedListOfComponentValues;
-exports.tokenize = tokenize;
 
 }));