Relax markup grammar to allow non-ASCII chars in tag names and attrib…

…utes. Fix #585
PrismJS · Jun 12, 2015 · 52fd55e · 52fd55e
1 parent 534f34c
commit 52fd55e
Show file tree

Hide file tree

Showing 4 changed files with 17 additions and 12 deletions.
diff --git a/components/prism-markup.js b/components/prism-markup.js
@@ -4,13 +4,13 @@ Prism.languages.markup = {
 	'doctype': /<!DOCTYPE.+?>/,
 	'cdata': /<!\[CDATA\[[\w\W]*?]]>/i,
 	'tag': {
-		pattern: /<\/?[\w:-]+\s*(?:\s+[\w:-]+(?:=(?:("|')(\\?[\w\W])*?\1|[^\s'">=]+))?\s*)*\/?>/i,
+		pattern: /<\/?[^\s>\/]+\s*(?:\s+[^\s>\/=]+(?:=(?:("|')(\\?[\w\W])*?\1|[^\s'">=]+))?\s*)*\/?>/i,
 		inside: {
 			'tag': {
-				pattern: /^<\/?[\w:-]+/i,
+				pattern: /^<\/?[^\s>\/]+/i,
 				inside: {
 					'punctuation': /^<\/?/,
-					'namespace': /^[\w-]+?:/
+					'namespace': /^[^\s>\/:]+:/
 				}
 			},
 			'attr-value': {
@@ -21,9 +21,9 @@ Prism.languages.markup = {
 			},
 			'punctuation': /\/?>/,
 			'attr-name': {
-				pattern: /[\w:-]+/,
+				pattern: /[^\s>\/]+/,
 				inside: {
-					'namespace': /^[\w-]+?:/
+					'namespace': /^[^\s>\/:]+:/
 				}
 			}
 

diff --git a/components/prism-markup.min.js b/components/prism-markup.min.js
diff --git a/examples/prism-markup.html b/examples/prism-markup.html
@@ -72,4 +72,9 @@ <h2>Invalid HTML</h2>
 <h2>Multi-line attribute values</h2>
 <pre><code>&lt;p title="foo
 bar
-baz"></code></pre>
+baz"></code></pre>
+
+<h2>XML tags with non-ASCII characters</h2>
+<pre><code>&lt;L&auml;ufer&gt;foo&lt;/L&auml;ufer&gt;
+&lt;tag l&auml;ufer="l&auml;ufer"&gt;bar&lt;/tag&gt;
+&lt;l&auml;ufer:tag&gt;baz&lt;/l&auml;ufer:tag&gt;</code></pre>
diff --git a/prism.js b/prism.js
@@ -441,13 +441,13 @@ Prism.languages.markup = {
 	'doctype': /<!DOCTYPE.+?>/,
 	'cdata': /<!\[CDATA\[[\w\W]*?]]>/i,
 	'tag': {
-		pattern: /<\/?[\w:-]+\s*(?:\s+[\w:-]+(?:=(?:("|')(\\?[\w\W])*?\1|[^\s'">=]+))?\s*)*\/?>/i,
+		pattern: /<\/?[^\s>\/]+\s*(?:\s+[^\s>\/=]+(?:=(?:("|')(\\?[\w\W])*?\1|[^\s'">=]+))?\s*)*\/?>/i,
 		inside: {
 			'tag': {
-				pattern: /^<\/?[\w:-]+/i,
+				pattern: /^<\/?[^\s>\/]+/i,
 				inside: {
 					'punctuation': /^<\/?/,
-					'namespace': /^[\w-]+?:/
+					'namespace': /^[^\s>\/:]+:/
 				}
 			},
 			'attr-value': {
@@ -458,9 +458,9 @@ Prism.languages.markup = {
 			},
 			'punctuation': /\/?>/,
 			'attr-name': {
-				pattern: /[\w:-]+/,
+				pattern: /[^\s>\/]+/,
 				inside: {
-					'namespace': /^[\w-]+?:/
+					'namespace': /^[^\s>\/:]+:/
 				}
 			}