fix(sax): Only apply HTML rules if mimeType is present

xmldom · Feb 16, 2022 · 9072016 · 9072016
1 parent 04c9f26
commit 9072016
Show file tree

Hide file tree

Showing 2 changed files with 40 additions and 6 deletions.
diff --git a/lib/sax.js b/lib/sax.js
@@ -1,4 +1,6 @@
-var NAMESPACE = require("./conventions").NAMESPACE;
+var conventions = require("./conventions");
+var NAMESPACE = conventions.NAMESPACE;
+var MIME_TYPE = conventions.MIME_TYPE;
 
 //[4]   	NameStartChar	   ::=   	":" | [A-Z] | "_" | [a-z] | [#xC0-#xD6] | [#xD8-#xF6] | [#xF8-#x2FF] | [#x370-#x37D] | [#x37F-#x1FFF] | [#x200C-#x200D] | [#x2070-#x218F] | [#x2C00-#x2FEF] | [#x3001-#xD7FF] | [#xF900-#xFDCF] | [#xFDF0-#xFFFD] | [#x10000-#xEFFFF]
 //[4a]   	NameChar	   ::=   	NameStartChar | "-" | "." | [0-9] | #xB7 | [#x0300-#x036F] | [#x203F-#x2040]
@@ -50,6 +52,7 @@ XMLReader.prototype = {
 	}
 }
 function parse(source,defaultNSMapCopy,entityMap,domBuilder,errorHandler){
+	var isHTML = MIME_TYPE.isHTML(domBuilder.mimeType);
 	function fixedFromCharCode(code) {
 		// String.prototype.fromCharCode does not supports
 		// > 2 bytes unicode chars directly
@@ -162,7 +165,15 @@ function parse(source,defaultNSMapCopy,entityMap,domBuilder,errorHandler){
 				var el = new ElementAttributes();
 				var currentNSMap = parseStack[parseStack.length-1].currentNSMap;
 				//elStartEnd
-				var end = parseElementStartPart(source,tagStart,el,currentNSMap,entityReplacer,errorHandler);
+				var end = parseElementStartPart(
+					source,
+					tagStart,
+					el,
+					currentNSMap,
+					entityReplacer,
+					errorHandler,
+					isHTML
+				)
 				var len = el.length;
 
 
@@ -191,7 +202,7 @@ function parse(source,defaultNSMapCopy,entityMap,domBuilder,errorHandler){
 					}
 				}
 
-				if (NAMESPACE.isHTML(el.uri) && !el.closed) {
+				if (isHTML && !el.closed) {
 					end = parseHtmlSpecialContent(source,end,el.tagName,entityReplacer,domBuilder)
 				} else {
 					end++;
@@ -222,7 +233,9 @@ function copyLocator(f,t){
  * @see #appendElement(source,elStartEnd,el,selfClosed,entityReplacer,domBuilder,parseStack);
  * @return end of the elementStartPart(end of elementEndPart for selfClosed el)
  */
-function parseElementStartPart(source,start,el,currentNSMap,entityReplacer,errorHandler){
+function parseElementStartPart(
+	source,start,el,currentNSMap,entityReplacer,errorHandler, isHTML
+){
 
 	/**
 	 * @param {string} qname
@@ -337,7 +350,7 @@ function parseElementStartPart(source,start,el,currentNSMap,entityReplacer,error
 					errorHandler.warning('attribute "'+value+'" missed quot(")!');
 					addAttribute(attrName, value, start)
 				}else{
-					if(!NAMESPACE.isHTML(currentNSMap['']) || !value.match(/^(?:disabled|checked|selected)$/i)){
+					if(!isHTML || !value.match(/^(?:disabled|checked|selected)$/i)){
 						errorHandler.warning('attribute "'+value+'" missed value!! "'+value+'" instead!!')
 					}
 					addAttribute(value, value, start)
@@ -385,7 +398,7 @@ function parseElementStartPart(source,start,el,currentNSMap,entityReplacer,error
 				//case S_ATTR_NOQUOT_VALUE:void();break;
 				case S_ATTR_SPACE:
 					var tagName =  el.tagName;
-					if (!NAMESPACE.isHTML(currentNSMap['']) || !attrName.match(/^(?:disabled|checked|selected)$/i)) {
+					if (!isHTML || !attrName.match(/^(?:disabled|checked|selected)$/i)) {
 						errorHandler.warning('attribute "'+attrName+'" missed value!! "'+attrName+'" instead2!!')
 					}
 					addAttribute(attrName, attrName, start);

diff --git a/test/html/normalize.test.js b/test/html/normalize.test.js
@@ -53,6 +53,27 @@ describe('html normalizer', () => {
 		})
 	})
 
+	it.each([
+		`<html xmlns="http://www.w3.org/1999/xhtml"><script>let message = " &amp; ETH";</script></html>`,
+		`<html><script>let message = " &amp; ETH";</script></html>`,
+	])(`should %s`, (xml) => {
+		const { errors, parser } = getTestParser()
+
+		const actual = parser.parseFromString(xml, 'application/xml')
+
+		expect(actual.documentElement.firstChild.textContent).toBe('let message = " & ETH";');
+	})
+	it.each([
+		`<html xmlns="http://www.w3.org/1999/xhtml"><script>let message = " &amp; ETH";</script></html>`,
+		`<html><script>let message = " &amp; ETH";</script></html>`,
+	])(`should %s`, (xml) => {
+		const { errors, parser } = getTestParser()
+
+		const actual = parser.parseFromString(xml, 'text/html')
+
+		expect(actual.documentElement.firstChild.textContent).toBe('let message = " &amp; ETH";');
+	})
+
 	it('European entities', () => {
 		const { errors, parser } = getTestParser()