Merge pull request #3 from Leimi/fix-attributes-parsing

Fix parser bug when having "/" char in a xml node attribute
felixgirault · Jan 10, 2018 · b3438ad · b3438ad
2 parents 828b751 + a38422f
commit b3438ad
Show file tree

Hide file tree

Showing 2 changed files with 52 additions and 0 deletions.
diff --git a/src/Parser.php b/src/Parser.php
@@ -33,6 +33,9 @@ class Parser {
 	//
 	const PARSING_SELF_CLOSING_TAG = 4;
 
+	//
+	const PARSING_ATTRIBUTE_VALUE = 5 ;
+
 	//
 	protected $_observers = [];
 
@@ -45,6 +48,9 @@ class Parser {
 	//
 	protected $_tagName = '';
 
+	//
+	protected $_attributeQuoteChar = '';
+
 	//
 	protected $_state = self::PARSING_TAG_CONTENTS;
 
@@ -96,6 +102,10 @@ public function parse($xml) {
 				case self::PARSING_SELF_CLOSING_TAG:
 					$this->_parseSelfClosingTag($char);
 					break;
+
+				case self::PARSING_ATTRIBUTE_VALUE:
+					$this->_parseAttributeValue($char);
+					break;
 			}
 
 			if (!$this->_continue) {
@@ -138,6 +148,12 @@ protected function _parseOpeningTag($char) {
 	 */
 	protected function _parseTagAttributes($char) {
 		switch ($char) {
+			case '"':
+			case '\'':
+				$this->_attributeQuoteChar = $char;
+				$this->_state = self::PARSING_ATTRIBUTE_VALUE;
+				break;
+
 			case '/':
 				$this->_state = self::PARSING_SELF_CLOSING_TAG;
 				break;
@@ -150,6 +166,18 @@ protected function _parseTagAttributes($char) {
 		}
 	}
 
+	/**
+	 * we are in an attribute value: don't parse anything here
+	 * just go back to the attributes parsing function when not in the value anymore
+	 */
+	protected function _parseAttributeValue($char) {
+		switch ($char) {
+			case $this->_attributeQuoteChar:
+				$this->_state = self::PARSING_TAG_ATTRIBUTES;
+				break;
+		}
+	}
+
 	/**
 	 *
 	 */

diff --git a/tests/NokogiriTest.php b/tests/NokogiriTest.php
@@ -121,6 +121,30 @@ public function testCutWithAutoClosingTags() {
 		);
 	}
 
+	public function testAttributeValues() {
+		$html = <<<HTML
+			<p><a href="www.attribu.te/with/slash">this is a link</a></p>
+HTML;
+		$expected = <<<HTML
+			<p><a href="www.attribu.te/with/slash">this</a></p>
+HTML;
+		$this->assertEquals(
+			$expected,
+			$this->Nokogiri->cut($html, 4, true)
+		);
+
+		$html = <<<HTML
+			<p><a title='simple quotes' aria-label="double quotes">this is a link</a></p>
+HTML;
+		$expected = <<<HTML
+			<p><a title='simple quotes' aria-label="double quotes">this</a></p>
+HTML;
+		$this->assertEquals(
+			$expected,
+			$this->Nokogiri->cut($html, 4, true)
+		);
+	}
+
 	/**
 	 *
 	 */