Merge pull request #192 from Zegnat/implied-url

Fix implied URL parsing

Author: aaronpk

Mf2/Parser.php

@@ -1083,35 +1083,31 @@ public function parseH(\DOMElement $e, $is_backcompat = false, $has_nested_mf =
 
 		}
 
-		// Check for u-url
-		if (!array_key_exists('url', $return) && !$is_backcompat) {
-			$url = null;
-			// Look for img @src
-			if ($e->tagName == 'a' or $e->tagName == 'area') {
-				$url = $e->getAttribute('href');
-			}
-
-			// Look for nested a @href
-			foreach ($this->xpath->query('./a[count(preceding-sibling::a)+count(following-sibling::a)=0]', $e) as $em) {
-				$emNames = mfNamesFromElement($em, 'h-');
-				if (empty($emNames)) {
-					$url = $em->getAttribute('href');
-					break;
-				}
-			}
-
-			// Look for nested area @src
-			foreach ($this->xpath->query('./area[count(preceding-sibling::area)+count(following-sibling::area)=0]', $e) as $em) {
-				$emNames = mfNamesFromElement($em, 'h-');
-				if (empty($emNames)) {
-					$url = $em->getAttribute('href');
-					break;
+		// Do we need to imply a url property?
+		// if no explicit "url" property, and no other explicit u-* properties, and no nested microformats
+		if (!array_key_exists('url', $return) && !in_array('u-', $prefixes) && !$has_nested_mf && !$is_backcompat) {
+			// a.h-x[href] or area.h-x[href]
+			if (($e->tagName === 'a' || $e->tagName === 'area') && $e->hasAttribute('href')) {
+				$return['url'][] = $this->resolveUrl($e->getAttribute('href'));
+			} else {
+				$xpaths = array(
+					// .h-x>a[href]:only-of-type:not[.h-*]
+					'./a[not(contains(concat(" ", @class), " h-")) and count(../a) = 1 and @href]',
+					// .h-x>area[href]:only-of-type:not[.h-*]
+					'./area[not(contains(concat(" ", @class), " h-")) and count(../area) = 1 and @href]',
+					// .h-x>:only-child:not[.h-*]>a[href]:only-of-type:not[.h-*]
+					'./*[not(contains(concat(" ", @class), " h-")) and count(../*) = 1 and count(a) = 1]/a[not(contains(concat(" ", @class), " h-")) and @href]',
+					// .h-x>:only-child:not[.h-*]>area[href]:only-of-type:not[.h-*]
+					'./*[not(contains(concat(" ", @class), " h-")) and count(../*) = 1 and count(area) = 1]/area[not(contains(concat(" ", @class), " h-")) and @href]'
+				);
+				foreach ($xpaths as $xpath) {
+					$url = $this->xpath->query($xpath, $e);
+					if ($url !== false && $url->length === 1) {
+						$return['url'][] = $this->resolveUrl($url->item(0)->getAttribute('href'));
+						break;
+					}
 				}
 			}
-
-			if (!is_null($url)) {
-				$return['url'][] = $this->resolveUrl($url);
-			}
 		}
 
 		// Make sure things are unique and in alphabetical order

tests/Mf2/ClassicMicroformatsTest.php

@@ -704,7 +704,7 @@ public function testMixedMf2andMf1Case3() {
 		$parser = new Parser($input);
 		$result = $parser->parse();
 
-		$this->assertCount(3, $result['items'][0]['properties']);
+		$this->assertCount(2, $result['items'][0]['properties']);
 		$this->assertArrayNotHasKey('street-address', $result['items'][0]['properties']);
 		$this->assertArrayNotHasKey('locality', $result['items'][0]['properties']);
 		$this->assertArrayNotHasKey('country-name', $result['items'][0]['properties']);

tests/Mf2/ParseImpliedTest.php

@@ -351,5 +351,18 @@ public function testBackcompatNoImpliedUrl() {
 		$this->assertArrayHasKey('content', $result['items'][0]['properties']);
 	}
 
+
+	/**
+	 * Don't imply u-url if there are other u-*
+	 * @see http://microformats.org/wiki/microformats2-parsing#parsing_for_implied_properties
+	 * @see https://github.com/microformats/php-mf2/issues/183
+	 */
+	public function testNoImpliedUrl() {
+		$input = '<div class="h-entry"> <h1 class="p-name"><a href="https://example.com/this-post">Title</a></h1> <div class="e-content"> <p> blah blah blah </p> </div> <a href="https://example.org/syndicate" class="u-syndication"></a> </div>';
+		$result = Mf2\parse($input);
+
+		$this->assertArrayNotHasKey('url', $result['items'][0]['properties']);
+	}
+
 }