Still more fixes

lib/Parser.php

@@ -272,7 +272,7 @@ class Parser {
                     $out['rel-urls'][$url][$attr] = trim($link->getAttribute($attr));
                 }
             }
-            if (strlen($text = $this->getCleanText($link, "p"))) {
+            if (!isset($out['rel-urls'][$url]['text']) && strlen($text = $this->getCleanText($link, "p"))) {
                 $out['rel-urls'][$url]['text'] = $text;
             }
             # if there is no "rels" key in that hash, add it with an empty array value
@@ -979,6 +979,10 @@ class Parser {
         foreach ($copy->getElementsByTagName("style") as $e) {
             $e->parentNode->removeChild($e);
         }
+        // also drop templates; their contents would not normally be included in textContent
+        foreach ($copy->getElementsByTagName("template") as $e) {
+            $e->parentNode->removeChild($e);
+        }
         # replacing any nested <img> elements with their alt attribute, if
         #   present; otherwise their src attribute, if present, adding a
         #   space at the beginning and end, resolving the URL if it’s
@@ -1016,7 +1020,7 @@ class Parser {
      * @param bool $considerChildren Whether or not child nodes are valid next nodes
      */
     protected function nextElement(\DOMElement $node, \DOMElement $root, bool $considerChildren): ?\DOMElement {
-        if ($considerChildren && $node->localName !== "template" && $node->hasChildNodes()) {
+        if ($considerChildren && $node->hasChildNodes()) {
             $node = $node->firstChild;
             $next = $node;
         } elseif ($node->isSameNode($root)) {
@@ -1024,7 +1028,7 @@ class Parser {
          } else {
             $next = $node->nextSibling;
         }
-        while ($next && !$next instanceof \DOMElement) {
+        while ($next && (!$next instanceof \DOMElement ||  $next->localName === "template")) {
             $next = $next->nextSibling;
         }
         while (!$next) {
@@ -1033,7 +1037,7 @@ class Parser {
                 return null;
             }
             $next = $node->nextSibling;
-            while ($next and !$next instanceof \DOMElement) {
+            while ($next && (!$next instanceof \DOMElement ||  $next->localName === "template")) {
                 $next = $next->nextSibling;
             }
         }

tests/cases/StandardTest.php

@@ -45,9 +45,6 @@ class StandardTest extends \PHPUnit\Framework\TestCase {
         "microformats-v2/h-event/time",
         "microformats-v2/h-review-aggregate/hevent",
         "microformats-v2/h-review-aggregate/simpleproperties",
-        "microformats-v2/mixed/ignoretemplate",
-        "microformats-v2/rel/duplicate-rels",
-        "microformats-v2/rel/varying-text-duplicate-rels",
     ];
 
     /** @dataProvider provideStandardTests */
@@ -64,6 +61,8 @@ class StandardTest extends \PHPUnit\Framework\TestCase {
             if (preg_match('#^https?://[^/]+$#', $v)) {
                 $v .= "/";
             }
+            // at least one test has spurious whitespace
+            $v = trim($v);
         });
         // parse input
         $dom = new DOMParser;
@@ -81,7 +80,7 @@ class StandardTest extends \PHPUnit\Framework\TestCase {
         foreach (new \RegexIterator(new \RecursiveIteratorIterator(new \RecursiveDirectoryIterator(self::BASE)), '/\.json$/') as $path) {
             $path = str_replace(self::BASE, "", $path->getPathname());
             $path =  preg_replace('/\.json$/', '', $path);
-            yield [$path];
+            yield $path => [$path];
         }
     }