Make HTML content break out of foreign content

Fixes html tags being nested inside svg and math tags. Uses the proposed spec changes from whatwg/html#6399 and whatwg/html#6455
TRowbotham · Mar 25, 2021 · 84108af · 84108af
1 parent f9c880d
commit 84108af
Show file tree

Hide file tree

Showing 2 changed files with 62 additions and 26 deletions.
diff --git a/src/Parser/HTML/InsertionMode/InForeignContentInsertionMode.php b/src/Parser/HTML/InsertionMode/InForeignContentInsertionMode.php
@@ -144,36 +144,25 @@ public function processToken(TreeBuilderContext $context, Token $token): void
                 )
             ) {
                 // Parse error.
-                // If the parser was originally created for the HTML fragment
-                // parsing algorithm, then act as described in the "any other start
-                // tag" entry below. (fragment case)
-                if ($context->parser->isFragmentCase) {
-                    $this->inForeignContentAnyOtherStartTag($context, $token);
 
-                    return;
-                }
-
-                // Pop an element from the stack of open elements, and then keep
-                // popping more elements from the stack of open elements until the
-                // current node is a MathML text integration point, an HTML
-                // integration point, or an element in the HTML namespace.
+                // While the current node is not a MathML text integration point, an HTML integration point, or an
+                // element in the HTML namespace, pop elements from the stack of open elements.
                 while (!$context->parser->openElements->isEmpty()) {
-                    $context->parser->openElements->pop();
                     $currentNode = $context->parser->openElements->bottom();
 
                     if (
                         $this->isMathMLTextIntegrationPoint($currentNode)
                         || $this->isHTMLIntegrationPoint($currentNode, $context->elementTokenMap)
-                        || (
-                            $currentNode instanceof Element
-                            && $currentNode->namespaceURI === Namespaces::HTML
-                        )
+                        || $currentNode->namespaceURI === Namespaces::HTML
                     ) {
                         break;
                     }
+
+                    $context->parser->openElements->pop();
                 }
 
-                // Then, reprocess the token.
+                // Reprocess the token according to the rules given in the section corresponding to the current
+                // insertion mode in HTML content.
                 $context->insertionMode->processToken($context, $token);
 
                 return;

diff --git a/tests/html5lib/test_data/tree-construction/foreign-fragment.dat b/tests/html5lib/test_data/tree-construction/foreign-fragment.dat
@@ -7,7 +7,7 @@
 #document-fragment
 svg path
 #document
-| <svg nobr>
+| <nobr>
 |   "X"
 
 #data
@@ -17,7 +17,7 @@ svg path
 #document-fragment
 svg path
 #document
-| <svg font>
+| <font>
 |   color=""
 | "X"
 
@@ -390,7 +390,7 @@ math mtext
 #document-fragment
 math annotation-xml
 #document
-| <math div>
+| <div>
 
 #data
 <figure></figure>
@@ -407,7 +407,7 @@ math annotation-xml
 #document-fragment
 math math
 #document
-| <math div>
+| <div>
 
 #data
 <figure></figure>
@@ -461,12 +461,11 @@ svg desc
 <div><h1>X</h1></div>
 #errors
 5: HTML start tag “div” in a foreign namespace context.
-9: HTML start tag “h1” in a foreign namespace context.
 #document-fragment
 svg svg
 #document
-| <svg div>
-|   <svg h1>
+| <div>
+|   <h1>
 |     "X"
 
 #data
@@ -476,7 +475,7 @@ svg svg
 #document-fragment
 svg svg
 #document
-| <svg div>
+| <div>
 
 #data
 <div></div>
@@ -557,3 +556,51 @@ svg desc
 svg desc
 #document
 | "X"
+
+#data
+<svg><p>
+#errors
+8: HTML start tag “p” in a foreign namespace context.
+#document-fragment
+div
+#document
+| <svg svg>
+| <p>
+
+#data
+<p>
+#errors
+3: HTML start tag “p” in a foreign namespace context.
+#document-fragment
+svg svg
+#document
+| <p>
+
+#data
+<body><foo>
+#errors
+6: HTML start tag “body” in a foreign namespace context.
+#document-fragment
+svg svg
+#document
+| <svg foo>
+
+#data
+<p><foo>
+#errors
+3: HTML start tag “p” in a foreign namespace context.
+#document-fragment
+svg svg
+#document
+| <p>
+|   <foo>
+
+#data
+<p></p><foo>
+#errors
+3: HTML start tag “p” in a foreign namespace context.
+#document-fragment
+svg svg
+#document
+| <p>
+| <svg foo>