Masterminds · bytestream · May 21, 2020 · Jan 11, 2023 · Jan 18, 2023 · Jan 18, 2023
diff --git a/.github/workflows/cs.yaml b/.github/workflows/cs.yaml
@@ -28,5 +28,5 @@ jobs:
 
       - name: cs fix
         run: |
-          wget -q https://github.com/FriendsOfPHP/PHP-CS-Fixer/releases/download/v2.13.1/php-cs-fixer.phar 
+          wget -q https://github.com/FriendsOfPHP/PHP-CS-Fixer/releases/download/v2.19.0/php-cs-fixer.phar 
           php php-cs-fixer.phar fix --dry-run --diff
diff --git a/.php_cs.dist b/.php_cs.dist
@@ -9,6 +9,7 @@ return PhpCsFixer\Config::create()
         '@Symfony' => true,
         'concat_space' => array('spacing' => 'one'),
         'phpdoc_annotation_without_dot' => false,
+        'array_syntax' => ['syntax' => 'long'],
     ))
     ->setFinder($finder)
 ;
diff --git a/src/HTML5.php b/src/HTML5.php
@@ -146,7 +146,6 @@ public function hasErrors()
      * Parse an input string.
      *
      * @param string $input
-     * @param array  $options
      *
      * @return \DOMDocument
      */

diff --git a/src/HTML5/Parser/DOMTreeBuilder.php b/src/HTML5/Parser/DOMTreeBuilder.php
@@ -231,8 +231,6 @@ public function fragment()
      *
      * This is used for handling Processor Instructions as they are
      * inserted. If omitted, PI's are inserted directly into the DOM tree.
-     *
-     * @param InstructionProcessor $proc
      */
     public function setInstructionProcessor(InstructionProcessor $proc)
     {
@@ -302,12 +300,28 @@ public function startTag($name, $attributes = array(), $selfClosing = false)
             case 'head':
                 if ($this->insertMode > static::IM_BEFORE_HEAD) {
                     $this->parseError('Unexpected head tag outside of head context.');
+                    // https://html.spec.whatwg.org/multipage/parsing.html#parsing-main-inbody
+                    // A start tag whose tag name is one of: "caption", "col", "colgroup", "frame", "head", "tbody",
+                    // "td", "tfoot", "th", "thead", "tr"
+                    // Parse error. Ignore the token.
+                    return 0;
                 } else {
                     $this->insertMode = static::IM_IN_HEAD;
                 }
                 break;
             case 'body':
-                $this->insertMode = static::IM_IN_BODY;
+                if ($this->insertMode >= static::IM_IN_BODY) {
+                    $this->parseError('Unexpected body tag outside of body context.');
+                    // https://html.spec.whatwg.org/multipage/parsing.html#parsing-main-inbody
+                    // A start tag whose tag name is "body"
+                    // Parse error.
+                    // If the second element on the stack of open elements is not a body element, if the stack of open elements has only one node on it, or if there is a template element on the stack of open elements, then ignore the token. (fragment case)
+                    // Otherwise, set the frameset-ok flag to "not ok"; then, for each attribute on the token, check to see if the attribute is already present on the body element (the second element) on the stack of open elements, and if it is not, add the attribute and its corresponding value to that element.
+                    return 0;
+                } else {
+                    $this->insertMode = static::IM_IN_BODY;
+                }
+
                 break;
             case 'svg':
                 $this->insertMode = static::IM_IN_SVG;
@@ -322,6 +336,12 @@ public function startTag($name, $attributes = array(), $selfClosing = false)
                 break;
         }
 
+        // Case when no <body> exists, note section on 'Anything else' below.
+        // https://html.spec.whatwg.org/multipage/parsing.html#the-after-head-insertion-mode
+        if ($this->insertMode === static::IM_AFTER_HEAD && 'head' !== $name && 'body' !== $name) {
+            $this->startTag('body');
+        }
+
         // Special case handling for SVG.
         if ($this->insertMode === static::IM_IN_SVG) {
             $lname = Elements::normalizeSvgElement($lname);
@@ -535,10 +555,18 @@ public function endTag($name)
 
         switch ($lname) {
             case 'head':
-                $this->insertMode = static::IM_AFTER_HEAD;
+                if ($this->insertMode <= static::IM_AFTER_HEAD) {
+                    $this->insertMode = static::IM_AFTER_HEAD;
+                } else {
+                    $this->parseError('Closing head tag encountered but not in head context.');
+                }
                 break;
             case 'body':
-                $this->insertMode = static::IM_AFTER_BODY;
+                if ($this->insertMode <= static::IM_AFTER_BODY || $this->insertMode >= static::IM_IN_SVG) {
+                    $this->insertMode = static::IM_AFTER_BODY;
+                } else {
+                    $this->parseError('Closing body tag encountered but not in body context.');
+                }
                 break;
             case 'svg':
             case 'mathml':
@@ -556,21 +584,20 @@ public function comment($cdata)
 
     public function text($data)
     {
-        // XXX: Hmmm.... should we really be this strict?
+        // https://html.spec.whatwg.org/multipage/parsing.html#the-before-head-insertion-mode
         if ($this->insertMode < static::IM_IN_HEAD) {
             // Per '8.2.5.4.3 The "before head" insertion mode' the characters
-            // " \t\n\r\f" should be ignored but no mention of a parse error. This is
-            // practical as most documents contain these characters. Other text is not
-            // expected here so recording a parse error is necessary.
+            // " \t\n\r\f" should be ignored .
             $dataTmp = trim($data, " \t\n\r\f");
             if (!empty($dataTmp)) {
-                // fprintf(STDOUT, "Unexpected insert mode: %d", $this->insertMode);
-                $this->parseError('Unexpected text. Ignoring: ' . $dataTmp);
+                $this->startTag('head');
+                $this->endTag('head');
+                $this->startTag('body');
+            } else {
+                return;
             }
-
-            return;
         }
-        // fprintf(STDOUT, "Appending text %s.", $data);
+
         $node = $this->doc->createTextNode($data);
         $this->current->appendChild($node);
     }

diff --git a/src/HTML5/Parser/Tokenizer.php b/src/HTML5/Parser/Tokenizer.php
@@ -726,6 +726,7 @@ protected function isCommentEnd()
         // Test for '!>'
         if ('!' == $this->scanner->current() && '>' == $this->scanner->peek()) {
             $this->scanner->consume(); // Consume the last '>'
+
             return true;
         }
         // Unread '-' and one of '!' or '>';

diff --git a/src/HTML5/Parser/TreeBuildingRules.php b/src/HTML5/Parser/TreeBuildingRules.php
@@ -80,7 +80,6 @@ public function evaluate($new, $current)
             case 'thead':
             case 'tfoot':
             case 'table': // Spec isn't explicit about this, but it's necessary.
-
                 return $this->closeIfCurrentMatches($new, $current, array(
                     'thead',
                     'tfoot',

diff --git a/test/HTML5/Html5Test.php b/test/HTML5/Html5Test.php
@@ -492,4 +492,47 @@ public function testAnchorTargetQueryParam()
             $res
         );
     }
+
+    /**
+     * Test for issue #166.
+     *
+     * @dataProvider tagOmissionProvider
+     */
+    public function testTagOmission($input, $expected)
+    {
+        $doc = $this->html5->loadHTML($input);
+        $this->assertCount(0, $this->html5->getErrors());
+
+        $out = $this->html5->saveHTML($doc);
+        $this->assertRegExp('|' . preg_quote($expected, '|') . '|', $out);
+    }
+
+    /**
+     * Tag omission test cases.
+     *
+     * @return \string[][]
+     */
+    public function tagOmissionProvider()
+    {
+        return array(
+            array(
+                '<!DOCTYPE html><html>Hello, This is a test.<br />Does it work this time?</html>',
+                '<html><head></head><body>Hello, This is a test.<br>Does it work this time?</body></html>',
+            ),
+            // test whitespace (\n)
+            array(
+                '<!DOCTYPE html>
+<html>
+<head><meta http-equiv="Content-Type" content="text/html; charset=utf-8"></head>
+<body>
+<br>
+</body>
+</html>',
+                '<head><meta http-equiv="Content-Type" content="text/html; charset=utf-8"></head>
+<body>
+<br>
+</body>',
+            ),
+        );
+    }
 }
diff --git a/test/HTML5/Parser/DOMTreeBuilderTest.php b/test/HTML5/Parser/DOMTreeBuilderTest.php
@@ -6,9 +6,9 @@
 
 namespace Masterminds\HTML5\Tests\Parser;
 
+use Masterminds\HTML5\Parser\DOMTreeBuilder;
 use Masterminds\HTML5\Parser\Scanner;
 use Masterminds\HTML5\Parser\Tokenizer;
-use Masterminds\HTML5\Parser\DOMTreeBuilder;
 
 /**
  * These tests are functional, not necessarily unit tests.
@@ -457,14 +457,34 @@ public function testText()
         $data = $wrapper->childNodes->item(0);
         $this->assertEquals(XML_TEXT_NODE, $data->nodeType);
         $this->assertEquals('test', $data->data);
+    }
 
+    public function testTextBeforeHeadNotAllowed()
+    {
         // The DomTreeBuilder has special handling for text when in before head mode.
-        $html = '<!DOCTYPE html><html>
-    Foo<head></head><body></body></html>';
+        $html = '<!DOCTYPE html><html>Foo<head></head><body>test</body></html>';
         $doc = $this->parse($html);
-        $this->assertEquals('Line 0, Col 0: Unexpected text. Ignoring: Foo', $this->errors[0]);
-        $headElement = $doc->documentElement->firstChild;
-        $this->assertEquals('head', $headElement->tagName);
+
+        $this->assertContains('Line 0, Col 0: Unexpected body tag outside of body context.', $this->errors);
+        $this->assertXmlStringEqualsXmlString($doc, '<html xmlns="http://www.w3.org/1999/xhtml"><head/><body>Footest</body></html>');
+    }
+
+    public function testHeadInBodyTriggersParseError()
+    {
+        $html = '<!DOCTYPE html><html><head></head><body><head></head>test</body></html>';
+        $doc = $this->parse($html);
+
+        $this->assertContains('Line 0, Col 0: Unexpected head tag outside of head context.', $this->errors);
+        $this->assertXmlStringEqualsXmlString($doc, '<html xmlns="http://www.w3.org/1999/xhtml"><head/><body>test</body></html>');
+    }
+
+    public function testBodyInBodyTriggersParseError()
+    {
+        $html = '<!DOCTYPE html><html><head></head><body>test<body>ba<br/>z</body></body></html>';
+        $doc = $this->parse($html);
+
+        $this->assertContains('Line 0, Col 0: Unexpected body tag outside of body context.', $this->errors);
+        $this->assertXmlStringEqualsXmlString($doc, '<html xmlns="http://www.w3.org/1999/xhtml"><head/><body>testba<br/>z</body></html>');
     }
 
     public function testParseErrors()

diff --git a/test/HTML5/Parser/ScannerTest.php b/test/HTML5/Parser/ScannerTest.php
@@ -6,8 +6,8 @@
 
 namespace Masterminds\HTML5\Tests\Parser;
 
-use Masterminds\HTML5\Parser\StringInputStream;
 use Masterminds\HTML5\Parser\Scanner;
+use Masterminds\HTML5\Parser\StringInputStream;
 
 class ScannerTest extends \Masterminds\HTML5\Tests\TestCase
 {

diff --git a/test/HTML5/Parser/TokenizerTest.php b/test/HTML5/Parser/TokenizerTest.php
@@ -2,9 +2,9 @@
 
 namespace Masterminds\HTML5\Tests\Parser;
 
-use Masterminds\HTML5\Parser\UTF8Utils;
 use Masterminds\HTML5\Parser\Scanner;
 use Masterminds\HTML5\Parser\Tokenizer;
+use Masterminds\HTML5\Parser\UTF8Utils;
 
 class TokenizerTest extends \Masterminds\HTML5\Tests\TestCase
 {

diff --git a/test/HTML5/Parser/TreeBuildingRulesTest.php b/test/HTML5/Parser/TreeBuildingRulesTest.php
@@ -6,10 +6,10 @@
 
 namespace Masterminds\HTML5\Tests\Parser;
 
-use Masterminds\HTML5\Parser\TreeBuildingRules;
-use Masterminds\HTML5\Parser\Tokenizer;
-use Masterminds\HTML5\Parser\Scanner;
 use Masterminds\HTML5\Parser\DOMTreeBuilder;
+use Masterminds\HTML5\Parser\Scanner;
+use Masterminds\HTML5\Parser\Tokenizer;
+use Masterminds\HTML5\Parser\TreeBuildingRules;
 
 /**
  * These tests are functional, not necessarily unit tests.

diff --git a/test/HTML5/Serializer/OutputRulesTest.php b/test/HTML5/Serializer/OutputRulesTest.php
@@ -2,9 +2,9 @@
 
 namespace Masterminds\HTML5\Tests\Serializer;
 
+use Masterminds\HTML5;
 use Masterminds\HTML5\Serializer\OutputRules;
 use Masterminds\HTML5\Serializer\Traverser;
-use Masterminds\HTML5;
 
 class OutputRulesTest extends \Masterminds\HTML5\Tests\TestCase
 {