Implement XML infoset coercion

lib/DOM/Document.php

@@ -96,12 +96,13 @@ class Document extends \DOMDocument {
                 $e->content = $this->createDocumentFragment();
             }
             return $e;
-        } catch (\DOMException) {
+        } catch (\DOMException $e) {
             // The element name is invalid for XML
             // Replace any offending characters with "UHHHHHH" where H is the
             //   uppercase hexadecimal digits of the character's code point
             $this->mangledElements = true;
-            $name = $this->CoerceName($name);
+            $name = $this->coerceName($name);
+            return parent::createElement($name, $value);
         }
     }
 
@@ -113,13 +114,39 @@ class Document extends \DOMDocument {
                 $e->content = $this->createDocumentFragment();
             }
             return $e;
-        } catch (\DOMException) {
-            throw $e;
+        } catch (\DOMException $e) {
             // The element name is invalid for XML
-            // Replace any offending characters with "UHHHHHH" where H is the
+            // Replace any offending characters with "UHHHHHH" where H are the
             //   uppercase hexadecimal digits of the character's code point
             $this->mangledElements = true;
-            $qualifiedName = $this->CoerceName($qualifiedName);
+            $qualifiedName = $this->coerceName($qualifiedName);
+            return parent::createElementNS($namespaceURI, $qualifiedName, $value);
+        }
+    }
+
+    public function createAttribute($name) {
+        try {
+            return parent::createAttribute($name);
+        } catch (\DOMException $e) {
+            // The element name is invalid for XML
+            // Replace any offending characters with "UHHHHHH" where H are the
+            //   uppercase hexadecimal digits of the character's code point
+            $this->mangledAttributes = true;
+            $name = $this->coerceName($name);
+            return parent::createAttribute($name);
+        }
+    }
+
+    public function createAttributeNS($namespaceURI, $qualifiedName) {
+        try {
+            return parent::createAttributeNS($namespaceURI, $qualifiedName);
+        } catch (\DOMException $e) {
+            // The element name is invalid for XML
+            // Replace any offending characters with "UHHHHHH" where H are the
+            //   uppercase hexadecimal digits of the character's code point
+            $this->mangledAttributes = true;
+            $qualifiedName = $this->coerceName($qualifiedName);
+            return parent::createAttributeNS($namespaceURI, $qualifiedName);
         }
     }
 

lib/DOM/Element.php

@@ -12,6 +12,32 @@ class Element extends \DOMElement {
 
     protected const SELF_CLOSING_ELEMENTS = ['area', 'base', 'basefont', 'bgsound', 'br', 'col', 'embed', 'frame', 'hr', 'img', 'input', 'link', 'meta', 'param', 'source', 'track', 'wbr'];
 
+    public function setAttribute($name, $value) {
+        try {
+            parent::setAttribute($name, $value);
+        } catch (\DOMException $e) {
+            // The attribute name is invalid for XML
+            // Replace any offending characters with "UHHHHHH" where H are the
+            //   uppercase hexadecimal digits of the character's code point
+            $this->ownerDocument->mangledAttributes = true;
+            $name = $this->coerceName($name);
+            parent::setAttribute($name, $value);
+        }
+    }
+    
+    public function setAttributeNS($namespaceURI, $qualifiedName, $value) {
+        try {
+            parent::setAttributeNS($namespaceURI, $qualifiedName, $value);
+        } catch (\DOMException $e) {
+            // The attribute name is invalid for XML
+            // Replace any offending characters with "UHHHHHH" where H are the
+            //   uppercase hexadecimal digits of the character's code point
+            $this->ownerDocument->mangledAttributes = true;
+            $qualifiedName = $this->coerceName($qualifiedName);
+            parent::setAttributeNS($namespaceURI, $qualifiedName, $value);
+        }
+    }
+    
     public function isMathMLTextIntegrationPoint(): bool {
         return (
             $this->namespaceURI === Parser::MATHML_NAMESPACE && (

lib/DOM/traits/EscapeString.php

@@ -23,12 +23,30 @@ trait EscapeString {
         return $string;
     }
 
-    protected function CoerceName(string $name): string {
+    protected function coerceName(string $name): string {
         // This matches the inverse of the production of NameChar in XML 1.0,
         //   with the added exclusion of ":" from allowed characters
         // See https://www.w3.org/TR/REC-xml/#NT-NameStartChar
-        preg_match_all('/[^\-\.0-9\x{B7}\x{300}-\x{36F}\x{203F}-\x{2040}A-Za-z_\x{C0}-\x{D6}\x{D8}-\x{F6}\x{F8}-\x{2FF}\x{370}-\x{37D}\x{37F}-\x{1FFF}\x{200C}-\x{200D}\x{2070}-\x{218F}\x{2C00}-\x{2FEF}\x{3001}-\x{D7FF}\x{F900}-\x{FDCF}\x{FDF0}-\x{FFFD}\x{10000}-\x{EFFFF}]/u', $name, $m, \PREG_OFFSET_CAPTURE);
-        var_export($m);
-        exit;
+        preg_match_all('/[^\-\.0-9\x{B7}\x{300}-\x{36F}\x{203F}-\x{2040}A-Za-z_\x{C0}-\x{D6}\x{D8}-\x{F6}\x{F8}-\x{2FF}\x{370}-\x{37D}\x{37F}-\x{1FFF}\x{200C}-\x{200D}\x{2070}-\x{218F}\x{2C00}-\x{2FEF}\x{3001}-\x{D7FF}\x{F900}-\x{FDCF}\x{FDF0}-\x{FFFD}\x{10000}-\x{EFFFF}]/u', $name, $m);
+        foreach (array_unique($m[0], \SORT_STRING) as $c) {
+            $esc = "U".str_pad(strtoupper(dechex(\IntlChar::ord($c))), 6, "0", \STR_PAD_LEFT);
+            $name = str_replace($c, $esc, $name);
+        }
+        // Apply stricter rules to the first character
+        if (preg_match('/^[^A-Za-z_\x{C0}-\x{D6}\x{D8}-\x{F6}\x{F8}-\x{2FF}\x{370}-\x{37D}\x{37F}-\x{1FFF}\x{200C}-\x{200D}\x{2070}-\x{218F}\x{2C00}-\x{2FEF}\x{3001}-\x{D7FF}\x{F900}-\x{FDCF}\x{FDF0}-\x{FFFD}\x{10000}-\x{EFFFF}]/u', $name, $m)) {
+            $c = $m[0];
+            $esc = "U".str_pad(strtoupper(dechex(\IntlChar::ord($c))), 6, "0", \STR_PAD_LEFT);
+            $name = $esc.substr($name, strlen($c));
+        }
+        return $name;
+    }
+
+    protected function uncoerceName(string $name): string {
+        preg_match_all('/U[0-9A-F]{6}/', $name, $m);
+        foreach (array_unique($m[0], \SORT_STRING) as $o) {
+            $c = \IntlChar::chr(hexdec(substr($o, 1)));
+            $name = str_replace($o, $c, $name);
+        }
+        return $name;
     }
 }

lib/TreeBuilder.php

@@ -3,7 +3,7 @@ declare(strict_types=1);
 namespace dW\HTML5;
 
 class TreeBuilder {
-    use ParseErrorEmitter;
+    use ParseErrorEmitter, EscapeString;
 
     public $debugLog = "";
 
@@ -292,6 +292,12 @@ class TreeBuilder {
             $this->debugLog .= "EMITTED: ".constant(get_class($token)."::NAME")."\n";
             return true;
         })());
+
+        // If element name coercison has occurred at some earlier point,
+        //   we must coerce all end tag names to match mangled start tags
+        if ($token instanceof EndTagToken && $this->DOM->mangledElements) {
+            $token->name = $this->coerceName($token->name);
+        }
         // Loop used for reprocessing.
         $iterations = 0;
         while (true) {
@@ -373,6 +379,14 @@ class TreeBuilder {
             return true;
         })());
 
+        // If attribute name coercison has occurred at some earlier point,
+        //   we must coerce all attributes on html and body start tags in
+        //   case they are relocated to existing elements
+        if ($token instanceof StartTagToken && $this->DOM->mangledAttributes && in_array($token->name, ["html", "body"])) {
+            foreach ($token->attributes as $attr) {
+                $attr->name = $this->coerceName($attr->name);
+            }
+        }
         # 13.2.6.4. The rules for parsing tokens in HTML content
         # 13.2.6.4.1. The "initial" insertion mode
         if ($insertionMode === self::INITIAL_MODE) {
@@ -4198,7 +4212,18 @@ class TreeBuilder {
         $element = $document->createElementNS($namespace, $localName);
         # Append each attribute in the given token to element.
         foreach ($token->attributes as $attr) {
-            $element->setAttributeNS(null, $attr->name, $attr->value);
+            $ns = null;
+            if ($namespace) {
+                // Determine the namespace URI for the prefix, if any
+                if (strpos($attr->name, "xml:") === 0) {
+                    $ns = Parser::XML_NAMESPACE;
+                } elseif (strpos($attr->name, "xmlns:") === 0) {
+                    $ns = Parser::XMLNS_NAMESPACE;
+                } elseif (strpos($attr->name, "xlink:") === 0) {
+                    $ns = Parser::XLINK_NAMESPACE;
+                }
+            }
+            $element->setAttributeNS($ns, $attr->name, $attr->value);
         }
         # If element has an xmlns attribute in the XMLNS namespace whose value
         #   is not exactly the same as the element's namespace, that is a

tests/cases/TestTreeConstructor.php

@@ -22,6 +22,8 @@ use dW\HTML5\TreeBuilder;
  * @covers \dW\HTML5\Stack
  */
 class TestTreeConstructor extends \PHPUnit\Framework\TestCase {
+    use \dW\HTML5\EscapeString;
+
     protected $out;
     protected $depth;
 
@@ -102,19 +104,6 @@ class TestTreeConstructor extends \PHPUnit\Framework\TestCase {
                 }
             }
         }
-        if (in_array($data, [
-            //'<!DOCTYPE html><html xml:lang=bar><html xml:lang=foo>',
-            //'<!DOCTYPE html><body xlink:href=foo><svg xlink:href=foo></svg>',
-            //'<!DOCTYPE html><body xlink:href=foo xml:lang=en><svg><g xml:lang=en xlink:href=foo></g></svg>',
-            //'<!DOCTYPE html><body xlink:href=foo xml:lang=en><svg><g xml:lang=en xlink:href=foo /></svg>',
-            //'<!DOCTYPE html><body xlink:href=foo xml:lang=en><svg><g xml:lang=en xlink:href=foo />bar</svg>',
-            //'<!DOCTYPE html><body xlink:href=foo><math xlink:href=foo></math>',
-            //'<!DOCTYPE html><body xlink:href=foo xml:lang=en><math><mi xml:lang=en xlink:href=foo></mi></math>',
-            //'<!DOCTYPE html><body xlink:href=foo xml:lang=en><math><mi xml:lang=en xlink:href=foo /></math>',
-            //'<!DOCTYPE html><body xlink:href=foo xml:lang=en><math><mi xml:lang=en xlink:href=foo />bar</math>',
-        ])) {
-            $skip = 'Requires implementation of the "Coercing an HTML DOM into an infoset" specification section';
-        }
         return [$exp, $patched, $skip];
     }
 
@@ -167,13 +156,20 @@ class TestTreeConstructor extends \PHPUnit\Framework\TestCase {
         } else {
             $prefix = "";
         }
-        $this->push("<".$prefix.$e->localName.">");
+        $localName = $this->uncoerceName($e->localName);
+        $this->push("<".$prefix.$localName.">");
         $this->depth++;
         $attr = [];
         foreach ($e->attributes as $a) {
-            $attr[$a->name] = $a->value;
+            $prefix = "";
+            if ($a->namespaceURI) {
+                $prefix = Parser::NAMESPACE_MAP[$a->namespaceURI];
+                assert((bool) $prefix, new \Exception("Prefix for namespace {$a->namespaceURI} is not defined"));
+                $prefix .= " ";
+            }
+            $attr[$prefix.$this->uncoerceName($a->name)] = $a->value;
         }
-        ksort($attr);
+        ksort($attr, \SORT_STRING);
         foreach ($attr as $k => $v) {
             $this->push($k.'="'.$v.'"');
         }