MensBeam
/
HTML-Parser
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
Browse Source

Simplify tree construction slightly

ns
J. King 3 years ago
parent
ce127d0192
commit
354f5acf0a
3 changed files with 120 additions and 102 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Unified View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 4
      lib/Parser.php
  2. 214
      lib/TreeBuilder.php
  3. 4
      tests/cases/TestTreeConstructor.php

4
lib/Parser.php
View File

@ -35,9 +35,7 @@ class Parser {
$errorHandler->setHandler(); $errorHandler->setHandler();
try { try {
// run the parser to completion // run the parser to completion
foreach ($tokenList as $token) { $treeBuilder->constructTree();
$treeBuilder->emitToken($token);
}
} finally { } finally {
// Restore error handling // Restore error handling
$errorHandler->clearHandler(); $errorHandler->clearHandler();

214
lib/TreeBuilder.php
View File

@ -288,76 +288,61 @@ class TreeBuilder {
} }
} }
public function emitToken(Token $token): void { public function constructTree(): void {
assert((function() use ($token) { foreach ($this->tokenList as $token) {
$this->debugLog .= "EMITTED: ".constant(get_class($token)."::NAME")."\n"; assert((function() use ($token) {
return true; $this->debugLog .= "EMITTED: ".constant(get_class($token)."::NAME")."\n";
})()); return true;
})());
// If element name coercison has occurred at some earlier point, // If element name coercison has occurred at some earlier point,
// we must coerce all end tag names to match mangled start tags // we must coerce all end tag names to match mangled start tags
if ($token instanceof EndTagToken && $this->DOM->mangledElements) { if ($token instanceof EndTagToken && $this->DOM->mangledElements) {
$token->name = $this->coerceName($token->name); $token->name = $this->coerceName($token->name);
} }
// Loop used for reprocessing. ProcessToken:
$iterations = 0;
while (true) {
assert($iterations++ < 50, new LoopException("Probable infinite loop detected in HTML content handling (outer reprocessing)"));
$adjustedCurrentNode = $this->stack->adjustedCurrentNode; $adjustedCurrentNode = $this->stack->adjustedCurrentNode;
$adjustedCurrentNodeName = $this->stack->adjustedCurrentNodeName; $adjustedCurrentNodeName = $this->stack->adjustedCurrentNodeName;
assert(!$adjustedCurrentNode || $adjustedCurrentNodeName, new \Exception("The adjusted current node must have a name if not null"));
$adjustedCurrentNodeNamespace = $this->stack->adjustedCurrentNodeNamespace; $adjustedCurrentNodeNamespace = $this->stack->adjustedCurrentNodeNamespace;
# 13.2.6 Tree construction # 13.2.6 Tree construction
# #
# As each token is emitted from the tokenizer, the user agent must follow the # As each token is emitted from the tokenizer, the user agent must follow the
# appropriate steps from the following list, known as the tree construction dispatcher: # appropriate steps from the following list, known as the tree construction dispatcher:
# if (
# If the stack of open elements is empty # If the stack of open elements is empty
if (count($this->stack) === 0 || !$this->stack->currentNode
# If the adjusted current node is an element in the HTML namespace # If the adjusted current node is an element in the HTML namespace
// PHP's DOM returns null when the namespace isn't specified... eg. HTML. // DEVIATION: For the purposes of this implementation the HTML namespace is null
is_null($adjustedCurrentNodeNamespace) || ( // rather than the XHTML namespace
# If the adjusted current node is a MathML text integration point and the token is a || $adjustedCurrentNodeNamespace === null
# start tag whose tag name is neither "mglyph" nor "malignmark" # If the adjusted current node is a MathML text integration
# If the adjusted current node is a MathML text integration point and the token is a # point and the token is a start tag whose tag name is
# character token # neither "mglyph" nor "malignmark"
$adjustedCurrentNode->isMathMLTextIntegrationPoint() && (( # If the adjusted current node is a MathML text integration
$token instanceof StartTagToken && ( # point and the token is a character token
$token->name !== 'mglyph' && $token->name !== 'malignmark' || ($adjustedCurrentNode->isMathMLTextIntegrationPoint() && (($token instanceof StartTagToken && ($token->name !== 'mglyph' && $token->name !== 'malignmark') || $token instanceof CharacterToken)))
) || # If the adjusted current node is an annotation-xml element
$token instanceof CharacterToken # in the MathML namespace and the token is a start tag
) # whose tag name is "svg"
) || ($adjustedCurrentNodeNamespace === Parser::MATHML_NAMESPACE && $adjustedCurrentNodeName === 'annotation-xml' && $token instanceof StartTagToken && $token->name === 'svg')
) || ( # If the adjusted current node is an HTML integration point
# If the adjusted current node is an annotation-xml element in the MathML namespace and # and the token is a start tag
# the token is a start tag whose tag name is "svg" # If the adjusted current node is an HTML integration point
$adjustedCurrentNodeNamespace === Parser::MATHML_NAMESPACE && # and the token is a character token
$adjustedCurrentNodeName === 'annotation-xml' && || ($adjustedCurrentNode->isHTMLIntegrationPoint() && ($token instanceof StartTagToken || $token instanceof CharacterToken))
$token instanceof StartTagToken && # If the token is an end-of-file token
$token->name === 'svg' || $token instanceof EOFToken
) || ( ) {
# If the adjusted current node is an HTML integration point and the token is a start tag # Process the token according to the rules given in the section
# If the adjusted current node is an HTML integration point and the token is a character # corresponding to the current insertion mode in HTML content.
# token
$adjustedCurrentNode->isHTMLIntegrationPoint() && (
$token instanceof StartTagToken || $token instanceof CharacterToken
)
) ||
# If the token is an end-of-file token
$token instanceof EOFToken) {
# Process the token according to the rules given in the section corresponding to
# the current insertion mode in HTML content.
$this->parseTokenInHTMLContent($token); $this->parseTokenInHTMLContent($token);
} }
# Otherwise # Otherwise
else { else {
# Process the token according to the rules given in the section for parsing # Process the token according to the rules given in the section
# tokens in foreign content. # for parsing tokens in foreign content.
// Returns false when needing to reprocess. $this->parseTokenInForeignContent($token);
if ($this->parseTokenInForeignContent($token) === false) {
continue;
}
} }
# When a start tag token is emitted with its self-closing flag set, if the flag # When a start tag token is emitted with its self-closing flag set, if the flag
# is not acknowledged when it is processed by the tree construction stage, that # is not acknowledged when it is processed by the tree construction stage, that
@ -365,7 +350,6 @@ class TreeBuilder {
if ($token instanceof StartTagToken && $token->selfClosing && !$token->selfClosingAcknowledged) { if ($token instanceof StartTagToken && $token->selfClosing && !$token->selfClosingAcknowledged) {
$this->error(ParseError::NON_VOID_HTML_ELEMENT_START_TAG_WITH_TRAILING_SOLIDUS, $token->name); $this->error(ParseError::NON_VOID_HTML_ELEMENT_START_TAG_WITH_TRAILING_SOLIDUS, $token->name);
} }
break;
} }
} }
@ -624,7 +608,8 @@ class TreeBuilder {
# A start tag whose tag name is "html" # A start tag whose tag name is "html"
elseif ($token instanceof StartTagToken && $token->name === 'html') { elseif ($token instanceof StartTagToken && $token->name === 'html') {
# Process the token using the rules for the "in body" insertion mode. # Process the token using the rules for the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# A start tag whose tag name is "head" # A start tag whose tag name is "head"
elseif ($token instanceof StartTagToken && $token->name === 'head') { elseif ($token instanceof StartTagToken && $token->name === 'head') {
@ -679,7 +664,8 @@ class TreeBuilder {
# A start tag whose tag name is "html" # A start tag whose tag name is "html"
if ($token->name === 'html') { if ($token->name === 'html') {
# Process the token using the rules for the "in body" insertion mode. # Process the token using the rules for the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# A start tag whose tag name is one of: "base", "basefont", "bgsound", "link" # A start tag whose tag name is one of: "base", "basefont", "bgsound", "link"
elseif ($token->name === 'base' || $token->name === 'basefont' || $token->name === 'bgsound' || $token->name === 'link') { elseif ($token->name === 'base' || $token->name === 'basefont' || $token->name === 'bgsound' || $token->name === 'link') {
@ -872,13 +858,15 @@ class TreeBuilder {
# A start tag whose tag name is "html" # A start tag whose tag name is "html"
if ($token->name === 'html') { if ($token->name === 'html') {
# Process the token using the rules for the "in body" insertion mode. # Process the token using the rules for the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# A start tag whose tag name is one of: "basefont", "bgsound", "link", "meta", # A start tag whose tag name is one of: "basefont", "bgsound", "link", "meta",
# "noframes", "style" # "noframes", "style"
elseif ($token->name === 'basefont' || $token->name === 'bgsound' || $token->name === 'link' || $token->name === 'meta' || $token->name === 'noframes' || $token->name === 'style'){ elseif ($token->name === 'basefont' || $token->name === 'bgsound' || $token->name === 'link' || $token->name === 'meta' || $token->name === 'noframes' || $token->name === 'style'){
# Process the token using the rules for the "in head" insertion mode. # Process the token using the rules for the "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# A start tag whose tag name is one of: "head", "noscript" # A start tag whose tag name is one of: "head", "noscript"
elseif ($token->name === 'head' || $token->name === 'noscript') { elseif ($token->name === 'head' || $token->name === 'noscript') {
@ -920,7 +908,8 @@ class TreeBuilder {
# A comment token # A comment token
elseif ($token instanceof CommentToken || $token instanceof WhitespaceToken) { elseif ($token instanceof CommentToken || $token instanceof WhitespaceToken) {
# Process the token using the rules for the "in head" insertion mode. # Process the token using the rules for the "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# Anything else # Anything else
else { else {
@ -966,7 +955,8 @@ class TreeBuilder {
# A start tag whose tag name is "html" # A start tag whose tag name is "html"
if ($token->name === 'html') { if ($token->name === 'html') {
# Process the token using the rules for the "in body" insertion mode. # Process the token using the rules for the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# A start tag whose tag name is "body" # A start tag whose tag name is "body"
elseif ($token->name === 'body') { elseif ($token->name === 'body') {
@ -1018,7 +1008,8 @@ class TreeBuilder {
# An end tag whose tag name is "template" # An end tag whose tag name is "template"
if ($token->name === 'template') { if ($token->name === 'template') {
# Process the token using the rules for the "in head" insertion mode. # Process the token using the rules for the "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# An end tag whose tag name is one of: "body", "html", "br" # An end tag whose tag name is one of: "body", "html", "br"
elseif ($token->name === 'body' || $token->name === 'html' || $token->name === 'br') { elseif ($token->name === 'body' || $token->name === 'html' || $token->name === 'br') {
@ -1107,7 +1098,8 @@ class TreeBuilder {
# "meta", "noframes", "script", "style", "template", "title" # "meta", "noframes", "script", "style", "template", "title"
elseif ($token->name === 'base' || $token->name === 'basefont' || $token->name === 'bgsound' || $token->name === 'link' || $token->name === 'meta' || $token->name === 'noframes' || $token->name === 'script' || $token->name === 'style' || $token->name === 'template' || $token->name === 'title') { elseif ($token->name === 'base' || $token->name === 'basefont' || $token->name === 'bgsound' || $token->name === 'link' || $token->name === 'meta' || $token->name === 'noframes' || $token->name === 'script' || $token->name === 'style' || $token->name === 'template' || $token->name === 'title') {
# Process the token using the rules for the "in head" insertion mode. # Process the token using the rules for the "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# A start tag whose tag name is "body" # A start tag whose tag name is "body"
elseif ($token->name === 'body') { elseif ($token->name === 'body') {
@ -1669,7 +1661,8 @@ class TreeBuilder {
# An end tag whose tag name is "template" # An end tag whose tag name is "template"
if ($token->name === 'template') { if ($token->name === 'template') {
# Process the token using the rules for the "in head" insertion mode. # Process the token using the rules for the "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# An end tag whose tag name is "body" # An end tag whose tag name is "body"
# An end tag whose tag name is "html" # An end tag whose tag name is "html"
@ -1889,7 +1882,8 @@ class TreeBuilder {
# in the next entry; i.e. act as if this was a "br" start tag token with # in the next entry; i.e. act as if this was a "br" start tag token with
# no attributes, rather than the end tag token that it actually is. # no attributes, rather than the end tag token that it actually is.
$this->error(ParseError::UNEXPECTED_END_TAG, $token->name); $this->error(ParseError::UNEXPECTED_END_TAG, $token->name);
return $this->parseTokenInHTMLContent(new StartTagToken("br"), $insertionMode); $token = new StartTagToken("br");
goto ProcessToken;
} }
# Any other end tag # Any other end tag
else { else {
@ -1925,7 +1919,8 @@ class TreeBuilder {
elseif ($token instanceof EOFToken) { elseif ($token instanceof EOFToken) {
# If the stack of template insertion modes is not empty, then process the token using the rules for the "in template" insertion mode. # If the stack of template insertion modes is not empty, then process the token using the rules for the "in template" insertion mode.
if (count($this->templateInsertionModes) !== 0) { if (count($this->templateInsertionModes) !== 0) {
return $this->parseTokenInHTMLContent($token, self::IN_TEMPLATE_MODE); $insertionMode = self::IN_TEMPLATE_MODE;
goto ProcessToken;
} }
# Otherwise, follow these steps: # Otherwise, follow these steps:
@ -2079,7 +2074,8 @@ class TreeBuilder {
# A start tag whose tag name is one of: "style", "script", "template" # A start tag whose tag name is one of: "style", "script", "template"
elseif ($token->name === "style" || $token->name === "script" || $token->name === "template") { elseif ($token->name === "style" || $token->name === "script" || $token->name === "template") {
# Process the token using the rules for the "in head" insertion mode. # Process the token using the rules for the "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# A start tag whose tag name is "input" # A start tag whose tag name is "input"
elseif ($token->name === "input") { elseif ($token->name === "input") {
@ -2158,7 +2154,8 @@ class TreeBuilder {
elseif ($token->name === "template") { elseif ($token->name === "template") {
# Process the token using the rules for the "in head" # Process the token using the rules for the "in head"
# insertion mode. # insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
// Any other end tag // Any other end tag
else { else {
@ -2169,7 +2166,8 @@ class TreeBuilder {
elseif ($token instanceof EOFToken) { elseif ($token instanceof EOFToken) {
# Process the token using the rules for the "in body" # Process the token using the rules for the "in body"
# insertion mode. # insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# Anything else # Anything else
else { else {
@ -2313,7 +2311,8 @@ class TreeBuilder {
# Anything else # Anything else
else { else {
# Process the token using the rules for the "in body" insertion mode. # Process the token using the rules for the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
} }
# 13.2.6.4.12 The "in column group" insertion mode # 13.2.6.4.12 The "in column group" insertion mode
@ -2339,7 +2338,8 @@ class TreeBuilder {
elseif ($token instanceof StartTagToken && $token->name === "html") { elseif ($token instanceof StartTagToken && $token->name === "html") {
# Process the token using the rules for the "in body" # Process the token using the rules for the "in body"
# insertion mode. # insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# A start tag whose tag name is "col" # A start tag whose tag name is "col"
elseif ($token instanceof StartTagToken && $token->name === "col") { elseif ($token instanceof StartTagToken && $token->name === "col") {
@ -2374,13 +2374,15 @@ class TreeBuilder {
elseif ($token instanceof TagToken && $token->name === "template") { elseif ($token instanceof TagToken && $token->name === "template") {
# Process the token using the rules for # Process the token using the rules for
# the "in head" insertion mode. # the "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# An end-of-file token # An end-of-file token
elseif ($token instanceof EOFToken) { elseif ($token instanceof EOFToken) {
# Process the token using the rules for # Process the token using the rules for
# the "in body" insertion mode. # the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# Anything else # Anything else
else { else {
@ -2487,7 +2489,8 @@ class TreeBuilder {
else { else {
# Process the token using the rules for # Process the token using the rules for
# the "in table" insertion mode. # the "in table" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_TABLE_MODE); $insertionMode = self::IN_TABLE_MODE;
goto ProcessToken;
} }
} }
# 13.2.6.4.14 The "in row" insertion mode # 13.2.6.4.14 The "in row" insertion mode
@ -2588,7 +2591,8 @@ class TreeBuilder {
else { else {
# Process the token using the rules for the # Process the token using the rules for the
# "in table" insertion mode. # "in table" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_TABLE_MODE); $insertionMode = self::IN_TABLE_MODE;
goto ProcessToken;
} }
} }
# 13.2.6.4.15 The "in cell" insertion mode # 13.2.6.4.15 The "in cell" insertion mode
@ -2663,7 +2667,8 @@ class TreeBuilder {
else { else {
# Process the token using the rules for # Process the token using the rules for
# the "in body" insertion mode. # the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
} }
# 13.2.6.4.16 The "in select" insertion mode # 13.2.6.4.16 The "in select" insertion mode
@ -2693,7 +2698,8 @@ class TreeBuilder {
# A start tag whose tag name is "html" # A start tag whose tag name is "html"
if ($token->name === "html") { if ($token->name === "html") {
# Process the token using the rules for the "in body" insertion mode. # Process the token using the rules for the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# A start tag whose tag name is "option" # A start tag whose tag name is "option"
elseif ($token->name === "option") { elseif ($token->name === "option") {
@ -2762,7 +2768,8 @@ class TreeBuilder {
elseif ($token->name === "script" || $token->name === "template") { elseif ($token->name === "script" || $token->name === "template") {
# Process the token using the rules for the # Process the token using the rules for the
# "in head" insertion mode. # "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
// Any other start tag // Any other start tag
else { else {
@ -2775,7 +2782,8 @@ class TreeBuilder {
# An end tag whose tag name is "template" # An end tag whose tag name is "template"
if ($token->name === "tenplate") { if ($token->name === "tenplate") {
# Process the token using the rules for the "in head" insertion mode. # Process the token using the rules for the "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# An end tag whose tag name is "optgroup" # An end tag whose tag name is "optgroup"
elseif ($token->name === "optgroup") { elseif ($token->name === "optgroup") {
@ -2834,7 +2842,8 @@ class TreeBuilder {
# An end-of-file token # An end-of-file token
elseif ($token instanceof EOFToken) { elseif ($token instanceof EOFToken) {
# Process the token using the rules for the "in body" insertion mode. # Process the token using the rules for the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# Anything else # Anything else
else { else {
@ -2884,7 +2893,8 @@ class TreeBuilder {
else { else {
# Process the token using the rules for the # Process the token using the rules for the
# "in select" insertion mode. # "in select" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_SELECT_MODE); $insertionMode = self::IN_SELECT_MODE;
goto ProcessToken;
} }
} }
# 13.2.6.4.18 The "in template" insertion mode # 13.2.6.4.18 The "in template" insertion mode
@ -2895,7 +2905,8 @@ class TreeBuilder {
if ($token instanceof CharacterToken || $token instanceof CommentToken || $token instanceof DOCTYPEToken) { if ($token instanceof CharacterToken || $token instanceof CommentToken || $token instanceof DOCTYPEToken) {
# Process the token using the rules for the # Process the token using the rules for the
# "in body" insertion mode. # "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# A start tag... # A start tag...
elseif ($token instanceof StartTagToken) { elseif ($token instanceof StartTagToken) {
@ -2905,7 +2916,8 @@ class TreeBuilder {
if (in_array($token->name, ["base", "basefont", "bgsound", "link", "meta", "noframes", "script", "style", "template", "title"])) { if (in_array($token->name, ["base", "basefont", "bgsound", "link", "meta", "noframes", "script", "style", "template", "title"])) {
# Process the token using the rules for the # Process the token using the rules for the
# "in head" insertion mode. # "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# A start tag whose tag name is one of: "caption", # A start tag whose tag name is one of: "caption",
# "colgroup", "tbody", "tfoot", "thead" # "colgroup", "tbody", "tfoot", "thead"
@ -2983,7 +2995,8 @@ class TreeBuilder {
elseif ($token instanceof EndTagToken && $token->name === "template") { elseif ($token instanceof EndTagToken && $token->name === "template") {
# Process the token using the rules for the # Process the token using the rules for the
# "in head" insertion mode. # "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# Any other end tag # Any other end tag
elseif ($token instanceof EndTagToken) { elseif ($token instanceof EndTagToken) {
@ -3024,7 +3037,8 @@ class TreeBuilder {
if ($token instanceof WhitespaceToken) { if ($token instanceof WhitespaceToken) {
# Process the token using the rules for # Process the token using the rules for
# the "in body" insertion mode. # the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# A comment token # A comment token
elseif ($token instanceof CommentToken) { elseif ($token instanceof CommentToken) {
@ -3041,7 +3055,8 @@ class TreeBuilder {
elseif ($token instanceof StartTagToken && $token->name === "html") { elseif ($token instanceof StartTagToken && $token->name === "html") {
# Process the token using the rules for # Process the token using the rules for
# the "in body" insertion mode. # the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# An end tag whose tag name is "html" # An end tag whose tag name is "html"
elseif ($token instanceof EndTagToken && $token->name === "html") { elseif ($token instanceof EndTagToken && $token->name === "html") {
@ -3104,7 +3119,8 @@ class TreeBuilder {
if ($token->name === "html") { if ($token->name === "html") {
# Process the token using the rules for # Process the token using the rules for
# the "in body" insertion mode. # the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# A start tag whose tag name is "frameset" # A start tag whose tag name is "frameset"
elseif ($token->name === "frameset") { elseif ($token->name === "frameset") {
@ -3124,7 +3140,8 @@ class TreeBuilder {
elseif ($token->name === "noframes") { elseif ($token->name === "noframes") {
# Process the token using the rules # Process the token using the rules
# for the "in head" insertion mode. # for the "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
// Any other start tag // Any other start tag
else { else {
@ -3204,7 +3221,8 @@ class TreeBuilder {
elseif ($token instanceof StartTagToken && $token->name === "html") { elseif ($token instanceof StartTagToken && $token->name === "html") {
# Process the token using the rules for # Process the token using the rules for
# the "in body" insertion mode. # the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# An end tag whose tag name is "html" # An end tag whose tag name is "html"
elseif ($token instanceof EndTagToken && $token->name === "html") { elseif ($token instanceof EndTagToken && $token->name === "html") {
@ -3215,7 +3233,8 @@ class TreeBuilder {
elseif ($token instanceof StartTagToken && $token->name === "noframes") { elseif ($token instanceof StartTagToken && $token->name === "noframes") {
# Process the token using the rules for # Process the token using the rules for
# the "in head" insertion mode. # the "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# An end-of-file token # An end-of-file token
elseif ($token instanceof EOFToken) { elseif ($token instanceof EOFToken) {
@ -3256,7 +3275,8 @@ class TreeBuilder {
elseif ($token instanceof DOCTYPEToken || $token instanceof WhitespaceToken || ($token instanceof StartTagToken && $token->name === "html")) { elseif ($token instanceof DOCTYPEToken || $token instanceof WhitespaceToken || ($token instanceof StartTagToken && $token->name === "html")) {
# Process the token using the rules for # Process the token using the rules for
# the "in body" insertion mode. # the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# An end-of-file token # An end-of-file token
elseif ($token instanceof EOFToken) { elseif ($token instanceof EOFToken) {
@ -3295,7 +3315,8 @@ class TreeBuilder {
elseif ($token instanceof DOCTYPEToken || $token instanceof WhitespaceToken || ($token instanceof StartTagToken && $token->name === "html")) { elseif ($token instanceof DOCTYPEToken || $token instanceof WhitespaceToken || ($token instanceof StartTagToken && $token->name === "html")) {
# Process the token using the rules for # Process the token using the rules for
# the "in body" insertion mode. # the "in body" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_BODY_MODE); $insertionMode = self::IN_BODY_MODE;
goto ProcessToken;
} }
# An end-of-file token # An end-of-file token
elseif ($token instanceof EOFToken) { elseif ($token instanceof EOFToken) {
@ -3306,7 +3327,8 @@ class TreeBuilder {
elseif ($token instanceof StartTagToken && $token->name === "noframes") { elseif ($token instanceof StartTagToken && $token->name === "noframes") {
# Process the token using the rules for # Process the token using the rules for
# the "in head" insertion mode. # the "in head" insertion mode.
return $this->parseTokenInHTMLContent($token, self::IN_HEAD_MODE); $insertionMode = self::IN_HEAD_MODE;
goto ProcessToken;
} }
# Anything else # Anything else
else { else {

4
tests/cases/TestTreeConstructor.php
View File

@ -73,9 +73,7 @@ class TestTreeConstructor extends \PHPUnit\Framework\TestCase {
$treeBuilder = new TreeBuilder($doc, $decoder, $tokenizer, $tokenList, $errorHandler, $stack, new TemplateInsertionModesStack, $fragmentContext); $treeBuilder = new TreeBuilder($doc, $decoder, $tokenizer, $tokenList, $errorHandler, $stack, new TemplateInsertionModesStack, $fragmentContext);
// run the tree builder // run the tree builder
try { try {
foreach($tokenList as $token) { $treeBuilder->constructTree();
$treeBuilder->emitToken($token);
}
} catch (\DOMException $e) { } catch (\DOMException $e) {
$this->markTestIncomplete('Requires implementation of the "Coercing an HTML DOM into an infoset" specification section'); $this->markTestIncomplete('Requires implementation of the "Coercing an HTML DOM into an infoset" specification section');
return; return;

Write Preview
Loading…
Cancel
Save