html/scripting-parser/parser.html

<!DOCTYPE HTML>
<html lang=en>
<head>
<title>Live Scripting HTML Parser</title>
<link rel=author href="http://suika.fam.cx/~wakaba/who?">
<link rel=license href="http://suika.fam.cx/c/gnu/gpl"
    title="GNU GPL2 or later">
<style>
  h1 {
    margin: 0;
    font-size: 150%;
  } 
  h2 {
    margin: 0;
    font-size: 100%;
  }
  p {
    margin: 0 1em;
  }
  textarea {
    width: 100%;
    -width: 99%;
    height: 10em;
  }
  output {
    display: block;
    font-family: monospace;
    white-space: -moz-pre-wrap;
    white-space: pre-wrap;
  }
</style>
<script>
  var delayedUpdater = 0;

  function update () {
    if (delayedUpdater) {
      clearTimeout (delayedUpdater);
      delayedUpdater = 0;
    }
    delayedUpdater = setTimeout (update2, 100);
  } // update

  function update2 () {
    var v = document.sourceElement.value;
    if (v != document.previousSourceText) {
      document.previousSourceText = v;
      document.links['permalink'].href
          = location.pathname + '?s=' + encodeURIComponent (v);
      document.links['ldvlink'].href
          = 'http://software.hixie.ch/utilities/js/live-dom-viewer/?'
          + encodeURIComponent (v);

      document.logElement.textContent = '';
      var p = new Parser (new InputStream (v));
      var doc = p.doc;
      p.parse ();
      
      log (dumpTree (doc, ''));
      
      if (p.hasAsyncScript) {
        log ('Some script codes are executed asynchronously; it means that the document might be rendered in different ways depending on the network condition and other factors');
      }
    }
  } // update2

  var logIndentLevel = 0;
  function log (s) {
    var indent = '';
    for (var i = 0; i < logIndentLevel; i++) {
      indent += '  ';
    }
    s = indent + s.replace (/\n/g, "\n" + indent);
    document.logElement.appendChild (document.createTextNode (s + "\n"));
  } // log

  function InputStream (s) {
    this.s = s;
  } // InputStream

  function Parser (i, doc) {
    this.parseMode = 'pcdata';
    if (!doc) {
      doc = new JSDocument (this);
      doc.manakaiIsHTML = true;
    }
    this.nextToken = [];
    this.doc = doc;
    this.openElements = [doc];
    this.input = i;
    this.scriptsExecutedAfterParsing = [];
    this.scriptsExecutedSoon = [];
    this.scriptsExecutedAsynchronously = [];
  } // Parser

  Parser.prototype.getNextToken = function () {
    if (this.nextToken.length) {
      return this.nextToken.shift ();
    }

    var p = this;
    var i = this.input;
    if (this.parseMode == 'cdata') {
      var tagName = this.endTagName;
      var token;
      if (p.insertionPoint <= 0) {
        return {type: 'abort'};
      }
      i.s = i.s.replace (/^([^<]+)/,
      function (s, t) {
        if (0 < p.insertionPoint && p.insertionPoint < t.length) {
          token = {type: 'char', value: t.substring (0, p.insertionPoint)};
          var ip = p.insertionPoint;
          p.insertionPoint = 0;
          return t.substring (ip, t.length);
        }
        token = {type: 'char', value: t};
        p.insertionPoint -= t.length;
        return '';
      });
      if (token) return token;
      var pattern = new RegExp ('^</' + tagName + '>', 'i');
      i.s = i.s.replace (pattern, function (s) {
        if (p.insertionPoint < s.length) {
          token = {type: 'abort'};
          return s;
        }
        token = {type: 'end-tag', value: tagName};
        p.insertionPoint -= s.length;
        return '';
      });
      if (token) return token;
      var m;
      if ((p.insertionPoint < ('</' + tagName).length) &&
          (m = i.s.match (/^<\/([A-Za-z]+)/))) {
        var v = m[1].substring (0, p.insertionPoint).toLowerCase ();
        if (v == tagName.substring (0, p.insertionPoint - '</'.length)) {
          return {type: 'abort'};
        }
      }
      i.s = i.s.replace (/^</,
      function (s) {
        token = {type: 'char', value: s};
        p.insertionPoint -= s.length;
        return '';
      });
      if (token) return token;
      return {type: 'eof'};
    }

    var token;
    i.s = i.s.replace (/^<\/([^>]+)(?:>|$)/, function (s, e) {
      if (p.insertionPoint < s.length ||
          (p.insertionPoint <= s.length &&
           s.substring (s.length - 1, s.length) != '>')) {
        token = {type: 'abort'};
        return s;
      }
      token = {type: 'end-tag', value: e.toLowerCase ()};
      p.insertionPoint -= s.length;
      return '';
    });
    if (token) return token;
    i.s = i.s.replace (/^<([^>]+)(?:>|$)/, function (s, e) {
      if (p.insertionPoint < s.length ||
          (p.insertionPoint <= s.length &&
           s.substring (s.length - 1, s.length) != '>')) {
        token = {type: 'abort'};
        return s;
      }
      var tagName;
      var attrs = {};
      e = e.replace (/^[\S]+/, function (v) {
        tagName = v.toLowerCase ();
        return '';
      });
      while (true) {
        var m = false;
        e = e.replace (/^\s*([^\s=]+)\s*(?:=\s*(?:"([^"]*)"|'([^']*)'|([^"'\s]*)))?/,
        function (x, attrName, attrValue1, attrValue2, attrValue3) {
          v = attrValue1 || attrValue2 || attrValue3;
          v = v.replace (/&quot;/g, '"').replace (/&apos;/g, "'")
              .replace (/&amp;/g, '&');
          attrs[attrName.toLowerCase ()] = v;
          m = true;
          return '';
        });
        if (!m) break;
      }
      if (e.length) {
        log ('Broken start tag: "' + e + '"');
      }
      token = {type: 'start-tag', value: tagName, attrs: attrs};
      p.insertionPoint -= s.length;
      return '';
    });
    if (token) return token;
    if (p.insertionPoint <= 0) {
      return {type: 'abort'};
    }
    i.s = i.s.replace (/^[^<]+/, function (s) {
      if (p.insertionPoint < s.length) {
        token = {type: 'char', value: s.substring (0, p.insertionPoint)};
        var ip = p.insertionPoint;
        p.insertionPoint = 0;
        return s.substring (ip, s.length);
      }
      token = {type: 'char', value: s};
      p.insertionPoint -= s.length;
      return '';
    });
    if (token) return token;
    i.s = i.s.replace (/^[\s\S]/, function (s) {
      token = {type: 'char', value: s};
      p.insertionPoint -= s.length;
      return '';
    });
    if (token) return token;
    return {type: 'eof'};
  } // getNextToken

  Parser.prototype.parse = function () {
    logIndentLevel++;
    log ('parse: start');

    while (true) {
      var token = this.getNextToken ();
      log ('token: ' + token.type + ' "' + token.value + '"');

      if (this.cdataEndTagRequired) {
        // Generic CDATA parsing algorithm

        if (token.type != 'abort') {
          // 7.
          if (token.type == 'end-tag' && token.value == this.endTagName) {
            // 7.1. Ignores it.
            //
          } else {
            // 7.2. Parse error.
            log ('Parse error: no </' + this.endTagName + '>');
            this.nextToken.unshift (token);
          }
          this.cdataEndTagRequired = false;
          continue;
        }
      }

      if (token.type == 'start-tag') {
        if (token.value == 'script') {
          // 1. Create an element for the token in the HTML namespace.
          var el = new JSElement (this.doc, token.value);
          if (token.attrs.async != null) el.async = true;
          if (token.attrs.defer != null) el.defer = true;
          if (token.attrs.src != null) el.src = token.attrs.src;

          // 2. Mark the element as being "parser-inserted".
          el.manakaiParserInserted = true;

          // 3. Switch the tokeniser's content model flag to the CDATA state.
          this.parseMode = 'cdata';
          this.endTagName = 'script';

          // 4.1. Collect all the character tokens.
          while (true) {
            var token = this.getNextToken ();
            log ('token: ' + token.type + ' "' + token.value + '"');

            if (token.type == 'char') {
              // 5. Append a single Text node to the script element node.
              el.manakaiAppendText (token.value);

            // 4.2. Until it returns a token that is not a character token, or
            // until it stops tokenising.
            } else if (token.type == 'eof' ||
                       token.type == 'end-tag' ||
                       token.type == 'abort') {
              // 6. Switched back to the PCDATA state.
              this.parseMode = 'pcdata';

              // 7.1. If the next token is not an end tag token with ...
              if (!(token.type == 'end-tag' && token.value == 'script')) {
                // 7.2. This is a parse error.
                log ('Parse error: no </' + 'script>');
                this.nextToken.unshift (token);

                // 7.3. Mark the script element as "already executed".
                el.manakaiAlreadyExecuted = true;
              } else {
                // 7.4. Ignore it.
                //
              }
              break;
            }
          }

          // 8.1. If the parser were originally created for the ...
          if (this.fragmentParsingMode) {
            // 8.2. Mark the script element as "already executed" and ...
            el.manakaiAlreadyExecuted = true;
            continue;
          }

          // 9.1. Let the old insertion point have the same value as the ...
          var oldInsertionPoint = this.insertionPoint;
          // 9.2. Let the insertion point be just before the next input ...
          this.setInsertionPoint (0);

          // 10. Append the new element to the current node.
          this.openElements[this.openElements.length - 1].appendChild (el);

          // 11. Let the insertion point have the value of the old ...

          oldInsertionPoint += this.insertionPoint;
          this.setInsertionPoint (oldInsertionPoint);

          // 12. If there is a pending external script
          while (this.pendingExternalScript) {
            // 12.1. If the tree construction stage is being called reentrantly
            if (this.reentrant) {
              log ('parse: abort (reentrance)');
              logIndentLevel--;
              return;

            // 12.2. Otherwise
            } else {
              // 1.
              var script = this.pendingExternalScript;
              this.pendingExternalScript = null;

              // 2. Pause until the script has completed loading.
              // 

              // 3. Let the insertion point to just before the next input char.
              this.setInsertionPoint (0);

              // 4. Execute the script.
              executeScript (this.doc, script);

              // 5. Let the insertion point be undefined again.
              this.setInsertionPoint (undefined);

              // 6. If there is once again a script that will execute ...
              //
            }
          }
        } else if (token.value == 'style' ||
                   token.value == 'noscript' ||
                   token.value == 'xmp') {
          // 1. Create an element for the token in the HTML namespace.
          var el = new JSElement (this.doc, token.value);

          // 2. Append the new element to the current node.
          this.openElements[this.openElements.length - 1].appendChild (el);

          // 3. Switch the tokeniser's content model flag to the CDATA state.
          this.parseMode = 'cdata';
          this.endTagName = token.value;

          // 4.1. Collect all the character tokens.
          while (true) {
            var token = this.getNextToken ();
            log ('token: ' + token.type + ' "' + token.value + '"');

            if (token.type == 'char') {
              // 5. Append a single Text node to the script element node.
              el.manakaiAppendText (token.value);

            // 4.2. Until it returns a token that is not a character token, or
            // until it stops tokenising.
            } else if (token.type == 'eof' ||
                       token.type == 'end-tag' ||
                       token.type == 'abort') {
              // 6. Switched back to the PCDATA state.
              this.parseMode = 'pcdata';

              if (token.type == 'abort') {
                this.cdataEndTagRequired = true;
                break;
              }

              // 7.1. If the next token is not an end tag token with ...
              if (!(token.type == 'end-tag' &&
                    token.value == this.endTagName)) {
                // 7.2. This is a parse error.
                log ('Parse error: no </' + this.endTagName + '>');
                this.nextToken.unshift (token);

                // 7.3. Mark the script element as "already executed".
                el.manakaiAlreadyExecuted = true;
              } else {
                // 7.4. Ignore it.
                //
              }
              break;
            }
          }
        } else {
          var el = new JSElement (this.doc, token.value);
          this.openElements[this.openElements.length - 1].appendChild (el);
          this.openElements.push (el);
        }
      } else if (token.type == 'end-tag') {
        if (this.openElements[this.openElements.length - 1].localName ==
            token.value) {
          this.openElements.pop ();
        } else {
          log ('parse error: unmatched end tag: ' + token.value);
        }
      } else if (token.type == 'char') {
        this.openElements[this.openElements.length - 1].manakaiAppendText
            (token.value);
      } else if (token.type == 'eof') {
        break;
      } else if (token.type == 'abort') {
        log ('parse: abort');
        logIndentLevel--;
        return;
      }
    }

    log ('stop parsing');

    // readyState = 'interactive'

    // "When a script completes loading" rules start applying.

    while (this.scriptsExecutedSoon.length > 0 ||
           this.scriptsExecutedAsynchronously.length > 0) {
      // Handle "list of scripts that will execute as soon as possible".
      while (this.scriptsExecutedSoon.length > 0) {
        var e = this.scriptsExecutedSoon.shift ();
  
        // If it has completed loading
        log ('Execute an external script not inserted by parser...');
        executeScript (this.doc, e);

        // NOTE: It MAY be executed before the end of the parsing, according
        // to the spec.
        this.hasAsyncScript = true;
      }

      // Handle "list of scripts that will execute asynchronously".
      while (this.scriptsExecutedAsynchronously.length > 0) {
        var e = this.scriptsExecutedAsynchronously.shift ();

        // Step 1.
        // We assume that all scripts have been loaded at this time.
  
        // Step 2.
        log ('Execute an asynchronous script...');
        executeScript (this.doc, e);

        // Step 3.
        //
 
        // Step 4.
        //

        this.hasAsyncScript = true;
      }
    }

    // Handle "list of scripts that will execute when the document has finished
    // parsing".
    var list = this.scriptsExecutedAfterParsing;
    while (list.length > 0) {
      // TODO: break unless completed loading

      // Step 1.
      //

      // Step 2. and Step 3.
      log ('Executing a |defer|red script...');
      executeScript (this.doc, list.shift ());

      // Step 4.
    }

    log ('DOMContentLoaded event fired');

    // "delays the load event" things has completed:
    // readyState = 'complete'
    log ('load event fired');

    logIndentLevel--;
  } // parse

  Parser.prototype.setInsertionPoint = function (ip) {
    if (ip == undefined || ip == null || isNaN (ip)) {
      log ('insertion point: set to undefined');
      this.insertionPoint = undefined;
    } else if (ip == this.input.s.length) {
      log ('insertion point: end of file');
      this.insertionPoint = ip;
    } else {
      log ('insertion point: set to ' + ip +
           ' (before "' + this.input.s.substring (0, 10) + '")');
      this.insertionPoint = ip;
    }
  }; // setInsertionPoint

  function JSDocument (p) {
    this.childNodes = [];
    this._parser = p;
  } // JSDocument

  function JSElement (doc, localName) {
    this.localName = localName;
    this.ownerDocument = doc;
    this.childNodes = [];
  } // JSElement

  JSDocument.prototype.appendChild = JSElement.prototype.appendChild =
  function (e) {
    this.childNodes.push (e);
    e.parentNode = this;

    if (e.localName == 'script') {
      logIndentLevel++;
      log ('Running a script: start');

      var doc = this.ownerDocument || this;
      var p = doc._parser;

      // 1.The script's type
      //

      // 2. The cript's character encoding
      //

      // 3.1. If without script
      //
      // 2.2. If the script element was created by an XML ... innerHTML ...
      //
      // 2.3. If the user agent does not support the scripting language ...
      //
      if (false) {
        // 2.5. Abort these steps at this point.
        log ('Running a script: aborted (noscript)');
        logIndentLevel--;
        return e;
      }

      // 4. Set the element's "already executed" flag.
      e.manakaiAlreadyExecuted = true;

      // 5. If the element has a src attribute, then a load for ...
      // TODO: load an external resource

      // 5. The first of the following options:

      if (/* TODO: If the document is still being parsed && */
          e.defer && !e.async) {
        // 6.1. 
        p.scriptsExecutedAfterParsing.push (e);
        log ('Running a script: aborted (defer)');
      } else if (e.async && e.src != null) {
        // 6.2.
        p.scriptsExecutedAsynchronously.push (e);
        log ('Running a script: aborted (async src)');
      } else if (e.async && e.src == null &&
                 p.scriptsExecutedAsynchronously.length > 0) {
        // 6.3.
        p.scriptsExecutedAsynchronously.push (e);
        log ('Running a script: aborted (async)');
      } else if (e.src != null && e.manakaiParserInserted) {
        // 6.4.
        if (p.pendingExternalScript) {
          log ('Error: There is a pending external script.');
        }
        p.pendingExternalScript = e;
        log ('Running a script: aborted (src parser-inserted)');
      } else if (e.src != null) {
        // 6.5.
        p.scriptsExecutedSoon.push (e);
        log ('Running a script: aborted (src)');
      } else {
        // 6.6.
        executeScript (doc, e); // even if other scripts are already executing.
      }

      log ('Running a script: end');
      logIndentLevel--;
    }

    return e;
  }; // appendChild

  function executeScript (doc, e) {
    log ('executing a script block: start');

    var s;
    if (e.src != null) {
      s = getExternalScript (e.src);

      // If the load resulted in an error, then ... firing an error event ...
      if (s == null) {
        log ('error event fired at the script element');
        return;
      }

      log ('External script loaded: "' + s + '"');
    } else {
      s = e.text;
    }

    // If the load was successful

    if (true) {
    // Scripting is enabled, Document.designMode is disabled,
    // Document is the active document in its browsing context

      parseAndRunScript (doc, s);
    }

    log ('load event fired at the script element');

    log ('executing a script block: end');
  } // executeScript

  function getExternalScript (uri) {
    if (uri.match (/^javascript:/i)) {
      var m;
      if (m = uri.match (/^javascript:\s*(?:'([^']*)'|"([^"]+)")\s*$/i)) {
        if (m[1]) {
          return unescapeJSLiteral (m[1]);
        } else if (m[2]) {
          return unescapeJSLiteral (m[2]);
        } else {
          return null;
        }
      } else {
        log ('Complex javascript: URI is not supported: <' + uri + '>');
        return null;
      }
    } else {
      log ('URI scheme not supported: <' + uri + '>');
      return null;
    }
  } // getExternalScript

  function parseAndRunScript (doc, s) {
    while (true) {
      var matched = false;
      s = s.replace (/^\s*document\.write\s*\(((?:'[^']*'|"[^"]*")\s*(?:,\s*(?:'[^']*'|"[^"]*"))*)\)\s*;\s*/, function (s, t) {
        matched = true;
        var args = [];
        t.replace (/('[^']*'|"[^"]*")/g, function (s, v) {
          args.push (unescapeJSLiteral (v.substring (1, v.length - 1)));
          return '';
        });
        doc.write.apply (doc, args);
        return '';
      });
      var noDocumentElement = false;
      s = s.replace (/^\s*var\s+s\s*=\s*document\.createElement\s*\(\s*['"]script['"]\s*\)\s*;\s*s\.src\s*=\s*(?:'([^']*)'|"([^"]*)")\s*;\s*document\.documentElement\.appendChild\s*\(\s*s\s*\)\s*;\s*/,
      function (s, t, u) {
        matched = true;
        var args = [unescapeJSLiteral (t ? t : u)];
        noDocumentElement = !doc._insertExternalScript.apply (doc, args);
        return '';
      });
      if (noDocumentElement) {
        log ('Script error: documentElement is null');
        break;
      }
      s = s.replace (/^\s*w\s*\(\s*document\.documentElement\.innerHTML\s*\)\s*;\s*/,
      function (s, t) {
        matched = true;
        log (dumpTree (doc, ''));
        return '';
      });
      if (s == '') break;
      if (!matched) {
        log ('Script parse error: "' + s + '"');
        break;
      }
    }
  } // parseAndRunScript

  function unescapeJSLiteral (s) {
    return s.replace (/\\u([0-9A-Fa-f]{4})/g, function (t, v) {
      return String.fromCharCode (parseInt ('0x' + v));
    });
  } // unescapeJSLiteral

  function JSText (data) {
    this.data = data;
  } // JSText

  JSDocument.prototype.manakaiAppendText =
  JSElement.prototype.manakaiAppendText =
  function (s) {
    if (this.childNodes.length > 0 &&
        this.childNodes[this.childNodes.length - 1] instanceof JSText) {
      this.childNodes[this.childNodes.length - 1].data += s;
    } else {
      this.childNodes.push (new JSText (s));
    }
  }; // manakaiAppendText

  JSDocument.prototype.open = function () {
    // Two or fewer arguments

    // Step 1.
    var type = arguments[0] || 'text/html';
    
    // Step 2.
    var replace = arguments[1] == 'replace';

    // Step 3.
    if (this._parser &&
        !this._parser.scriptCreated &&
        this._parser.input.insertionPoint != undefined) {
      log ('document.open () in parsing mode is ignored');
      return this;
    }

    // Step 4.
    log ('onbeforeunload event fired');
    log ('onunload event fired');

    // Step 5.
    if (this._parser) {
      // Discard the parser.
    }

    // Step 6.
    log ('document cleared by document.open ()');
    this.childNodes = [];

    // Step 7.
    this._parser = new Parser (new InputStream (''), this);
    this._parser.scriptCreated = true;

    // Step 8.
    this.manakaiIsHTML = true;

    // Step 9.
    // If not text/html, ...

    // Step 10.
    if (!replace) {
      // History      
    }

    // Step 11.
    this._parser.setInsertionPoint (this._parser.input.s.length);

    // Step 12.
    return this;
  }; // document.open

  JSDocument.prototype.write = function () {
    log ('document.write: start');
    logIndentLevel++;

    var p = this._parser;

    // 1. If the insertion point is undefined, the open() method must be ...
    if (isNaN (p.insertionPoint) || p.insertionPoint == undefined) {
      this.open ();
      p = this._parser;
    }

    // 2. ... inserted into the input stream just before the insertion point.
    var s = Array.join (arguments, '');
    log ('document.write: insert "' + s + '"' +
         ' before "' +
         p.input.s.substring (p.insertionPoint, p.insertionPoint + 10) + '"');
    p.input.s = p.input.s.substring (0, p.insertionPoint) + s
        + p.input.s.substring (p.insertionPoint, p.input.s.length);
    p.insertionPoint += s.length;

    // 3. If there is a pending external script
    if (p.pendingExternalScript) {
      log ('document.write: processed later (there is an unprocessed <script src>)');
      logIndentLevel--;
      log ('document.write: return');
      return;
    }

    // 4. Process the characters that were inserted, ...
    var originalReentrant = p.reentrant;
    p.reentrant = true;
    p.parse ();
    p.reentrant = originalReentrant;
    // TODO: "Abort the processing of any nested invokations of the tokeniser,
    // yielding control back to the caller." (<script> parsing).  Do we need
    // to do something here?

    // 5. Return
    logIndentLevel--;
    log ('document.write: return');

    return;
  }; // document.write

  JSDocument.prototype._insertExternalScript = function (uri) {
    var s = new JSElement (this, 'script');
    s.src = uri;
    if (this.documentElement) {
      this.documentElement.appendChild (s);
      return true;
    } else {
      return false;
    }
  }; // _insertExternalScript

  JSDocument.prototype.__defineGetter__ ('documentElement', function () {
    var cn = this.childNodes;
    for (var i = 0; i < cn.length; i++) {
      if (cn[i] instanceof JSElement) {
        return cn[i]
      }
    }
    return null;
  });

  JSElement.prototype.__defineGetter__ ('text', function () {
    var r = '';
    for (var i = 0; i < this.childNodes.length; i++) {
      if (this.childNodes[i] instanceof JSText) {
        r += this.childNodes[i].data;
      }
    }
    return r;
  });

  function dumpTree (n, indent) {
    var r = '';
    for (var i = 0; i < n.childNodes.length; i++) {
      var node = n.childNodes[i];
      if (node instanceof JSElement) {
        r += '| ' + indent + node.localName + '\n';
        if (node.async) r += '| ' + indent + '  async=""\n';
        if (node.defer) r += '| ' + indent + '  defer=""\n';
        if (node.src != null) {
          r += '| ' + indent + '  src="' + node.src + '"\n';
        }
        r += dumpTree (node, indent + '  ');
      } else if (node instanceof JSText) {
        r += '| ' + indent + '"' + node.data + '"\n';
      } else {
        r += '| ' + indent + node + '\n';
      }
    }
    return r;
  } // dumpTree
</script>
</head>
<body onload="
  document.sourceElement = document.getElementsByTagName ('textarea')[0];

  var q = location.search;
  if (q != null) {
    q = q.substring (1).split (/;/);
    for (var i = 0; i < q.length; i++) {
      var v = q[i].split (/=/, 2);
      v[0] = decodeURIComponent (v[0]);
      v[1] = decodeURIComponent (v[1] || '');
      if (v[0] == 's') {
        document.sourceElement.value = v[1];
      }
    }
  }

  document.logElement = document.getElementsByTagName ('output')[0];
  update ();
">
<h1>Live Scripting <abbr title="Hypertext Markup Language">HTML</abbr>
Parser</h1>

<h2>Markup to test
(<a href=data:, id=permalink rel=bookmark>permalink</a>,
<a href="http://software.hixie.ch/utilities/js/live-dom-viewer/"
    id=ldvlink>Live <abbr title="Document Object Model">DOM</abbr>
    Viewer</a>)</h2>
<p>
<textarea onkeydown=" update () " onchange=" update () " oninput=" update () ">&lt;html>
&lt;head>&lt;/head>&lt;body>
&lt;p>
&lt;script>
document.write ('aaaaaaa&lt;/p>&lt;script>document.write("cccccc");&lt;/', 'script>bbbbbb');
&lt;/script>
&lt;p>
</textarea>

<h2 id=log>Log</h2>
<p><output></output>

<h2 id=notes>Notes</h2>

<p>This is a <em>simplified</em> implementation of 
<a href="http://www.whatwg.org/specs/web-apps/current-work/#parsing">HTML5 
Parsing Algorithm</a>.  It only implements script-related part of the 
algorithm.  Especially, this parser:
<ul>
<li>Does not support <code>DOCTYPE</code> and comment tokens.
<li>Does not support entities except for <code>&amp;quot;</code>,
<code>&amp;apos;</code>, and <code>&amp;amp;</code> in <code>script</code>
<code>src</code> attribute value.
<li>Does not support omissions of start or end tags, the <abbr>AAA</abbr>
algorithm, and so on.
<li>Does not raise parse errors for invalid attribute specifications in start 
or end tags.
<li>Does not support RCDATA elements (<code>title</code> and
<code>textarea</code>).
<li>Does not strip the first newline in <code>pre</code>,
<code>listing</code>, and <code>textarea</code> elements.
<li>Does not support <code>&lt;!--</code>..<code>--></code> parsing rule
in CDATA/RCDATA elements.
<li>Does not support foreign (SVG or MathML) elements.
<li>Only supports <code>script</code> <code>type</code>
<code>text/javascript</code>.  <code>type</code> and <code>language</code>
attributes are ignored.
<li>Only supports limited statements.  It must consist of zero or more
of statements looking similar to the following statements, possibly
introduced, followed, or separated by white space characters:
  <ul>
  <li><code>document.write ("<var>string</var>", ["<var>string</var>", ...]);</code>.
  <li><code>var s = document.createElement ("script");
            s.src = "<var>string</var>";
            document.documentElement.appendChild (s);</code>
  <li><code>w (document.documentElement.innerHTML);</code> (This statement
  can be used to dump the document, even when the document has no
  document element.  The output format is the tree dump format used
  in html5lib test data, not <abbr>HTML</abbr>.)
  </ul>
Note that strings may be delimited by <code>'</code>s instead of
<code>"</code>s.
<li>Only supports <code>javascript:</code> 
<abbr title="Uniform Resourace Identifiers">URI</abbr> scheme in the 
<code>src</code> attribute of the <code>script</code> element.  In addition, 
the <abbr title="Uniform Resource Identifiers">URI</abbr> must be conform to 
the regular expression <code>^javascript:\s*(?:"[^"]*"|'[^']*')\s*$</code>.
<li>Only supports <code>\u<var>HHHH</var></code> escapes in JavaScript 
string literals.
<li>Does not handle <i>stop parsing</i> phase correctly if the document is 
replaced by <code>document.open ()</code> call.  In other word, delayed 
(deferred or asynchronous) script executions and event firings might be
treated in a wrong way if a <code>document.open ()</code> invocation
is implicitly done by <code>document.write ()</code> in a delayed script.
</ul>

<p>For some reason, this parser does not work in browsers that do
not support JavaScript 1.5.

<!-- TODO: |src| attribute value should refer the value at the time
when it is inserted into the document, not the value when the script is 
executed.  Currently it does not matter, since we don't allow dynamic
modification to the |src| content/DOM attribute value yet. -->

<p>See also
<a href="http://suika.fam.cx/gate/2005/sw/Live%20Scripting%20HTML%20Parser">SuikaWiki:
Live Scripting HTML Parser</a>.

</body>
</html>
<!-- $Date: 2008/08/31 09:50:49 $ -->
<!--

Copyright 2008 Wakaba <w@suika.fam.cx>

This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 2
of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.

-->