/[suikacvs]/markup/html/scripting-parser/parser.html
Suika

Diff of /markup/html/scripting-parser/parser.html

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 1.14 by wakaba, Tue Apr 29 02:50:00 2008 UTC revision 1.17 by wakaba, Fri May 16 10:29:25 2008 UTC
# Line 65  Line 65 
65    
66    var logIndentLevel = 0;    var logIndentLevel = 0;
67    function log (s) {    function log (s) {
68        var indent = '';
69      for (var i = 0; i < logIndentLevel; i++) {      for (var i = 0; i < logIndentLevel; i++) {
70        s = '  ' + s;        indent += '  ';
71      }      }
72        s = indent + s.replace (/\n/g, "\n" + indent);
73      document.logElement.appendChild (document.createTextNode (s + "\n"));      document.logElement.appendChild (document.createTextNode (s + "\n"));
74    } // log    } // log
75    
# Line 81  Line 83 
83        doc = new JSDocument (this);        doc = new JSDocument (this);
84        doc.manakaiIsHTML = true;        doc.manakaiIsHTML = true;
85      }      }
86        this.nextToken = [];
87      this.doc = doc;      this.doc = doc;
88      this.openElements = [doc];      this.openElements = [doc];
89      this.input = i;      this.input = i;
# Line 90  Line 93 
93    } // Parser    } // Parser
94    
95    Parser.prototype.getNextToken = function () {    Parser.prototype.getNextToken = function () {
96        if (this.nextToken.length) {
97          return this.nextToken.shift ();
98        }
99    
100      var p = this;      var p = this;
101      var i = this.input;      var i = this.input;
102      if (this.parseMode == 'cdata') {      if (this.parseMode == 'cdata') {
# Line 144  Line 151 
151      i.s = i.s.replace (/^<\/([^>]+)(?:>|$)/, function (s, e) {      i.s = i.s.replace (/^<\/([^>]+)(?:>|$)/, function (s, e) {
152        if (p.insertionPoint < s.length ||        if (p.insertionPoint < s.length ||
153            (p.insertionPoint <= s.length &&            (p.insertionPoint <= s.length &&
154             s.substring (s.length - 1, 1) != '>')) {             s.substring (s.length - 1, s.length) != '>')) {
155          token = {type: 'abort'};          token = {type: 'abort'};
156          return s;          return s;
157        }        }
# Line 156  Line 163 
163      i.s = i.s.replace (/^<([^>]+)(?:>|$)/, function (s, e) {      i.s = i.s.replace (/^<([^>]+)(?:>|$)/, function (s, e) {
164        if (p.insertionPoint < s.length ||        if (p.insertionPoint < s.length ||
165            (p.insertionPoint <= s.length &&            (p.insertionPoint <= s.length &&
166             s.substring (s.length - 1, 1) != '>')) {             s.substring (s.length - 1, s.length) != '>')) {
167          token = {type: 'abort'};          token = {type: 'abort'};
168          return s;          return s;
169        }        }
# Line 219  Line 226 
226        var token = this.getNextToken ();        var token = this.getNextToken ();
227        log ('token: ' + token.type + ' "' + token.value + '"');        log ('token: ' + token.type + ' "' + token.value + '"');
228    
229          if (this.cdataEndTagRequired) {
230            // Generic CDATA parsing algorithm
231    
232            if (token.type != 'abort') {
233              // 7.
234              if (token.type == 'end-tag' && token.value == this.endTagName) {
235                // 7.1. Ignores it.
236                //
237              } else {
238                // 7.2. Parse error.
239                log ('Parse error: no </' + this.endTagName + '>');
240                this.nextToken.unshift (token);
241              }
242              this.cdataEndTagRequired = false;
243              continue;
244            }
245          }
246    
247        if (token.type == 'start-tag') {        if (token.type == 'start-tag') {
248          if (token.value == 'script') {          if (token.value == 'script') {
249            // 1. Create an element for the token in the HTML namespace.            // 1. Create an element for the token in the HTML namespace.
# Line 255  Line 280 
280                if (!(token.type == 'end-tag' && token.value == 'script')) {                if (!(token.type == 'end-tag' && token.value == 'script')) {
281                  // 7.2. This is a parse error.                  // 7.2. This is a parse error.
282                  log ('Parse error: no </' + 'script>');                  log ('Parse error: no </' + 'script>');
283                    this.nextToken.unshift (token);
284    
285                  // 7.3. Mark the script element as "already executed".                  // 7.3. Mark the script element as "already executed".
286                  el.manakaiAlreadyExecuted = true;                  el.manakaiAlreadyExecuted = true;
# Line 346  Line 372 
372                // 6. Switched back to the PCDATA state.                // 6. Switched back to the PCDATA state.
373                this.parseMode = 'pcdata';                this.parseMode = 'pcdata';
374    
375                  if (token.type == 'abort') {
376                    this.cdataEndTagRequired = true;
377                    break;
378                  }
379    
380                // 7.1. If the next token is not an end tag token with ...                // 7.1. If the next token is not an end tag token with ...
381                if (!(token.type == 'end-tag' &&                if (!(token.type == 'end-tag' &&
382                      token.value == this.endTagName)) {                      token.value == this.endTagName)) {
383                  // 7.2. This is a parse error.                  // 7.2. This is a parse error.
384                  log ('Parse error: no </' + this.endTagName + '>');                  log ('Parse error: no </' + this.endTagName + '>');
385                    this.nextToken.unshift (token);
386    
387                  // 7.3. Mark the script element as "already executed".                  // 7.3. Mark the script element as "already executed".
388                  el.manakaiAlreadyExecuted = true;                  el.manakaiAlreadyExecuted = true;
# Line 501  Line 533 
533        // 2.4. If the script element has its "already executed" flag set        // 2.4. If the script element has its "already executed" flag set
534        if (e.manakaiAlreadyExecuted) {        if (e.manakaiAlreadyExecuted) {
535          // 2.5. Abort these steps at this point.          // 2.5. Abort these steps at this point.
536          log ('Running a script: aborted');          log ('Running a script: aborted (already executed)');
537          logIndentLevel--;          logIndentLevel--;
538          return e;          return e;
539        }        }
# Line 566  Line 598 
598      }      }
599    
600      // If the load was successful      // If the load was successful
     log ('load event fired at the script element');  
601    
602      if (true) {      if (true) {
603      // Scripting is enabled, Document.designMode is disabled,      // Scripting is enabled, Document.designMode is disabled,
# Line 575  Line 606 
606        parseAndRunScript (doc, s);        parseAndRunScript (doc, s);
607      }      }
608    
609        log ('load event fired at the script element');
610    
611      log ('executing a script block: end');      log ('executing a script block: end');
612    } // executeScript    } // executeScript
613    
# Line 612  Line 645 
645          doc.write.apply (doc, args);          doc.write.apply (doc, args);
646          return '';          return '';
647        });        });
648        s = s.replace (/^\s*var\s+s\s*=\s*document\.createElement\s*\(\s*['"]script['"]\s*\)\s*;\s*s\.src\s*=\s*(?:'(javascript:[^']*)'|"(javascript:[^"]*)")\s*;\s*document\.documentElement\.appendChild\s*\(\s*s\s*\)\s*;\s*/,        var noDocumentElement = false;
649          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*/,
650        function (s, t, u) {        function (s, t, u) {
651          matched = true;          matched = true;
652          var args = [unescapeJSLiteral (t ? t : u)];          var args = [unescapeJSLiteral (t ? t : u)];
653          doc._insertExternalScript.apply (doc, args);          noDocumentElement = !doc._insertExternalScript.apply (doc, args);
654            return '';
655          });
656          if (noDocumentElement) {
657            log ('Script error: documentElement is null');
658            break;
659          }
660          s = s.replace (/^\s*w\s*\(\s*document\.documentElement\.innerHTML\s*\)\s*;\s*/,
661          function (s, t) {
662            matched = true;
663            log (dumpTree (doc, ''));
664          return '';          return '';
665        });        });
666        if (s == '') break;        if (s == '') break;
# Line 701  Line 745 
745    }; // document.open    }; // document.open
746    
747    JSDocument.prototype.write = function () {    JSDocument.prototype.write = function () {
748        log ('document.write: start');
749      logIndentLevel++;      logIndentLevel++;
750    
751      var p = this._parser;      var p = this._parser;
# Line 724  Line 769 
769      if (p.scriptExecutedAfterParserResumes) {      if (p.scriptExecutedAfterParserResumes) {
770        log ('document.write: processed later (there is an unprocessed <script src>)');        log ('document.write: processed later (there is an unprocessed <script src>)');
771        logIndentLevel--;        logIndentLevel--;
772          log ('document.write: return');
773        return;        return;
774      }      }
775    
# Line 737  Line 783 
783      // to do something here?      // to do something here?
784    
785      // 5. Return      // 5. Return
786        logIndentLevel--;
787      log ('document.write: return');      log ('document.write: return');
788    
     logIndentLevel--;  
789      return;      return;
790    }; // document.write    }; // document.write
791    
792    JSDocument.prototype._insertExternalScript = function (uri) {    JSDocument.prototype._insertExternalScript = function (uri) {
793      var s = new JSElement (this, 'script');      var s = new JSElement (this, 'script');
794      s.src = uri;      s.src = uri;
795      this.documentElement.appendChild (s);      if (this.documentElement) {
796          this.documentElement.appendChild (s);
797          return true;
798        } else {
799          return false;
800        }
801    }; // _insertExternalScript    }; // _insertExternalScript
802    
803    JSDocument.prototype.__defineGetter__ ('documentElement', function () {    JSDocument.prototype.__defineGetter__ ('documentElement', function () {
# Line 846  algorithm.  Especially, this parser: Line 897  algorithm.  Especially, this parser:
897  algorithm, and so on.  algorithm, and so on.
898  <li>Does not raise parse errors for invalid attribute specifications in start  <li>Does not raise parse errors for invalid attribute specifications in start
899  or end tags.  or end tags.
900  <li>Does not support PCDATA elements (<code>title</code> and  <li>Does not support RCDATA elements (<code>title</code> and
901  <code>textarea</code>).  <code>textarea</code>).
902  <li>Does not strip the first newline in <code>pre</code> elements.  <li>Does not strip the first newline in <code>pre</code>,
903    <code>listing</code>, and <code>textarea</code> elements.
904  <li>Does not support <code>&lt;!--</code>..<code>--></code> parsing rule  <li>Does not support <code>&lt;!--</code>..<code>--></code> parsing rule
905  in <code>script</code> element.  in CDATA/RCDATA elements.
906  <li>Does not support foreign (SVG or MathML) elements.  <li>Does not support foreign (SVG or MathML) elements.
907  <li>Only supports <code>script</code> <code>type</code>  <li>Only supports <code>script</code> <code>type</code>
908  <code>text/javascript</code>.  <code>type</code> and <code>language</code>  <code>text/javascript</code>.  <code>type</code> and <code>language</code>
# Line 863  introduced, followed, or separated by wh Line 915  introduced, followed, or separated by wh
915    <li><code>var s = document.createElement ("script");    <li><code>var s = document.createElement ("script");
916              s.src = "<var>string</var>";              s.src = "<var>string</var>";
917              document.documentElement.appendChild (s);</code>              document.documentElement.appendChild (s);</code>
918      <li><code>w (document.documentElement.innerHTML);</code> (This statement
919      can be used to dump the document, even when the document has no
920      document element.  The output format is the tree dump format used
921      in html5lib test data, not <abbr>HTML</abbr>.)
922    </ul>    </ul>
923  Note that strings may be delimited by <code>'</code>s instead of  Note that strings may be delimited by <code>'</code>s instead of
924  <code>"</code>s.  <code>"</code>s.
Legend:
Removed from v.1.14  
changed lines
  Added in v.1.17
admin@suikawiki.org
 ViewVC Help
Powered by ViewVC 1.1.24