/[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 1.2 by wakaba, Sun Apr 20 07:48:00 2008 UTC revision 1.5 by wakaba, Fri Apr 25 11:40:56 2008 UTC
# Line 13  Line 13 
13    output {    output {
14      display: block;      display: block;
15      font-family: monospace;      font-family: monospace;
16      white-space: pre;      white-space: -moz-pre-wrap;
17        white-space: pre-wrap;
18    }    }
19  </style>  </style>
20  <script>  <script>
21    function update () {    function update () {
22      document.logElement.textContent = '';      document.logElement.textContent = '';
23      var p = new Parser (new InputStream (document.sourceElement.value));      var p = new Parser (new InputStream (document.sourceElement.value));
24        var doc = p.doc;
25      p.parse ();      p.parse ();
26      log (dumpTree (p.doc, ''));      log (dumpTree (doc, ''));
27    } // update    } // update
28    
29    function log (s) {    function log (s) {
# Line 32  Line 34 
34      this.s = s;      this.s = s;
35    } // InputStream    } // InputStream
36    
37    function Parser (i) {    function Parser (i, doc) {
38      this.parseMode = 'pcdata';      this.parseMode = 'pcdata';
39      this.doc = new JSDocument (this);      if (!doc) {
40      this.openElements = [this.doc];        doc = new JSDocument (this);
41          doc.manakaiIsHTML = true;
42        }
43        this.doc = doc;
44        this.openElements = [doc];
45      this.in = i;      this.in = i;
46        this.scriptsExecutedAfterParsing = [];
47    } // Parser    } // Parser
48    
49    Parser.prototype.getNextToken = function () {    Parser.prototype.getNextToken = function () {
50        var p = this;
51      var i = this.in;      var i = this.in;
52      if (this.parseMode == 'script') {      if (this.parseMode == 'script') {
53        var token;        var token;
54        i.s = i.s.replace (/^([\s\S]+?)<\/[Ss][Cc][Rr][Ii][Pp][Tt]>/,        if (p.insertionPoint <= 0) {
55            return {type: 'abort'};
56          }
57          i.s = i.s.replace (/^([^<]+)/,
58        function (s, t) {        function (s, t) {
59            if (0 < p.insertionPoint && p.insertionPoint < t.length) {
60              token = {type: 'char', value: t.substring (0, p.insertionPoint)};
61              var ip = p.insertionPoint;
62              p.insertionPoint = 0;
63              return t.substring (ip, t.length);
64            }
65          token = {type: 'char', value: t};          token = {type: 'char', value: t};
66          return '<' + '/script>';          p.insertionPoint -= t.length;
67            return '';
68        });        });
69        if (token) return token;        if (token) return token;
70        i.s = i.s.replace (/^<\/[Ss][Cc][Rr][Ii][Pp][Tt]>/, function () {        i.s = i.s.replace (/^<\/[Ss][Cc][Rr][Ii][Pp][Tt]>/, function (s) {
71            if (p.insertionPoint < s.length) {
72              token = {type: 'abort'};
73              return s;
74            }
75          token = {type: 'end-tag', value: 'script'};          token = {type: 'end-tag', value: 'script'};
76            p.insertionPoint -= s.length;
77            return '';
78          });
79          if (token) return token;
80          var m;
81          if ((p.insertionPoint < '</script'.length) &&
82              (m = i.s.match (/^<\/([SCRIPTscript]+)/))) {
83            var v = m[1].substring (0, p.insertionPoint).toLowerCase ();
84            if (v == 'script'.substring (0, p.insertionPoint - '</'.length)) {
85              return {type: 'abort'};
86            }
87          }
88          i.s = i.s.replace (/^</,
89          function (s) {
90            token = {type: 'char', value: s};
91            p.insertionPoint -= s.length;
92          return '';          return '';
93        });        });
94        if (token) return token;        if (token) return token;
# Line 58  Line 96 
96      }      }
97    
98      var token;      var token;
99      i.s = i.s.replace (/^<\/([^>]+)>/, function (s, e) {      i.s = i.s.replace (/^<\/([^>]+)(?:>|$)/, function (s, e) {
100          if (p.insertionPoint < s.length ||
101              (p.insertionPoint <= s.length &&
102               s.substring (s.length - 1, 1) != '>')) {
103            token = {type: 'abort'};
104            return s;
105          }
106        token = {type: 'end-tag', value: e.toLowerCase ()};        token = {type: 'end-tag', value: e.toLowerCase ()};
107          p.insertionPoint -= s.length;
108        return '';        return '';
109      });      });
110      if (token) return token;      if (token) return token;
111      i.s = i.s.replace (/^<([^>]+)>/, function (s, e) {      i.s = i.s.replace (/^<([^>]+)(?:>|$)/, function (s, e) {
112        token = {type: 'start-tag', value: e.toLowerCase ()};        if (p.insertionPoint < s.length ||
113              (p.insertionPoint <= s.length &&
114               s.substring (s.length - 1, 1) != '>')) {
115            token = {type: 'abort'};
116            return s;
117          }
118          var tagName;
119          var attrs = {};
120          e = e.replace (/^[\S]+/, function (v) {
121            tagName = v.toLowerCase ();
122            return '';
123          });
124          e = e.replace (/^\s*(\S+)\s*(?:=\s*"([^"]*)"|'([^']*)'|([^"']+))?/,
125          function (x, attrName, attrValue1, attrValue2, attrValue3) {
126            attrs[attrName] = attrValue1 || attrValue2 || attrValue3;
127            return '';
128          });
129          token = {type: 'start-tag', value: tagName, attrs: attrs};
130          p.insertionPoint -= s.length;
131        return '';        return '';
132      });      });
133      if (token) return token;      if (token) return token;
134        if (p.insertionPoint <= 0) {
135          return {type: 'abort'};
136        }
137      i.s = i.s.replace (/^[^<]+/, function (s) {      i.s = i.s.replace (/^[^<]+/, function (s) {
138          if (p.insertionPoint < s.length) {
139            token = {type: 'char', value: s.substring (0, p.insertionPoint)};
140            var ip = p.insertionPoint;
141            p.insertionPoint = 0;
142            return s.substring (ip, s.length);
143          }
144        token = {type: 'char', value: s};        token = {type: 'char', value: s};
145          p.insertionPoint -= s.length;
146        return '';        return '';
147      });      });
148      if (token) return token;      if (token) return token;
149      i.s = i.s.replace (/^[\s\S]/, function (s) {      i.s = i.s.replace (/^[\s\S]/, function (s) {
150        token = {type: 'char', value: s};        token = {type: 'char', value: s};
151          p.insertionPoint -= s.length;
152        return '';        return '';
153      });      });
154      if (token) return token;      if (token) return token;
# Line 92  Line 166 
166          if (token.value == 'script') {          if (token.value == 'script') {
167            // 1. Create an element for the token in the HTML namespace.            // 1. Create an element for the token in the HTML namespace.
168            var el = new JSElement (this.doc, token.value);            var el = new JSElement (this.doc, token.value);
169              if (token.attrs.async != null) el.async = true;
170              if (token.attrs.defer != null) el.defer = true;
171              if (token.attrs.src != null) el.src = token.attrs.src;
172    
173            // 2. Mark the element as being "parser-inserted".            // 2. Mark the element as being "parser-inserted".
174            el.manakaiParserInserted = true;            el.manakaiParserInserted = true;
# Line 109  Line 186 
186                el.manakaiAppendText (token.value);                el.manakaiAppendText (token.value);
187    
188              // 4.2. Until it returns a token that is not a character token, or              // 4.2. Until it returns a token that is not a character token, or
189              // TODO: 4.3. Until it stops tokenising.              // until it stops tokenising.
190              } else if (token.type == 'eof' ||              } else if (token.type == 'eof' ||
191                         (token.type == 'end-tag' && token.value == 'script')) {                         (token.type == 'end-tag' && token.value == 'script') ||
192                           token.type == 'abort') {
193                // 6. Switched back to the PCDATA state.                // 6. Switched back to the PCDATA state.
194                this.parseMode = 'pcdata';                this.parseMode = 'pcdata';
195    
# Line 138  Line 216 
216            }            }
217    
218            // 9.1. Let the old insertion point have the same value as the ...            // 9.1. Let the old insertion point have the same value as the ...
219              var oldInsertionPoint = this.insertionPoint;
220            // 9.2. Let the insertion point be just before the next input ...            // 9.2. Let the insertion point be just before the next input ...
221              this.setInsertionPoint (0);
222    
223            // 10. Append the new element to the current node.            // 10. Append the new element to the current node.
224            this.openElements[this.openElements.length - 1].appendChild (el);            this.openElements[this.openElements.length - 1].appendChild (el);
225    
226            // 11. Let the insertion point have the value of the old ...            // 11. Let the insertion point have the value of the old ...
227              oldInsertionPoint += this.insertionPoint;
228              this.setInsertionPoint (oldInsertionPoint);
229    
230            // 12. If there is a script that will execute as soon as ...            // 12. If there is a script that will execute as soon as ...
231                        
# Line 161  Line 242 
242          } else {          } else {
243            log ('parse error: unmatched end tag: ' + token.value);            log ('parse error: unmatched end tag: ' + token.value);
244          }          }
245          } else if (token.type == 'char') {
246            this.openElements[this.openElements.length - 1].manakaiAppendText
247                (token.value);
248        } else if (token.type == 'eof') {        } else if (token.type == 'eof') {
249          break;          break;
250          } else if (token.type == 'abort') {
251            log ('parse: abort');
252            return;
253        }        }
254      }      }
255    
256      log ('stop parsing');      log ('stop parsing');
257    
258        // readyState = 'interactive'
259    
260        // "When a script completes loading" rules start applying.
261    
262        // TODO: Handles "list of scripts that will execute as soon as possible"
263        // and "list of scripts that will execute asynchronously"
264    
265        // Handle "list of scripts that will execute when the document has finished
266        // parsing".
267        var list = this.scriptsExecutedAfterParsing;
268        while (list.length > 0) {
269          // TODO: break unless completed loading
270    
271          // Step 1.
272          //
273    
274          // Step 2. and Step 3.
275          log ('Executing a |defer|red script...');
276          executeScript (this.doc, list.shift ());
277    
278          // Step 4.
279        }
280    
281        log ('DOMContentLoaded event fired');
282    
283        // "delays tha load event" things has completed:
284        // readyState = 'complete'
285        log ('load event fired');
286    } // parse    } // parse
287    
288      Parser.prototype.setInsertionPoint = function (ip) {
289        if (ip == undefined || ip == null || isNaN (ip)) {
290          log ('insertion point: set to undefined');
291          this.insertionPoint = undefined;
292        } else if (ip == this.in.s.length) {
293          log ('insertion point: end of file');
294          this.insertionPoint = ip;
295        } else {
296          log ('insertion point: set to ' + ip +
297               ' (before "' + this.in.s.substring (0, 10) + '")');
298          this.insertionPoint = ip;
299        }
300      }; // setInsertionPoint
301    
302    function JSDocument (p) {    function JSDocument (p) {
303      this.childNodes = [];      this.childNodes = [];
304      this._parser = p;      this._parser = p;
# Line 186  Line 316 
316      e.parentNode = this;      e.parentNode = this;
317    
318      if (e.localName == 'script') {      if (e.localName == 'script') {
319        log ('start running a script');        log ('Running a script: start');
320    
321        var doc = this.ownerDocument;        var doc = this.ownerDocument || this;
322        var p = doc._parser;        var p = doc._parser;
323    
324        // 1. Script type        // 1. Script type
# Line 203  Line 333 
333        // 2.4. If the script element has its "already executed" flag set        // 2.4. If the script element has its "already executed" flag set
334        if (e.manakaiAlreadyExecuted) {        if (e.manakaiAlreadyExecuted) {
335          // 2.5. Abort these steps at this point.          // 2.5. Abort these steps at this point.
336          log ('running a script: aborted');          log ('Running a script: aborted');
337          return e;          return e;
338        }        }
339    
# Line 218  Line 348 
348        // 5.1.        // 5.1.
349        if (/* TODO: If the document is still being parsed && */        if (/* TODO: If the document is still being parsed && */
350            e.defer && !e.async) {            e.defer && !e.async) {
351          // TODO          p.scriptsExecutedAfterParsing.push (e);
352            log ('Running a script: aborted (defer)');
353        } else if (e.async && e.src != null) {        } else if (e.async && e.src != null) {
354          // TODO          // TODO
355        } else if (e.async && e.src == null        } else if (e.async && e.src == null
# Line 232  Line 363 
363          executeScript (doc, e); // even if other scripts are already executing.          executeScript (doc, e); // even if other scripts are already executing.
364        }        }
365    
366        log ('end running a script');        log ('Running a script: end');
367      }      }
368    
369      return e;      return e;
# Line 299  Line 430 
430      }      }
431    }; // manakaiAppendText    }; // manakaiAppendText
432    
433      JSDocument.prototype.open = function () {
434        // Two or fewer arguments
435    
436        // Step 1.
437        var type = arguments[0] || 'text/html';
438        
439        // Step 2.
440        var replace = arguments[1] == 'replace';
441    
442        // Step 3.
443        if (this._parser &&
444            !this._parser.scriptCreated &&
445            this._parser.in.insertionPoint != undefined) {
446          log ('document.open () in parsing mode is ignored');
447          return this;
448        }
449    
450        // Step 4.
451        log ('onbeforeunload event fired');
452        log ('onunload event fired');
453    
454        // Step 5.
455        if (this._parser) {
456          // Discard the parser.
457        }
458    
459        // Step 6.
460        log ('document cleared by document.open ()');
461        this.childNodes = [];
462    
463        // Step 7.
464        this._parser = new Parser (new InputStream (''), this);
465        this._parser.scriptCreated = true;
466    
467        // Step 8.
468        this.manakaiIsHTML = true;
469    
470        // Step 9.
471        // If not text/html, ...
472    
473        // Step 10.
474        if (!replace) {
475          // History      
476        }
477    
478        // Step 11.
479        this._parser.setInsertionPoint (this._parser.in.s.length);
480    
481        // Step 12.
482        return this;
483      }; // document.open
484    
485    JSDocument.prototype.write = function () {    JSDocument.prototype.write = function () {
486        var p = this._parser;
487    
488      // 1. If the insertion point is undefined, the open() method must be ...      // 1. If the insertion point is undefined, the open() method must be ...
489      //      if (isNaN (p.insertionPoint) || p.insertionPoint == undefined) {
490          this.open ();
491          p = this._parser;
492        }
493    
494      // 2. ... inserted into the input stream just before the insertion point.      // 2. ... inserted into the input stream just before the insertion point.
495      log ('document.write: insert "' + Array.join (arguments, '') + '"');      var s = Array.join (arguments, '');
496        log ('document.write: insert "' + s + '"' +
497             ' before "' + p.in.s.substring (p.insertionPoint, p.insertionPoint + 10) + '"');
498        p.in.s = p.in.s.substring (0, p.insertionPoint) + s
499            + p.in.s.substring (p.insertionPoint, p.in.s.length);
500        p.insertionPoint += s.length;
501    
502      // 3. If there is a script that will execute as soon as the parser resumes      // 3. If there is a script that will execute as soon as the parser resumes
503      // TODO      // TODO
504    
505      // 4. Process the characters that were inserted, ...      // 4. Process the characters that were inserted, ...
506        p.parse ();
507    
508      // 5. Return      // 5. Return
509      log ('document.write: return');      log ('document.write: return');
# Line 332  Line 526 
526        var node = n.childNodes[i];        var node = n.childNodes[i];
527        if (node instanceof JSElement) {        if (node instanceof JSElement) {
528          r += '| ' + indent + node.localName + '\n';          r += '| ' + indent + node.localName + '\n';
529            if (node.async) r += '| ' + indent + '  async=""\n';
530            if (node.defer) r += '| ' + indent + '  defer=""\n';
531            if (node.src) r += '| ' + indent + '  src="' + node.src + '"\n';
532          r += dumpTree (node, indent + '  ');          r += dumpTree (node, indent + '  ');
533        } else if (node instanceof JSText) {        } else if (node instanceof JSText) {
534          r += '| ' + indent + '"' + node.data + '"\n';          r += '| ' + indent + '"' + node.data + '"\n';
# Line 353  Line 550 
550  &lt;head>&lt;/head>&lt;body>  &lt;head>&lt;/head>&lt;body>
551  &lt;p>  &lt;p>
552  &lt;script>  &lt;script>
553  document.write ('aaaaaaa&lt;/p>\n&lt;script>\ndocument.write("cccccc")\n&lt;/', 'script>\nbbbbbb');  document.write ('aaaaaaa&lt;/p>&lt;script>document.write("cccccc");&lt;/', 'script>bbbbbb');
554  &lt;/script>  &lt;/script>
555  &lt;p>  &lt;p>
556  </textarea>  </textarea>
Legend:
Removed from v.1.2  
changed lines
  Added in v.1.5
admin@suikawiki.org
 ViewVC Help
Powered by ViewVC 1.1.24