git.roojs.org Git - gnome.introspection-doc-generator/blob - JSDOC/TokenReader.vala

   1 //<script type="text/javascript">
   2
   3
   4
   5
   6 //const Token   = imports.Token.Token;
   7 //const Lang    = imports.Lang.Lang;
   8
   9 /**
  10         @class Search a {@link JSDOC.TextStream} for language tokens.
  11 */
  12
  13 namespace JSDOC {
  14
  15     public class TokenArray: Object {
  16
  17         Gee.ArrayList<Token> tokens;
  18
  19         public TokenArray()
  20         {
  21             this.items = new Gee.ArrayList<Token>();
  22         }
  23
  24         public Token? last() {
  25             if (this.tokens > 0) {
  26                 return this.tokens[this.tokens.length-1];
  27             }
  28             return null;
  29         }
  30         public Token? lastSym = function() {
  31             for (var i = this.tokens.length-1; i >= 0; i--) {
  32                 if (!(this.tokens.get(i).is("WHIT") || this.tokens.get(i).is("COMM")))  {
  33                     return this.tokens.get(i);
  34                 }
  35             }
  36             return null;
  37         }
  38     }
  39
  40
  41     public class TokenReader : Object
  42     {
  43
  44
  45
  46         /*
  47          *
  48          * I wonder if this will accept the prop: value, prop2 :value construxtor if we do not define one...
  49          */
  50
  51         /** @cfg {Boolean} collapseWhite merge multiple whitespace/comments into a single token **/
  52         public bool collapseWhite = false, // only reduces white space...
  53         /** @cfg {Boolean} keepDocs keep JSDOC comments **/
  54         public bool keepDocs = true,
  55         /** @cfg {Boolean} keepWhite keep White space **/
  56         public bool keepWhite = false,
  57         /** @cfg {Boolean} keepComments  keep all comments **/
  58         public bool keepComments = false,
  59         /** @cfg {Boolean} sepIdents seperate identifiers (eg. a.b.c into ['a', '.', 'b', '.', 'c'] ) **/
  60         public bool sepIdents = false,
  61         /** @cfg {String} filename name of file being parsed. **/
  62         public string filename = "";
  63         /** @config {Boolean} ignoreBadGrammer do not throw errors if we find stuff that might break compression **/
  64         public bool ignoreBadGrammer = false,
  65
  66
  67         int line = 0;
  68
  69         /**
  70          * tokenize a stream
  71          * @return {Array} of tokens
  72          *
  73          * ts = new TextStream(File.read(str));
  74          * tr = TokenReader({ keepComments : true, keepWhite : true });
  75          * tr.tokenize(ts)
  76          *
  77          */
  78         public TokenArray tokenize(TextStream stream)
  79         {
  80             this.line =1;
  81             var tokens = new TokenArray();
  82
  83
  84             while (!stream.look().eof) {
  85                 if (this.read_mlcomment(stream, tokens)) continue;
  86                 if (this.read_slcomment(stream, tokens)) continue;
  87                 if (this.read_dbquote(stream, tokens))   continue;
  88                 if (this.read_snquote(stream, tokens))   continue;
  89                 if (this.read_regx(stream, tokens))      continue;
  90                 if (this.read_numb(stream, tokens))      continue;
  91                 if (this.read_punc(stream, tokens))      continue;
  92                 if (this.read_newline(stream, tokens))   continue;
  93                 if (this.read_space(stream, tokens))     continue;
  94                 if (this.read_word(stream, tokens))      continue;
  95
  96                 // if execution reaches here then an error has happened
  97                 tokens.push(new Token(stream.next(), "TOKN", "UNKNOWN_TOKEN", this.line));
  98             }
  99
 100
 101
 102             return tokens;
 103         },
 104
 105         /**
 106          * findPuncToken - find the id of a token (previous to current)
 107          * need to back check syntax..
 108          *
 109          * @arg {Array} tokens the array of tokens.
 110          * @arg {String} token data (eg. '(')
 111          * @arg {Number} offset where to start reading from
 112          * @return {Number} position of token
 113          */
 114         findPuncToken : function(tokens, data, n) {
 115             n = n || tokens.length -1;
 116             var stack = 0;
 117             while (n > -1) {
 118
 119                 if (!stack && tokens[n].data == data) {
 120                     return n;
 121                 }
 122
 123                 if (tokens[n].data  == ')' || tokens[n].data  == '}') {
 124                     stack++;
 125                     n--;
 126                     continue;
 127                 }
 128                 if (stack && (tokens[n].data  == '{' || tokens[n].data  == '(')) {
 129                     stack--;
 130                     n--;
 131                     continue;
 132                 }
 133
 134
 135                 n--;
 136             }
 137             return -1;
 138         },
 139         /**
 140          * lastSym - find the last token symbol
 141          * need to back check syntax..
 142          *
 143          * @arg {Array} tokens the array of tokens.
 144          * @arg {Number} offset where to start..
 145          * @return {Token} the token
 146          */
 147         lastSym : function(tokens, n) {
 148             for (var i = n-1; i >= 0; i--) {
 149                 if (!(tokens[i].is("WHIT") || tokens[i].is("COMM"))) return tokens[i];
 150             }
 151             return null;
 152         },
 153
 154
 155
 156         /**
 157             @returns {Boolean} Was the token found?
 158          */
 159         read_word : function(/**JSDOC.TokenStream*/stream, tokens) {
 160             var found = "";
 161             while (!stream.look().eof && Lang.isWordChar(stream.look())) {
 162                 found += stream.next();
 163             }
 164
 165             if (found === "") {
 166                 return false;
 167             }
 168
 169             var name;
 170             if ((name = Lang.keyword(found))) {
 171                 if (found == 'return' && tokens.lastSym().data == ')') {
 172                     //Seed.print('@' + tokens.length);
 173                     var n = this.findPuncToken(tokens, ')');
 174                     //Seed.print(')@' + n);
 175                     n = this.findPuncToken(tokens, '(', n-1);
 176                     //Seed.print('(@' + n);
 177
 178                     var lt = this.lastSym(tokens, n);
 179                     print(JSON.stringify(lt));
 180                     if (lt.type != 'KEYW' || ['IF', 'WHILE'].indexOf(lt.name) < -1) {
 181                         if (!this.ignoreBadGrammer) {
 182                             throw {
 183                                 name : "ArgumentError",
 184                                 message: "\n" + this.filename + ':' + this.line + " Error - return found after )"
 185                             }
 186                         }
 187                     }
 188
 189
 190
 191                 }
 192
 193                 tokens.push(new Token(found, "KEYW", name, this.line));
 194                 return true;
 195             }
 196             if (!this.sepIdents || found.indexOf('.') < 0 ) {
 197                 tokens.push(new Token(found, "NAME", "NAME", this.line));
 198                 return true;
 199             }
 200             var n = found.split('.');
 201             var p = false;
 202             var _this = this;
 203             n.forEach(function(nm) {
 204                 if (p) {
 205                     tokens.push(new Token('.', "PUNC", "DOT", _this.line));
 206                 }
 207                 p=true;
 208                 tokens.push(new Token(nm, "NAME", "NAME", _this.line));
 209             });
 210             return true;
 211
 212
 213         },
 214
 215         /**
 216             @returns {Boolean} Was the token found?
 217          */
 218         read_punc : function(/**JSDOC.TokenStream*/stream, tokens) {
 219             var found = "";
 220             var name;
 221             while (!stream.look().eof && Lang.punc(found+stream.look())) {
 222                 found += stream.next();
 223             }
 224
 225
 226             if (found === "") {
 227                 return false;
 228             }
 229
 230             if ((found == '}' || found == ']') && tokens.lastSym().data == ',') {
 231                 //print("Error - comma found before " + found);
 232                 //print(JSON.stringify(tokens.lastSym(), null,4));
 233                 if (this.ignoreBadGrammer) {
 234                     print("\n" + this.filename + ':' + this.line + " Error - comma found before " + found);
 235                 } else {
 236
 237                     throw {
 238                         name : "ArgumentError",
 239                         message: "\n" + this.filename + ':' + this.line + " Error - comma found before " + found
 240                     }
 241                 }
 242             }
 243
 244             tokens.push(new Token(found, "PUNC", Lang.punc(found), this.line));
 245             return true;
 246
 247         },
 248
 249         /**
 250             @returns {Boolean} Was the token found?
 251          */
 252         read_space : function(/**JSDOC.TokenStream*/stream, tokens) {
 253             var found = "";
 254
 255             while (!stream.look().eof && Lang.isSpace(stream.look()) && !Lang.isNewline(stream.look())) {
 256                 found += stream.next();
 257             }
 258
 259             if (found === "") {
 260                 return false;
 261             }
 262             //print("WHITE = " + JSON.stringify(found));
 263             if (this.collapseWhite) found = " ";
 264             if (this.keepWhite) tokens.push(new Token(found, "WHIT", "SPACE", this.line));
 265             return true;
 266
 267         },
 268
 269         /**
 270             @returns {Boolean} Was the token found?
 271          */
 272         read_newline : function(/**JSDOC.TokenStream*/stream, tokens) {
 273             var found = "";
 274             var line = this.line;
 275             while (!stream.look().eof && Lang.isNewline(stream.look())) {
 276                 this.line++;
 277                 found += stream.next();
 278             }
 279
 280             if (found === "") {
 281                 return false;
 282             }
 283             //this.line++;
 284             if (this.collapseWhite) {
 285                 found = "\n";
 286             }
 287              if (this.keepWhite) {
 288                 var last = tokens ? tokens.pop() : false;
 289                 if (last && last.name != "WHIT") {
 290                     tokens.push(last);
 291                 }
 292
 293                 tokens.push(new Token(found, "WHIT", "NEWLINE", line));
 294             }
 295             return true;
 296         },
 297
 298         /**
 299             @returns {Boolean} Was the token found?
 300          */
 301         read_mlcomment : function(/**JSDOC.TokenStream*/stream, tokens) {
 302             if (stream.look() == "/" && stream.look(1) == "*") {
 303                 var found = stream.next(2);
 304                 var c = '';
 305                 var line = this.line;
 306                 while (!stream.look().eof && !(stream.look(-1) == "/" && stream.look(-2) == "*")) {
 307                     c = stream.next();
 308                     if (c == "\n") this.line++;
 309                     found += c;
 310                 }
 311
 312                 // to start doclet we allow /** or /*** but not /**/ or /****
 313                 if (/^\/\*\*([^\/]|\*[^*])/.test(found) && this.keepDocs) tokens.push(new Token(found, "COMM", "JSDOC", this.line));
 314                 else if (this.keepComments) tokens.push(new Token(found, "COMM", "MULTI_LINE_COMM", line));
 315                 return true;
 316             }
 317             return false;
 318         },
 319
 320         /**
 321             @returns {Boolean} Was the token found?
 322          */
 323         read_slcomment : function(/**JSDOC.TokenStream*/stream, tokens) {
 324             var found;
 325             if (
 326                 (stream.look() == "/" && stream.look(1) == "/" && (found=stream.next(2)))
 327                 ||
 328                 (stream.look() == "<" && stream.look(1) == "!" && stream.look(2) == "-" && stream.look(3) == "-" && (found=stream.next(4)))
 329             ) {
 330                 var line = this.line;
 331                 while (!stream.look().eof && !Lang.isNewline(stream.look())) {
 332                     found += stream.next();
 333                 }
 334                 if (!stream.look().eof) {
 335                     found += stream.next();
 336                 }
 337                 if (this.keepComments) {
 338                     tokens.push(new Token(found, "COMM", "SINGLE_LINE_COMM", line));
 339                 }
 340                 this.line++;
 341                 return true;
 342             }
 343             return false;
 344         },
 345
 346         /**
 347             @returns {Boolean} Was the token found?
 348          */
 349         read_dbquote : function(/**JSDOC.TokenStream*/stream, tokens) {
 350             if (stream.look() == "\"") {
 351                 // find terminator
 352                 var string = stream.next();
 353
 354                 while (!stream.look().eof) {
 355                     if (stream.look() == "\\") {
 356                         if (Lang.isNewline(stream.look(1))) {
 357                             do {
 358                                 stream.next();
 359                             } while (!stream.look().eof && Lang.isNewline(stream.look()));
 360                             string += "\\\n";
 361                         }
 362                         else {
 363                             string += stream.next(2);
 364                         }
 365                     }
 366                     else if (stream.look() == "\"") {
 367                         string += stream.next();
 368                         tokens.push(new Token(string, "STRN", "DOUBLE_QUOTE", this.line));
 369                         return true;
 370                     }
 371                     else {
 372                         string += stream.next();
 373                     }
 374                 }
 375             }
 376             return false; // error! unterminated string
 377         },
 378
 379         /**
 380             @returns {Boolean} Was the token found?
 381          */
 382         read_snquote : function(/**JSDOC.TokenStream*/stream, tokens) {
 383             if (stream.look() == "'") {
 384                 // find terminator
 385                 var string = stream.next();
 386
 387                 while (!stream.look().eof) {
 388                     if (stream.look() == "\\") { // escape sequence
 389                         string += stream.next(2);
 390                     }
 391                     else if (stream.look() == "'") {
 392                         string += stream.next();
 393                         tokens.push(new Token(string, "STRN", "SINGLE_QUOTE", this.line));
 394                         return true;
 395                     }
 396                     else {
 397                         string += stream.next();
 398                     }
 399                 }
 400             }
 401             return false; // error! unterminated string
 402         },
 403
 404         /**
 405             @returns {Boolean} Was the token found?
 406          */
 407         read_numb : function(/**JSDOC.TokenStream*/stream, tokens) {
 408             if (stream.look() === "0" && stream.look(1) == "x") {
 409                 return this.read_hex(stream, tokens);
 410             }
 411
 412             var found = "";
 413
 414             while (!stream.look().eof && Lang.isNumber(found+stream.look())){
 415                 found += stream.next();
 416             }
 417
 418             if (found === "") {
 419                 return false;
 420             }
 421             else {
 422                 if (/^0[0-7]/.test(found)) tokens.push(new Token(found, "NUMB", "OCTAL", this.line));
 423                 else tokens.push(new Token(found, "NUMB", "DECIMAL", this.line));
 424                 return true;
 425             }
 426         },
 427         /*t:
 428             requires("../lib/JSDOC/TextStream.js");
 429             requires("../lib/JSDOC/Token.js");
 430             requires("../lib/JSDOC/Lang.js");
 431
 432             plan(3, "testing read_numb");
 433
 434             //// setup
 435             var src = "function foo(num){while (num+8.0 >= 0x20 && num < 0777){}}";
 436             var tr = new TokenReader();
 437             var tokens = tr.tokenize(new TextStream(src));
 438
 439             var hexToken, octToken, decToken;
 440             for (var i = 0; i < tokens.length; i++) {
 441                 if (tokens[i].name == "HEX_DEC") hexToken = tokens[i];
 442                 if (tokens[i].name == "OCTAL") octToken = tokens[i];
 443                 if (tokens[i].name == "DECIMAL") decToken = tokens[i];
 444             }
 445             ////
 446
 447             is(decToken.data, "8.0", "decimal number is found in source.");
 448             is(hexToken.data, "0x20", "hexdec number is found in source (issue #99).");
 449             is(octToken.data, "0777", "octal number is found in source.");
 450         */
 451
 452         /**
 453             @returns {Boolean} Was the token found?
 454          */
 455         read_hex : function(/**JSDOC.TokenStream*/stream, tokens) {
 456             var found = stream.next(2);
 457
 458             while (!stream.look().eof) {
 459                 if (Lang.isHexDec(found) && !Lang.isHexDec(found+stream.look())) { // done
 460                     tokens.push(new Token(found, "NUMB", "HEX_DEC", this.line));
 461                     return true;
 462                 }
 463                 else {
 464                     found += stream.next();
 465                 }
 466             }
 467             return false;
 468         },
 469
 470         /**
 471             @returns {Boolean} Was the token found?
 472          */
 473         read_regx : function(/**JSDOC.TokenStream*/stream, tokens) {
 474             var last;
 475             if (
 476                 stream.look() == "/"
 477                 &&
 478                 (
 479
 480                     (
 481                         !(last = tokens.lastSym()) // there is no last, the regex is the first symbol
 482                         ||
 483                         (
 484                                !last.is("NUMB")
 485                             && !last.is("NAME")
 486                             && !last.is("RIGHT_PAREN")
 487                             && !last.is("RIGHT_BRACKET")
 488                         )
 489                     )
 490                 )
 491             ) {
 492                 var regex = stream.next();
 493
 494                 while (!stream.look().eof) {
 495                     if (stream.look() == "\\") { // escape sequence
 496                         regex += stream.next(2);
 497                     }
 498                     else if (stream.look() == "/") {
 499                         regex += stream.next();
 500
 501                         while (/[gmi]/.test(stream.look())) {
 502                             regex += stream.next();
 503                         }
 504
 505                         tokens.push(new Token(regex, "REGX", "REGX", this.line));
 506                         return true;
 507                     }
 508                     else {
 509                         regex += stream.next();
 510                     }
 511                 }
 512                 // error: unterminated regex
 513             }
 514             return false;
 515         }
 516 });