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         public Gee.ArrayList<Token> tokens;
  18         public int length {
  19             get { return this.tokens.size }
  20         }
  21
  22         public TokenArray()
  23         {
  24             this.items = new Gee.ArrayList<Token>();
  25         }
  26
  27         public Token? last() {
  28             if (this.tokens > 0) {
  29                 return this.tokens[this.tokens.length-1];
  30             }
  31             return null;
  32         }
  33         public Token? lastSym () {
  34             for (var i = this.tokens.length-1; i >= 0; i--) {
  35                 if (!(this.tokens.get(i).is("WHIT") || this.tokens.get(i).is("COMM")))  {
  36                     return this.tokens.get(i);
  37                 }
  38             }
  39             return null;
  40         }
  41         public void push (Token t) {
  42             this.tokens.add(t);
  43         }
  44         public Token? pop ()
  45         {
  46             if (this.size > 0) {
  47                 return this.tokens.remove_at(this.size-1);
  48             }
  49             return null;
  50         }
  51
  52         public Token get(int i) {
  53             return this.tokens.get(i);
  54         }
  55     }
  56
  57     errordomain TokenReader_Error {
  58             ArgumentError
  59     }
  60
  61
  62     public class TokenReader : Object
  63     {
  64
  65
  66
  67         /*
  68          *
  69          * I wonder if this will accept the prop: value, prop2 :value construxtor if we do not define one...
  70          */
  71
  72         /** @cfg {Boolean} collapseWhite merge multiple whitespace/comments into a single token **/
  73         public bool collapseWhite = false, // only reduces white space...
  74         /** @cfg {Boolean} keepDocs keep JSDOC comments **/
  75         public bool keepDocs = true,
  76         /** @cfg {Boolean} keepWhite keep White space **/
  77         public bool keepWhite = false,
  78         /** @cfg {Boolean} keepComments  keep all comments **/
  79         public bool keepComments = false,
  80         /** @cfg {Boolean} sepIdents seperate identifiers (eg. a.b.c into ['a', '.', 'b', '.', 'c'] ) **/
  81         public bool sepIdents = false,
  82         /** @cfg {String} filename name of file being parsed. **/
  83         public string filename = "";
  84         /** @config {Boolean} ignoreBadGrammer do not throw errors if we find stuff that might break compression **/
  85         public bool ignoreBadGrammer = false,
  86
  87
  88         int line = 0;
  89
  90         /**
  91          * tokenize a stream
  92          * @return {Array} of tokens
  93          *
  94          * ts = new TextStream(File.read(str));
  95          * tr = TokenReader({ keepComments : true, keepWhite : true });
  96          * tr.tokenize(ts)
  97          *
  98          */
  99         public TokenArray tokenize(TextStream stream)
 100         {
 101             this.line =1;
 102             var tokens = new TokenArray();
 103
 104             bool eof;
 105             while (!stream.lookEOF()) {
 106
 107
 108                 if (this.read_mlcomment(stream, tokens)) continue;
 109                 if (this.read_slcomment(stream, tokens)) continue;
 110                 if (this.read_dbquote(stream, tokens))   continue;
 111                 if (this.read_snquote(stream, tokens))   continue;
 112                 if (this.read_regx(stream, tokens))      continue;
 113                 if (this.read_numb(stream, tokens))      continue;
 114                 if (this.read_punc(stream, tokens))      continue;
 115                 if (this.read_newline(stream, tokens))   continue;
 116                 if (this.read_space(stream, tokens))     continue;
 117                 if (this.read_word(stream, tokens))      continue;
 118
 119                 // if execution reaches here then an error has happened
 120                 tokens.push(
 121                         new Token(stream.next(), "TOKN", "UNKNOWN_TOKEN", this.line)
 122                 );
 123             }
 124
 125
 126
 127             return tokens;
 128         }
 129
 130         /**
 131          * findPuncToken - find the id of a token (previous to current)
 132          * need to back check syntax..
 133          *
 134          * @arg {Array} tokens the array of tokens.
 135          * @arg {String} token data (eg. '(')
 136          * @arg {Number} offset where to start reading from
 137          * @return {Number} position of token
 138          */
 139         public int findPuncToken(TokenArray tokens, string data, int n)
 140         {
 141             n = n || tokens.length -1;
 142             var stack = 0;
 143             while (n > -1) {
 144
 145                 if (!stack && tokens.get(n).data == data) {
 146                     return n;
 147                 }
 148
 149                 if (tokens.get(n).data  == ')' || tokens.get(n).data  == '}') {
 150                     stack++;
 151                     n--;
 152                     continue;
 153                 }
 154                 if (stack && (tokens.get(n).data  == '{' || tokens.get(n).data  == '(')) {
 155                     stack--;
 156                     n--;
 157                     continue;
 158                 }
 159
 160
 161                 n--;
 162             }
 163             return -1;
 164         }
 165         /**
 166          * lastSym - find the last token symbol
 167          * need to back check syntax..
 168          *
 169          * @arg {Array} tokens the array of tokens.
 170          * @arg {Number} offset where to start..
 171          * @return {Token} the token
 172          */
 173         public Token lastSym(TokenArray tokens, int n)
 174         {
 175             for (var i = n-1; i >= 0; i--) {
 176                 if (!(tokens.get(i).is("WHIT") || tokens.get(i).is("COMM"))) {
 177                     return tokens.get(i);
 178                 }
 179             }
 180             return null;
 181         }
 182
 183
 184
 185         /**
 186             @returns {Boolean} Was the token found?
 187          */
 188         public bool read_word (TokenStream stream, TokenArray tokens)
 189         {
 190             string found = "";
 191             while (!stream.lookEOF() && Lang.isWordChar(stream.look())) {
 192                 found += stream.next();
 193             }
 194
 195             if (found == "") {
 196                 return false;
 197             }
 198
 199             var name = Lang.keyword(found);
 200             if (name != null) {
 201
 202                 // look for "()return" ?? why ???
 203                 var ls = tokens.lastSym();
 204                 if (found == "return" && ls != null && ls.data == ")") {
 205                     //Seed.print('@' + tokens.length);
 206                     var n = this.findPuncToken(tokens, ")");
 207                     //Seed.print(')@' + n);
 208                     n = this.findPuncToken(tokens, "(", n-1);
 209                     //Seed.print('(@' + n);
 210
 211                     var lt = this.lastSym(tokens, n);
 212
 213                     //print(JSON.stringify(lt));
 214                     if (lt.type != "KEYW" || ["IF", 'WHILE'].indexOf(lt.name) < -1) {
 215                         if (!this.ignoreBadGrammer) {
 216                             throw new TokenReader_Error.ArgumentError(
 217                                 this.filename + ":" + this.line + " Error - return found after )"
 218                             );
 219                         }
 220                     }
 221
 222
 223
 224                 }
 225
 226                 tokens.push(new Token(found, "KEYW", name, this.line));
 227                 return true;
 228             }
 229
 230             if (!this.sepIdents || found.indexOf('.') < 0 ) {
 231                 tokens.push(new Token(found, "NAME", "NAME", this.line));
 232                 return true;
 233             }
 234             var n = found.split('.');
 235             var p = false;
 236             foreach (unowned string nm in n) {
 237                 if (p) {
 238                     tokens.push(new Token('.', "PUNC", "DOT", this.line));
 239                 }
 240                 p=true;
 241                 tokens.push(new Token(nm, "NAME", "NAME", this.line));
 242             }
 243             return true;
 244
 245
 246         }
 247
 248         /**
 249             @returns {Boolean} Was the token found?
 250          */
 251         public bool read_punc (TokenStream stream, TokenArray tokens)
 252         {
 253             string found = "";
 254             var name;
 255             while (!stream.lookEOF() && Lang.punc(found + stream.look()).length > 0) {
 256                 found += stream.next();
 257             }
 258
 259
 260             if (found === "") {
 261                 return false;
 262             }
 263
 264             var ls = tokens.lastSym();
 265
 266             if ((found == "}" || found == "]") && ls != null && ls.data == ",") {
 267                 //print("Error - comma found before " + found);
 268                 //print(JSON.stringify(tokens.lastSym(), null,4));
 269                 if (this.ignoreBadGrammer) {
 270                     print("\n" + this.filename + ':' + this.line + " Error - comma found before " + found);
 271                 } else {
 272                     throw new TokenReader_Error.ArgumentError(
 273                                 this.filename + ":" + this.line + "  comma found before " + found
 274
 275                     );
 276
 277                 }
 278             }
 279
 280             tokens.push(new Token(found, "PUNC", Lang.punc(found), this.line));
 281             return true;
 282
 283         }
 284
 285         /**
 286             @returns {Boolean} Was the token found?
 287          */
 288         public bool read_space  (TokenStream stream, TokenArray tokens)
 289         {
 290             var found = "";
 291
 292             while (!stream.lookEOF() && Lang.isSpace(stream.look()) && !Lang.isNewline(stream.look())) {
 293                 found += stream.next();
 294             }
 295
 296             if (found === "") {
 297                 return false;
 298             }
 299             //print("WHITE = " + JSON.stringify(found));
 300
 301
 302             if (this.collapseWhite) {
 303                 found = " "; // this might work better if it was a '\n' ???
 304             }
 305             if (this.keepWhite) {
 306                 tokens.push(new Token(found, "WHIT", "SPACE", this.line));
 307             }
 308             return true;
 309
 310         }
 311
 312         /**
 313             @returns {Boolean} Was the token found?
 314          */
 315         public bool read_newline  (TokenStream stream, TokenArray tokens)
 316             var found = "";
 317             var line = this.line;
 318             while (!stream.lookEOF() && Lang.isNewline(stream.look())) {
 319                 this.line++;
 320                 found += stream.next();
 321             }
 322
 323             if (found === "") {
 324                 return false;
 325             }
 326
 327             // if we found a new line, then we could check if previous character was a ';' - if so we can drop it.
 328             // otherwise generally keep it.. in which case it should reduce our issue with stripping new lines..
 329
 330
 331             //this.line++;
 332             if (this.collapseWhite) {
 333                 found = "\n"; // reduces multiple line breaks into a single one...
 334             }
 335
 336             if (this.keepWhite) {
 337                 var last = tokens.pop();
 338                 if (last != null && last.name != "WHIT") {
 339                     tokens.push(last);
 340                 }
 341                 // replaces last new line...
 342                 tokens.push(new Token(found, "WHIT", "NEWLINE", line));
 343             }
 344             return true;
 345         },
 346
 347         /**
 348             @returns {Boolean} Was the token found?
 349          */
 350         read_mlcomment : function(/**JSDOC.TokenStream*/stream, tokens) {
 351             if (stream.look() == "/" && stream.look(1) == "*") {
 352                 var found = stream.next(2);
 353                 var c = '';
 354                 var line = this.line;
 355                 while (!stream.look().eof && !(stream.look(-1) == "/" && stream.look(-2) == "*")) {
 356                     c = stream.next();
 357                     if (c == "\n") this.line++;
 358                     found += c;
 359                 }
 360
 361                 // to start doclet we allow /** or /*** but not /**/ or /****
 362                 if (/^\/\*\*([^\/]|\*[^*])/.test(found) && this.keepDocs) tokens.push(new Token(found, "COMM", "JSDOC", this.line));
 363                 else if (this.keepComments) tokens.push(new Token(found, "COMM", "MULTI_LINE_COMM", line));
 364                 return true;
 365             }
 366             return false;
 367         },
 368
 369         /**
 370             @returns {Boolean} Was the token found?
 371          */
 372         read_slcomment : function(/**JSDOC.TokenStream*/stream, tokens) {
 373             var found;
 374             if (
 375                 (stream.look() == "/" && stream.look(1) == "/" && (found=stream.next(2)))
 376                 ||
 377                 (stream.look() == "<" && stream.look(1) == "!" && stream.look(2) == "-" && stream.look(3) == "-" && (found=stream.next(4)))
 378             ) {
 379                 var line = this.line;
 380                 while (!stream.look().eof && !Lang.isNewline(stream.look())) {
 381                     found += stream.next();
 382                 }
 383                 if (!stream.look().eof) {
 384                     found += stream.next();
 385                 }
 386                 if (this.keepComments) {
 387                     tokens.push(new Token(found, "COMM", "SINGLE_LINE_COMM", line));
 388                 }
 389                 this.line++;
 390                 return true;
 391             }
 392             return false;
 393         },
 394
 395         /**
 396             @returns {Boolean} Was the token found?
 397          */
 398         read_dbquote : function(/**JSDOC.TokenStream*/stream, tokens) {
 399             if (stream.look() == "\"") {
 400                 // find terminator
 401                 var string = stream.next();
 402
 403                 while (!stream.look().eof) {
 404                     if (stream.look() == "\\") {
 405                         if (Lang.isNewline(stream.look(1))) {
 406                             do {
 407                                 stream.next();
 408                             } while (!stream.look().eof && Lang.isNewline(stream.look()));
 409                             string += "\\\n";
 410                         }
 411                         else {
 412                             string += stream.next(2);
 413                         }
 414                     }
 415                     else if (stream.look() == "\"") {
 416                         string += stream.next();
 417                         tokens.push(new Token(string, "STRN", "DOUBLE_QUOTE", this.line));
 418                         return true;
 419                     }
 420                     else {
 421                         string += stream.next();
 422                     }
 423                 }
 424             }
 425             return false; // error! unterminated string
 426         },
 427
 428         /**
 429             @returns {Boolean} Was the token found?
 430          */
 431         read_snquote : function(/**JSDOC.TokenStream*/stream, tokens) {
 432             if (stream.look() == "'") {
 433                 // find terminator
 434                 var string = stream.next();
 435
 436                 while (!stream.look().eof) {
 437                     if (stream.look() == "\\") { // escape sequence
 438                         string += stream.next(2);
 439                     }
 440                     else if (stream.look() == "'") {
 441                         string += stream.next();
 442                         tokens.push(new Token(string, "STRN", "SINGLE_QUOTE", this.line));
 443                         return true;
 444                     }
 445                     else {
 446                         string += stream.next();
 447                     }
 448                 }
 449             }
 450             return false; // error! unterminated string
 451         },
 452
 453         /**
 454             @returns {Boolean} Was the token found?
 455          */
 456         read_numb : function(/**JSDOC.TokenStream*/stream, tokens) {
 457             if (stream.look() === "0" && stream.look(1) == "x") {
 458                 return this.read_hex(stream, tokens);
 459             }
 460
 461             var found = "";
 462
 463             while (!stream.look().eof && Lang.isNumber(found+stream.look())){
 464                 found += stream.next();
 465             }
 466
 467             if (found === "") {
 468                 return false;
 469             }
 470             else {
 471                 if (/^0[0-7]/.test(found)) tokens.push(new Token(found, "NUMB", "OCTAL", this.line));
 472                 else tokens.push(new Token(found, "NUMB", "DECIMAL", this.line));
 473                 return true;
 474             }
 475         },
 476         /*t:
 477             requires("../lib/JSDOC/TextStream.js");
 478             requires("../lib/JSDOC/Token.js");
 479             requires("../lib/JSDOC/Lang.js");
 480
 481             plan(3, "testing read_numb");
 482
 483             //// setup
 484             var src = "function foo(num){while (num+8.0 >= 0x20 && num < 0777){}}";
 485             var tr = new TokenReader();
 486             var tokens = tr.tokenize(new TextStream(src));
 487
 488             var hexToken, octToken, decToken;
 489             for (var i = 0; i < tokens.length; i++) {
 490                 if (tokens[i].name == "HEX_DEC") hexToken = tokens[i];
 491                 if (tokens[i].name == "OCTAL") octToken = tokens[i];
 492                 if (tokens[i].name == "DECIMAL") decToken = tokens[i];
 493             }
 494             ////
 495
 496             is(decToken.data, "8.0", "decimal number is found in source.");
 497             is(hexToken.data, "0x20", "hexdec number is found in source (issue #99).");
 498             is(octToken.data, "0777", "octal number is found in source.");
 499         */
 500
 501         /**
 502             @returns {Boolean} Was the token found?
 503          */
 504         read_hex : function(/**JSDOC.TokenStream*/stream, tokens) {
 505             var found = stream.next(2);
 506
 507             while (!stream.look().eof) {
 508                 if (Lang.isHexDec(found) && !Lang.isHexDec(found+stream.look())) { // done
 509                     tokens.push(new Token(found, "NUMB", "HEX_DEC", this.line));
 510                     return true;
 511                 }
 512                 else {
 513                     found += stream.next();
 514                 }
 515             }
 516             return false;
 517         },
 518
 519         /**
 520             @returns {Boolean} Was the token found?
 521          */
 522         read_regx : function(/**JSDOC.TokenStream*/stream, tokens) {
 523             var last;
 524             if (
 525                 stream.look() == "/"
 526                 &&
 527                 (
 528
 529                     (
 530                         !(last = tokens.lastSym()) // there is no last, the regex is the first symbol
 531                         ||
 532                         (
 533                                !last.is("NUMB")
 534                             && !last.is("NAME")
 535                             && !last.is("RIGHT_PAREN")
 536                             && !last.is("RIGHT_BRACKET")
 537                         )
 538                     )
 539                 )
 540             ) {
 541                 var regex = stream.next();
 542
 543                 while (!stream.look().eof) {
 544                     if (stream.look() == "\\") { // escape sequence
 545                         regex += stream.next(2);
 546                     }
 547                     else if (stream.look() == "/") {
 548                         regex += stream.next();
 549
 550                         while (/[gmi]/.test(stream.look())) {
 551                             regex += stream.next();
 552                         }
 553
 554                         tokens.push(new Token(regex, "REGX", "REGX", this.line));
 555                         return true;
 556                     }
 557                     else {
 558                         regex += stream.next();
 559                     }
 560                 }
 561                 // error: unterminated regex
 562             }
 563             return false;
 564         }
 565 });