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         public bool read_mlcomment  (TokenStream stream, TokenArray tokens)
 351         {
 352             if (stream.look() != "/") {
 353                 return false;
 354             }
 355             if (stream.look(1) != "*") {
 356                 return false;
 357             }
 358             var found = stream.next(2);
 359             var c = '';
 360             var line = this.line;
 361             while (!stream.lookEOF() && !(stream.look(-1) == "/" && stream.look(-2) == "*")) {
 362                 c = stream.next();
 363                 if (c == "\n") {
 364                     this.line++;
 365                 }
 366                 found += c;
 367             }
 368
 369             // to start doclet we allow /** or /*** but not /**/ or /****
 370             //if (found.length /^\/\*\*([^\/]|\*[^*])/.test(found) && this.keepDocs) {
 371             if ((this.keepDocs && found.length > 4 && found.index_of("/**") == 0 && found[3] != "/") {
 372                 tokens.push(new Token(found, "COMM", "JSDOC", this.line));
 373             } else if (this.keepComments) {
 374                 tokens.push(new Token(found, "COMM", "MULTI_LINE_COMM", line));
 375             }
 376             return true;
 377
 378         }
 379
 380         /**
 381             @returns {Boolean} Was the token found?
 382          */
 383          public bool read_slcomment  (TokenStream stream, TokenArray tokens)
 384          {
 385             var found = "";
 386             if (
 387                 (stream.look() == "/" && stream.look(1) == "/" && (found=stream.next(2)))
 388                 ||
 389                 (stream.look() == "<" && stream.look(1) == "!" && stream.look(2) == "-" && stream.look(3) == "-" && (found=stream.next(4)))
 390             ) {
 391                 var line = this.line;
 392                 while (!stream.lookEOF() && !Lang.isNewline(stream.look())) {
 393                     found += stream.next();
 394                 }
 395                 //if (!stream.lookEOF()) { // what? << eat the EOL?
 396                     found += stream.next();
 397                 //}
 398                 if (this.keepComments) {
 399                     tokens.push(new Token(found, "COMM", "SINGLE_LINE_COMM", line));
 400                 }
 401                 this.line++;
 402                 return true;
 403             }
 404             return false;
 405         }
 406
 407         /**
 408             @returns {Boolean} Was the token found?
 409          */
 410         public bool read_dbquote  (TokenStream stream, TokenArray tokens)
 411         {
 412             if (stream.look() != "\"") {
 413                 return false;
 414             }
 415                 // find terminator
 416             var str = stream.next();
 417
 418             while (!stream.lookEOF()) {
 419                 if (stream.look() == "\\") {
 420                     if (Lang.isNewline(stream.look(1))) {
 421                         do {
 422                             stream.next();
 423                         } while (!stream.lookEOF() && Lang.isNewline(stream.look()));
 424                         str += "\\\n";
 425                     }
 426                     else {
 427                         str += stream.next(2);
 428                     }
 429                     continue;
 430                 }
 431                 if (stream.look() == "\"") {
 432                     str += stream.next();
 433                     tokens.push(new Token(str, "STRN", "DOUBLE_QUOTE", this.line));
 434                     return true;
 435                 }
 436
 437                 str += stream.next();
 438
 439             }
 440             return false;
 441         },
 442
 443         /**
 444             @returns {Boolean} Was the token found?
 445          */
 446         public bool read_snquote  (TokenStream stream, TokenArray tokens)
 447         {
 448             if (stream.look() != "'") {
 449                 return false;
 450             }
 451             // find terminator
 452             var str = stream.next();
 453
 454             while (!stream.look().eof) {
 455                 if (stream.look() == "\\") { // escape sequence
 456                     str += stream.next(2);
 457                     continue;
 458                 }
 459                 if (stream.look() == "'") {
 460                     str += stream.next();
 461                     tokens.push(new Token(str, "STRN", "SINGLE_QUOTE", this.line));
 462                     return true;
 463                 }
 464                 str += stream.next();
 465
 466             }
 467             return false;
 468         }
 469
 470
 471         /**
 472             @returns {Boolean} Was the token found?
 473          */
 474         public bool read_numb  (TokenStream stream, TokenArray tokens)
 475         {
 476             if (stream.look() === "0" && stream.look(1) == "x") {
 477                 return this.read_hex(stream, tokens);
 478             }
 479
 480             var found = "";
 481
 482             while (!stream.lookEOF() && Lang.isNumber(found+stream.look())){
 483                 found += stream.next();
 484             }
 485
 486             if (found === "") {
 487                 return false;
 488             }
 489             if (GLib.Regex.match_simple("^0[0-7]", found)) {
 490                 tokens.push(new Token(found, "NUMB", "OCTAL", this.line));
 491                 return true;
 492             }
 493             tokens.push(new Token(found, "NUMB", "DECIMAL", this.line));
 494             return true;
 495
 496         }
 497
 498         /**
 499             @returns {Boolean} Was the token found?
 500          */
 501         public bool read_hex  (TokenStream stream, TokenArray tokens)
 502         {
 503             var found = stream.next(2);
 504
 505             while (!stream.lookEOF()) {
 506                 if (Lang.isHexDec(found) && !Lang.isHexDec(found+stream.look())) { // done
 507                     tokens.push(new Token(found, "NUMB", "HEX_DEC", this.line));
 508                     return true;
 509                 }
 510
 511                 found += stream.next();
 512
 513             }
 514             return false;
 515         },
 516
 517         /**
 518             @returns {Boolean} Was the token found?
 519          */
 520         public bool read_regx (TokenStream stream, TokenArray tokens)
 521         {
 522             Token last;
 523             if (stream.look() != "/") {
 524                 return false;
 525             }
 526             var last = tokens.lastSym();
 527             if (
 528                 (last == null)
 529                 ||
 530                 (
 531                        !last.is("NUMB")   // stuff that can not appear before a regex..
 532                     && !last.is("NAME")
 533                     && !last.is("RIGHT_PAREN")
 534                     && !last.is("RIGHT_BRACKET")
 535                 )
 536             )  {
 537                 var regex = stream.next();
 538
 539                 while (!stream.lookEOF()) {
 540                     if (stream.look() == "\\") { // escape sequence
 541                         regex += stream.next(2);
 542                         continue;
 543                     }
 544                     if (stream.look() == "/") {
 545                         regex += stream.next();
 546
 547                         while (GLib.Regex.match_simple("[gmi]", stream.look()) {
 548                             regex += stream.next();
 549                         }
 550
 551                         tokens.push(new Token(regex, "REGX", "REGX", this.line));
 552                         return true;
 553                     }
 554
 555                     regex += stream.next();
 556
 557                 }
 558                 // error: unterminated regex
 559             }
 560             return false;
 561         }
 562     }
 563 }