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