module/language/ecmascript/parse.scm

   1 ;;; ECMAScript for Guile
   2
   3 ;; Copyright (C) 2009 Free Software Foundation, Inc.
   4
   5 ;; This program is free software; you can redistribute it and/or modify
   6 ;; it under the terms of the GNU General Public License as published by
   7 ;; the Free Software Foundation; either version 2, or (at your option)
   8 ;; any later version.
   9 ;;
  10 ;; This program is distributed in the hope that it will be useful,
  11 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  12 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  13 ;; GNU General Public License for more details.
  14 ;;
  15 ;; You should have received a copy of the GNU General Public License
  16 ;; along with this program; see the file COPYING.  If not, write to
  17 ;; the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  18 ;; Boston, MA 02111-1307, USA.
  19
  20 ;;; Code:
  21
  22 (define-module (language ecmascript parse)
  23   #:use-module (language ecmascript parse-lalr)
  24   #:use-module (language ecmascript tokenize)
  25   #:export (read-ecmascript read-ecmascript/1 parse-ecmascript))
  26
  27 (define (syntax-error message . args)
  28   (apply throw 'SyntaxError message args))
  29
  30 (define (read-ecmascript port)
  31   (parse-ecmascript (make-tokenizer port) syntax-error))
  32
  33 (define (read-ecmascript/1 port)
  34   (parse-ecmascript (make-tokenizer/1 port) syntax-error))
  35
  36 (define *eof-object*
  37   (call-with-input-string "" read-char))
  38
  39 (define parse-ecmascript
  40   (lalr-parser
  41    ;; terminal (i.e. input) token types
  42    (lbrace rbrace lparen rparen lbracket rbracket dot semicolon comma <
  43     > <= >= == != === !== + - * % ++ -- << >> >>> & bor ^ ! ~ && or ?
  44     colon = += -= *= %= <<= >>= >>>= &= bor= ^= / /=
  45
  46     break else new var case finally return void catch for switch while
  47     continue function this with default if throw delete in try do
  48     instanceof typeof null true false
  49
  50     Identifier StringLiteral NumericLiteral RegexpLiteral)
  51
  52
  53    (Program (SourceElements) -> $1
  54             (*eoi*) -> *eof-object*)
  55
  56    ;;
  57    ;; Verily, here we define statements. Expressions are defined
  58    ;; afterwards.
  59    ;;
  60
  61    (SourceElement (Statement) -> $1
  62                   (FunctionDeclaration) -> $1)
  63
  64    (FunctionDeclaration (function Identifier lparen rparen lbrace FunctionBody rbrace) -> `(var (,$2 (lambda () ,$6)))
  65                         (function Identifier lparen FormalParameterList rparen lbrace FunctionBody rbrace) -> `(var (,$2 (lambda ,$4 ,$7))))
  66    (FunctionExpression (function lparen rparen lbrace FunctionBody rbrace) -> `(lambda () ,$5)
  67                        (function Identifier lparen rparen lbrace FunctionBody rbrace) -> `(lambda () ,$6)
  68                        (function lparen FormalParameterList rparen lbrace FunctionBody rbrace) -> `(lambda ,$3 ,$6)
  69                        (function Identifier lparen FormalParameterList rparen lbrace FunctionBody rbrace) -> `(lambda ,$4 ,$7))
  70    (FormalParameterList (Identifier) -> `(,$1)
  71                         (FormalParameterList comma Identifier) -> `(,@$1 ,$3))
  72    (SourceElements (SourceElement) -> $1
  73                    (SourceElements SourceElement) -> (if (and (pair? $1) (eq? (car $1) 'begin))
  74                                                          `(begin ,@(cdr $1) ,$2)
  75                                                          `(begin ,$1 ,$2)))
  76    (FunctionBody (SourceElements) -> $1)
  77
  78    (Statement (Block) -> $1
  79               (VariableStatement) -> $1
  80               (EmptyStatement) -> $1
  81               (ExpressionStatement) -> $1
  82               (IfStatement) -> $1
  83               (IterationStatement) -> $1
  84               (ContinueStatement) -> $1
  85               (BreakStatement) -> $1
  86               (ReturnStatement) -> $1
  87               (WithStatement) -> $1
  88               (LabelledStatement) -> $1
  89               (SwitchStatement) -> $1
  90               (ThrowStatement) -> $1
  91               (TryStatement) -> $1)
  92
  93    (Block (lbrace StatementList rbrace) -> `(block ,$2))
  94    (StatementList (Statement) -> $1
  95                   (StatementList Statement) -> (if (and (pair? $1) (eq? (car $1) 'begin))
  96                                                    `(begin ,@(cdr $1) ,$2)
  97                                                    `(begin ,$1 ,$2)))
  98
  99    (VariableStatement (var VariableDeclarationList) -> `(var ,@$2))
 100    (VariableDeclarationList (VariableDeclaration) -> `(,$1)
 101                             (VariableDeclarationList comma VariableDeclaration) -> `(,@$1 ,$2))
 102    (VariableDeclarationListNoIn (VariableDeclarationNoIn) -> `(,$1)
 103                                 (VariableDeclarationListNoIn comma VariableDeclarationNoIn) -> `(,@$1 ,$2))
 104    (VariableDeclaration (Identifier) -> `(,$1)
 105                         (Identifier Initialiser) -> `(,$1 ,$2))
 106    (VariableDeclarationNoIn (Identifier) -> `(,$1)
 107                             (Identifier Initialiser) -> `(,$1 ,$2))
 108    (Initialiser (= AssignmentExpression) -> $2)
 109    (InitialiserNoIn (= AssignmentExpressionNoIn) -> $2)
 110
 111    (EmptyStatement (semicolon) -> '(begin))
 112
 113    (ExpressionStatement (Expression semicolon) -> $1)
 114
 115    (IfStatement (if lparen Expression rparen Statement else Statement) -> `(if ,$3 ,$5 ,$7)
 116                 (if lparen Expression rparen Statement) -> `(if ,$3 ,$5))
 117
 118    (IterationStatement (do Statement while lparen Expression rparen semicolon) -> `(do ,$2 ,$5)
 119
 120                        (while lparen Expression rparen Statement) -> `(while ,$3 ,$5)
 121
 122                        (for lparen semicolon semicolon rparen Statement) -> `(for #f #f #f ,$6)
 123                        (for lparen semicolon semicolon Expression rparen Statement) -> `(for #f #f ,$5 ,$7)
 124                        (for lparen semicolon Expression semicolon rparen Statement) -> `(for #f ,$4 #f ,$7)
 125                        (for lparen semicolon Expression semicolon Expression rparen Statement) -> `(for #f ,$4 ,$6 ,$8)
 126
 127                        (for lparen ExpressionNoIn semicolon semicolon rparen Statement) -> `(for ,$3 #f #f ,$7)
 128                        (for lparen ExpressionNoIn semicolon semicolon Expression rparen Statement) -> `(for ,$3 #f ,$6 ,$8)
 129                        (for lparen ExpressionNoIn semicolon Expression semicolon rparen Statement) -> `(for ,$3 ,$5 #f ,$8)
 130                        (for lparen ExpressionNoIn semicolon Expression semicolon Expression rparen Statement) -> `(for ,$3 ,$5 ,$7 ,$9)
 131
 132                        (for lparen var VariableDeclarationListNoIn semicolon semicolon rparen Statement) -> `(for (var ,@$4) #f #f ,$8)
 133                        (for lparen var VariableDeclarationListNoIn semicolon semicolon Expression rparen Statement) -> `(for (var ,@$4) #f ,$7 ,$9)
 134                        (for lparen var VariableDeclarationListNoIn semicolon Expression semicolon rparen Statement) -> `(for (var ,@$4) ,$6 #f ,$9)
 135                        (for lparen var VariableDeclarationListNoIn semicolon Expression semicolon Expression rparen Statement) -> `(for (var ,@$4) ,$6 ,$8 ,$10)
 136
 137                        (for lparen LeftHandSideExpression in Expression rparen Statement) -> `(for-in ,$3 ,$5 ,$7)
 138                        (for lparen var VariableDeclarationNoIn in Expression rparen Statement) -> `(begin (var ,$4) (for-in (ref ,@$4) ,$6 ,$8)))
 139
 140    (ContinueStatement (continue Identifier semicolon) -> `(continue ,$2)
 141                       (continue semicolon) -> `(continue))
 142
 143    (BreakStatement (break Identifier semicolon) -> `(break ,$2)
 144                    (break semicolon) -> `(break))
 145
 146    (ReturnStatement (return Expression semicolon) -> `(return ,$2)
 147                     (return semicolon) -> `(return))
 148
 149    (WithStatement (with lparen Expression rparen Statement) -> `(with ,$3 ,$5))
 150
 151    (SwitchStatement (switch lparen Expression rparen CaseBlock) -> `(switch ,$3 ,@$5))
 152    (CaseBlock (lbrace rbrace) -> '()
 153               (lbrace CaseClauses rbrace) -> $2
 154               (lbrace CaseClauses DefaultClause rbrace) -> `(,@$2 ,@$3)
 155               (lbrace DefaultClause rbrace) -> `(,$2)
 156               (lbrace DefaultClause CaseClauses rbrace) -> `(,@$2 ,@$3))
 157    (CaseClauses (CaseClause) -> `(,$1)
 158                 (CaseClauses CaseClause) -> `(,@$1 ,$2))
 159    (CaseClause (case Expression colon) -> `(case ,$2)
 160                (case Expression colon StatementList) -> `(case ,$2 ,$4))
 161    (DefaultClause (default colon) -> `(default)
 162                   (default colon StatementList) -> `(default ,$3))
 163
 164    (LabelledStatement (Identifier colon Statement) -> `(label ,$1 ,$3))
 165
 166    (ThrowStatement (throw Expression semicolon) -> `(throw ,$2))
 167
 168    (TryStatement (try Block Catch) -> `(try ,$2 ,$3 #f)
 169                  (try Block Finally) -> `(try ,$2 #f ,$3)
 170                  (try Block Catch Finally) -> `(try ,$2 ,$3 ,$4))
 171    (Catch (catch lparen Identifier rparen Block) -> `(catch ,$3 ,$5))
 172    (Finally (finally Block) -> `(finally ,$2))
 173
 174    ;;
 175    ;; As promised, expressions. We build up to Expression bottom-up, so
 176    ;; as to get operator precedence right.
 177    ;;
 178
 179    (PrimaryExpression (this) -> 'this
 180                       (null) -> 'null
 181                       (true) -> 'true
 182                       (false) -> 'false
 183                       (Identifier) -> `(ref ,$1)
 184                       (StringLiteral) -> `(string ,$1)
 185                       (RegexpLiteral) -> `(regexp ,$1)
 186                       (NumericLiteral) -> `(number ,$1)
 187                       (ArrayLiteral) -> $1
 188                       (ObjectLiteral) -> $1
 189                       (lparen Expression rparen) -> $2)
 190
 191    (ArrayLiteral (lbracket rbracket) -> '(array)
 192                  (lbracket Elision rbracket) -> '(array ,@$2)
 193                  (lbracket ElementList rbracket) -> `(array ,@$2)
 194                  (lbracket ElementList comma rbracket) -> `(array ,@$2)
 195                  (lbracket ElementList comma Elision rbracket) -> `(array ,@$2))
 196    (ElementList (AssignmentExpression) -> `(,$1)
 197                 (Elision AssignmentExpression) -> `(,@$1 ,$2)
 198                 (ElementList comma AssignmentExpression) -> `(,@$1 ,$3)
 199                 (ElementList comma Elision AssignmentExpression) -> `(,@$1 ,@$3 ,$4))
 200    (Elision (comma) -> '((number 0))
 201             (Elision comma) -> `(,@$1 (number 0)))
 202
 203    (ObjectLiteral (lbrace rbrace) -> `(object)
 204                   (lbrace PropertyNameAndValueList rbrace) -> `(object ,@$2))
 205    (PropertyNameAndValueList (PropertyName colon AssignmentExpression) -> `((,$1 ,$3))
 206                              (PropertyNameAndValueList comma PropertyName colon AssignmentExpression) -> `(,@$1 (,$3 ,$5)))
 207    (PropertyName (Identifier) -> $1
 208                  (StringLiteral) -> (string->symbol $1)
 209                  (NumericLiteral) -> $1)
 210
 211    (MemberExpression (PrimaryExpression) -> $1
 212                      (FunctionExpression) -> $1
 213                      (MemberExpression lbracket Expression rbracket) -> `(aref ,$1 ,$3)
 214                      (MemberExpression dot Identifier) -> `(pref ,$1 ,$3)
 215                      (new MemberExpression Arguments) -> `(new ,$2 ,$3))
 216
 217    (NewExpression (MemberExpression) -> $1
 218                   (new NewExpression) -> `(new ,$2 ()))
 219
 220    (CallExpression (MemberExpression Arguments) -> `(call ,$1 ,$2)
 221                    (CallExpression Arguments) -> `(call ,$1 ,$2)
 222                    (CallExpression lbracket Expression rbracket) -> `(aref ,$1 ,$3)
 223                    (CallExpression dot Identifier) -> `(pref ,$1 ,$3))
 224    (Arguments (lparen rparen) -> '()
 225               (lparen ArgumentList rparen) -> $2)
 226    (ArgumentList (AssignmentExpression) -> `(,$1)
 227                  (ArgumentList comma AssignmentExpression) -> `(,@$1 ,$3))
 228
 229    (LeftHandSideExpression (NewExpression) -> $1
 230                            (CallExpression) -> $1)
 231
 232    (PostfixExpression (LeftHandSideExpression) -> $1
 233                       (LeftHandSideExpression ++) -> `(postinc ,$1)
 234                       (LeftHandSideExpression --) -> `(postdec ,$1))
 235
 236    (UnaryExpression (PostfixExpression) -> $1
 237                     (delete UnaryExpression) -> `(delete ,$2)
 238                     (void UnaryExpression) -> `(void ,$2)
 239                     (typeof UnaryExpression) -> `(typeof ,$2)
 240                     (++ UnaryExpression) -> `(preinc ,$2)
 241                     (-- UnaryExpression) -> `(predec ,$2)
 242                     (+ UnaryExpression) -> `(+ ,$2)
 243                     (- UnaryExpression) -> `(- ,$2)
 244                     (~ UnaryExpression) -> `(~ ,$2)
 245                     (! UnaryExpression) -> `(! ,$2))
 246
 247    (MultiplicativeExpression (UnaryExpression) -> $1
 248                              (MultiplicativeExpression * UnaryExpression) -> `(* ,$1 ,$3)
 249                              (MultiplicativeExpression / UnaryExpression) -> `(/ ,$1 ,$3)
 250                              (MultiplicativeExpression % UnaryExpression) -> `(% ,$1 ,$3))
 251
 252    (AdditiveExpression (MultiplicativeExpression) -> $1
 253                        (AdditiveExpression + MultiplicativeExpression) -> `(+ ,$1 ,$3)
 254                        (AdditiveExpression - MultiplicativeExpression) -> `(- ,$1 ,$3))
 255
 256    (ShiftExpression (AdditiveExpression) -> $1
 257                     (ShiftExpression << MultiplicativeExpression) -> `(<< ,$1 ,$3)
 258                     (ShiftExpression >> MultiplicativeExpression) -> `(>> ,$1 ,$3)
 259                     (ShiftExpression >>> MultiplicativeExpression) -> `(>>> ,$1 ,$3))
 260
 261    (RelationalExpression (ShiftExpression) -> $1
 262                          (RelationalExpression < ShiftExpression) -> `(< ,$1 ,$3)
 263                          (RelationalExpression > ShiftExpression) -> `(> ,$1 ,$3)
 264                          (RelationalExpression <= ShiftExpression) -> `(<= ,$1 ,$3)
 265                          (RelationalExpression >= ShiftExpression) -> `(>= ,$1 ,$3)
 266                          (RelationalExpression instanceof ShiftExpression) -> `(instanceof ,$1 ,$3)
 267                          (RelationalExpression in ShiftExpression) -> `(in ,$1 ,$3))
 268
 269    (RelationalExpressionNoIn (ShiftExpression) -> $1
 270                              (RelationalExpressionNoIn < ShiftExpression) -> `(< ,$1 ,$3)
 271                              (RelationalExpressionNoIn > ShiftExpression) -> `(> ,$1 ,$3)
 272                              (RelationalExpressionNoIn <= ShiftExpression) -> `(<= ,$1 ,$3)
 273                              (RelationalExpressionNoIn >= ShiftExpression) -> `(>= ,$1 ,$3)
 274                              (RelationalExpressionNoIn instanceof ShiftExpression) -> `(instanceof ,$1 ,$3))
 275
 276    (EqualityExpression (RelationalExpression) -> $1
 277                        (EqualityExpression == RelationalExpression) -> `(== ,$1 ,$3)
 278                        (EqualityExpression != RelationalExpression) -> `(!= ,$1 ,$3)
 279                        (EqualityExpression === RelationalExpression) -> `(=== ,$1 ,$3)
 280                        (EqualityExpression !== RelationalExpression) -> `(!== ,$1 ,$3))
 281
 282    (EqualityExpressionNoIn (RelationalExpressionNoIn) -> $1
 283                            (EqualityExpressionNoIn == RelationalExpressionNoIn) -> `(== ,$1 ,$3)
 284                            (EqualityExpressionNoIn != RelationalExpressionNoIn) -> `(!= ,$1 ,$3)
 285                            (EqualityExpressionNoIn === RelationalExpressionNoIn) -> `(=== ,$1 ,$3)
 286                            (EqualityExpressionNoIn !== RelationalExpressionNoIn) -> `(!== ,$1 ,$3))
 287
 288    (BitwiseANDExpression (EqualityExpression) -> $1
 289                          (BitwiseANDExpression & EqualityExpression) -> `(& ,$1 ,$3))
 290    (BitwiseANDExpressionNoIn (EqualityExpressionNoIn) -> $1
 291                              (BitwiseANDExpressionNoIn & EqualityExpressionNoIn) -> `(& ,$1 ,$3))
 292
 293    (BitwiseXORExpression (BitwiseANDExpression) -> $1
 294                          (BitwiseXORExpression ^ BitwiseANDExpression) -> `(^ ,$1 ,$3))
 295    (BitwiseXORExpressionNoIn (BitwiseANDExpressionNoIn) -> $1
 296                              (BitwiseXORExpressionNoIn ^ BitwiseANDExpressionNoIn) -> `(^ ,$1 ,$3))
 297
 298    (BitwiseORExpression (BitwiseXORExpression) -> $1
 299                         (BitwiseORExpression bor BitwiseXORExpression) -> `(bor ,$1 ,$3))
 300    (BitwiseORExpressionNoIn (BitwiseXORExpressionNoIn) -> $1
 301                             (BitwiseORExpressionNoIn bor BitwiseXORExpressionNoIn) -> `(bor ,$1 ,$3))
 302
 303    (LogicalANDExpression (BitwiseORExpression) -> $1
 304                          (LogicalANDExpression && BitwiseORExpression) -> `(and ,$1 ,$3))
 305    (LogicalANDExpressionNoIn (BitwiseORExpressionNoIn) -> $1
 306                              (LogicalANDExpressionNoIn && BitwiseORExpressionNoIn) -> `(and ,$1 ,$3))
 307
 308    (LogicalORExpression (LogicalANDExpression) -> $1
 309                         (LogicalORExpression or LogicalANDExpression) -> `(or ,$1 ,$3))
 310    (LogicalORExpressionNoIn (LogicalANDExpressionNoIn) -> $1
 311                             (LogicalORExpressionNoIn or LogicalANDExpressionNoIn) -> `(or ,$1 ,$3))
 312
 313    (ConditionalExpression (LogicalORExpression) -> $1
 314                           (LogicalORExpression ? AssignmentExpression colon AssignmentExpression) -> `(if ,$1 ,$3 ,$5))
 315    (ConditionalExpressionNoIn (LogicalORExpressionNoIn) -> $1
 316                               (LogicalORExpressionNoIn ? AssignmentExpressionNoIn colon AssignmentExpressionNoIn) -> `(if ,$1 ,$3 ,$5))
 317
 318    (AssignmentExpression (ConditionalExpression) -> $1
 319                          (LeftHandSideExpression AssignmentOperator AssignmentExpression) -> `(,$2 ,$1 ,$3))
 320    (AssignmentExpressionNoIn (ConditionalExpressionNoIn) -> $1
 321                              (LeftHandSideExpression AssignmentOperator AssignmentExpressionNoIn) -> `(,$2 ,$1 ,$3))
 322    (AssignmentOperator (=) -> '=
 323                        (*=) -> '*=
 324                        (/=) -> '/=
 325                        (%=) -> '%=
 326                        (+=) -> '+=
 327                        (-=) -> '-=
 328                        (<<=) -> '<<=
 329                        (>>=) -> '>>=
 330                        (>>>=) -> '>>>=
 331                        (&=) -> '&=
 332                        (^=) -> '^=
 333                        (bor=) -> 'bor=)
 334
 335    (Expression (AssignmentExpression) -> $1
 336                (Expression comma AssignmentExpression) -> `(begin ,$1 ,$3))
 337    (ExpressionNoIn (AssignmentExpressionNoIn) -> $1
 338                    (ExpressionNoIn comma AssignmentExpressionNoIn) -> `(begin ,$1 ,$3))))