3 * Copyright (c) 1991-2002, 2003, 2004 Larry Wall
5 * You may distribute under the terms of either the GNU General Public
6 * License or the Artistic License, as specified in the README file.
11 * 'I see,' laughed Strider. 'I look foul and feel fair. Is that it?
12 * All that is gold does not glitter, not all those who wander are lost.'
15 /* This file holds the grammar for the Perl language. If edited, you need
16 * to run regen_perly.pl, which re-creates the files perly.h, perly.tab
17 * and perly.act which are derived from this.
19 * The main job of of this grammar is to call the various newFOO()
20 * functions in op.c to build a syntax tree of OP structs.
21 * It relies on the lexer in toke.c to do the tokenizing.
24 /* Make the parser re-entrant. */
39 %token <opval> WORD METHOD FUNCMETH THING PMFUNC PRIVATEREF
40 %token <opval> FUNC0SUB UNIOPSUB LSTOPSUB
42 %token <ival> FORMAT SUB ANONSUB PACKAGE USE
43 %token <ival> WHILE UNTIL IF UNLESS ELSE ELSIF CONTINUE FOR
44 %token <ival> LOOPEX DOTDOT
45 %token <ival> FUNC0 FUNC1 FUNC UNIOP LSTOP
46 %token <ival> RELOP EQOP MULOP ADDOP
47 %token <ival> DOLSHARP DO HASHBRACK NOAMP
48 %token <ival> LOCAL MY MYSUB REQUIRE
51 %type <ival> prog decl format startsub startanonsub startformsub mintro
52 %type <ival> progstart remember mremember '&' savescope
53 %type <opval> block mblock lineseq line loop cond else
54 %type <opval> expr term subscripted scalar ary hsh arylen star amper sideff
55 %type <opval> argexpr nexpr texpr iexpr mexpr mnexpr miexpr
56 %type <opval> listexpr listexprcom indirob listop method
57 %type <opval> formname subname proto subbody cont my_scalar
58 %type <opval> subattrlist myattrlist mysubrout myattrterm myterm
59 %type <opval> termbinop termunop anonymous termdo
65 %left <ival> OROP DOROP
68 %nonassoc LSTOP LSTOPSUB
70 %right <ival> ASSIGNOP
79 %nonassoc UNIOP UNIOPSUB
84 %right '!' '~' UMINUS REFGEN
86 %nonassoc PREINC PREDEC POSTINC POSTDEC
94 /* The whole program */
97 { $$ = $1; newPROG(block_end($1,$2)); }
100 /* An ordinary block */
101 block : '{' remember lineseq '}'
102 { if (PL_copline > (line_t)$1)
103 PL_copline = (line_t)$1;
104 $$ = block_end($2, $3); }
107 remember: /* NULL */ /* start a full lexical scope */
108 { $$ = block_start(TRUE); }
113 PL_expect = XSTATE; $$ = block_start(TRUE);
118 mblock : '{' mremember lineseq '}'
119 { if (PL_copline > (line_t)$1)
120 PL_copline = (line_t)$1;
121 $$ = block_end($2, $3); }
124 mremember: /* NULL */ /* start a partial lexical scope */
125 { $$ = block_start(FALSE); }
128 savescope: /* NULL */ /* remember stack pos in case of error */
129 { $$ = PL_savestack_ix; }
131 /* A collection of "lines" in the program */
136 | lineseq savescope line
138 $$ = append_list(OP_LINESEQ,
139 (LISTOP*)$1, (LISTOP*)$3);
140 PL_pad_reset_pending = TRUE;
141 if ($1 && $3) PL_hints |= HINT_BLOCK_SCOPE; }
144 /* A "line" in the program */
146 { $$ = newSTATEOP(0, $1, $2); }
147 | loop /* loops add their own labels */
149 { if ($1 != Nullch) {
150 $$ = newSTATEOP(0, $1, newOP(OP_NULL, 0));
156 PL_expect = XSTATE; }
158 { $$ = newSTATEOP(0, $1, $2);
159 PL_expect = XSTATE; }
162 /* An expression which may have a side-effect */
168 { $$ = newLOGOP(OP_AND, 0, $3, $1); }
170 { $$ = newLOGOP(OP_OR, 0, $3, $1); }
172 { $$ = newLOOPOP(OPf_PARENS, 1, scalar($3), $1); }
174 { $$ = newLOOPOP(OPf_PARENS, 1, $3, $1);}
176 { $$ = newFOROP(0, Nullch, (line_t)$2,
177 Nullop, $3, $1, Nullop); }
180 /* else and elsif blocks */
184 { ($2)->op_flags |= OPf_PARENS; $$ = scope($2); }
185 | ELSIF '(' mexpr ')' mblock else
186 { PL_copline = (line_t)$1;
187 $$ = newCONDOP(0, $3, scope($5), $6);
188 PL_hints |= HINT_BLOCK_SCOPE; }
191 /* Real conditional expressions */
192 cond : IF '(' remember mexpr ')' mblock else
193 { PL_copline = (line_t)$1;
195 newCONDOP(0, $4, scope($6), $7)); }
196 | UNLESS '(' remember miexpr ')' mblock else
197 { PL_copline = (line_t)$1;
199 newCONDOP(0, $4, scope($6), $7)); }
202 /* Continue blocks */
209 /* Loops: while, until, for, and a bare block */
210 loop : label WHILE '(' remember texpr ')' mintro mblock cont
211 { PL_copline = (line_t)$2;
214 newWHILEOP(0, 1, (LOOP*)Nullop,
215 $2, $5, $8, $9, $7))); }
216 | label UNTIL '(' remember iexpr ')' mintro mblock cont
217 { PL_copline = (line_t)$2;
220 newWHILEOP(0, 1, (LOOP*)Nullop,
221 $2, $5, $8, $9, $7))); }
222 | label FOR MY remember my_scalar '(' mexpr ')' mblock cont
224 newFOROP(0, $1, (line_t)$2, $5, $7, $9, $10)); }
225 | label FOR scalar '(' remember mexpr ')' mblock cont
227 newFOROP(0, $1, (line_t)$2, mod($3, OP_ENTERLOOP),
229 | label FOR '(' remember mexpr ')' mblock cont
231 newFOROP(0, $1, (line_t)$2, Nullop, $5, $7, $8)); }
232 | label FOR '(' remember mnexpr ';' texpr ';' mintro mnexpr ')'
234 /* basically fake up an initialize-while lineseq */
236 PL_copline = (line_t)$2;
237 forop = newSTATEOP(0, $1,
238 newWHILEOP(0, 1, (LOOP*)Nullop,
242 forop = append_elem(OP_LINESEQ,
243 newSTATEOP(0, ($1?savepv($1):Nullch),
248 $$ = block_end($4, forop); }
249 | label block cont /* a block is a loop that happens once */
250 { $$ = newSTATEOP(0, $1,
251 newWHILEOP(0, 1, (LOOP*)Nullop,
252 NOLINE, Nullop, $2, $3, 0)); }
255 /* determine whether there are any new my declarations */
257 { $$ = (PL_min_intro_pending &&
258 PL_max_intro_pending >= PL_min_intro_pending);
261 /* Normal expression */
267 /* Boolean expression */
268 texpr : /* NULL means true */
269 { (void)scan_num("1", &yylval); $$ = yylval.opval; }
273 /* Inverted boolean expression */
275 { $$ = invert(scalar($1)); }
278 /* Expression with its own lexical scope */
280 { $$ = $1; intro_my(); }
284 { $$ = $1; intro_my(); }
288 { $$ = $1; intro_my(); }
291 /* Optional "MAIN:"-style loop labels */
297 /* Some kind of declaration - does not take part in the parse tree */
310 format : FORMAT startformsub formname block
311 { newFORM($2, $3, $4); }
314 formname: WORD { $$ = $1; }
315 | /* NULL */ { $$ = Nullop; }
318 /* Unimplemented "my sub foo { }" */
319 mysubrout: MYSUB startsub subname proto subattrlist subbody
320 { newMYSUB($2, $3, $4, $5, $6); }
323 /* Subroutine definition */
324 subrout : SUB startsub subname proto subattrlist subbody
325 { newATTRSUB($2, $3, $4, $5, $6); }
328 startsub: /* NULL */ /* start a regular subroutine scope */
329 { $$ = start_subparse(FALSE, 0); }
332 startanonsub: /* NULL */ /* start an anonymous subroutine scope */
333 { $$ = start_subparse(FALSE, CVf_ANON); }
336 startformsub: /* NULL */ /* start a format subroutine scope */
337 { $$ = start_subparse(TRUE, 0); }
340 /* Name of a subroutine - must be a bareword, could be special */
341 subname : WORD { const char *const name = SvPV_nolen_const(((SVOP*)$1)->op_sv);
342 if (strEQ(name, "BEGIN") || strEQ(name, "END")
343 || strEQ(name, "INIT") || strEQ(name, "CHECK"))
344 CvSPECIAL_on(PL_compcv);
348 /* Subroutine prototype */
354 /* Optional list of subroutine attributes */
355 subattrlist: /* NULL */
363 /* List of attributes for a "my" variable declaration */
364 myattrlist: COLONATTR THING
370 /* Subroutine body - either null or a block */
371 subbody : block { $$ = $1; }
372 | ';' { $$ = Nullop; PL_expect = XSTATE; }
375 package : PACKAGE WORD ';'
380 { CvSPECIAL_on(PL_compcv); /* It's a BEGIN {} */ }
381 WORD WORD listexpr ';'
382 { utilize($1, $2, $4, $5, $6); }
385 /* Ordinary expressions; logical combinations */
386 expr : expr ANDOP expr
387 { $$ = newLOGOP(OP_AND, 0, $1, $3); }
389 { $$ = newLOGOP($2, 0, $1, $3); }
391 { $$ = newLOGOP(OP_DOR, 0, $1, $3); }
392 | argexpr %prec PREC_LOW
395 /* Expressions are a list of terms joined by commas */
396 argexpr : argexpr ','
399 { $$ = append_elem(OP_LIST, $1, $3); }
400 | term %prec PREC_LOW
404 listop : LSTOP indirob argexpr /* map {...} @args or print $fh @args */
405 { $$ = convert($1, OPf_STACKED,
406 prepend_elem(OP_LIST, newGVREF($1,$2), $3) ); }
407 | FUNC '(' indirob expr ')' /* print ($fh @args */
408 { $$ = convert($1, OPf_STACKED,
409 prepend_elem(OP_LIST, newGVREF($1,$3), $4) ); }
410 | term ARROW method '(' listexprcom ')' /* $foo->bar(list) */
411 { $$ = convert(OP_ENTERSUB, OPf_STACKED,
413 prepend_elem(OP_LIST, scalar($1), $5),
414 newUNOP(OP_METHOD, 0, $3))); }
415 | term ARROW method /* $foo->bar */
416 { $$ = convert(OP_ENTERSUB, OPf_STACKED,
417 append_elem(OP_LIST, scalar($1),
418 newUNOP(OP_METHOD, 0, $3))); }
419 | METHOD indirob listexpr /* new Class @args */
420 { $$ = convert(OP_ENTERSUB, OPf_STACKED,
422 prepend_elem(OP_LIST, $2, $3),
423 newUNOP(OP_METHOD, 0, $1))); }
424 | FUNCMETH indirob '(' listexprcom ')' /* method $object (@args) */
425 { $$ = convert(OP_ENTERSUB, OPf_STACKED,
427 prepend_elem(OP_LIST, $2, $4),
428 newUNOP(OP_METHOD, 0, $1))); }
429 | LSTOP listexpr /* print @args */
430 { $$ = convert($1, 0, $2); }
431 | FUNC '(' listexprcom ')' /* print (@args) */
432 { $$ = convert($1, 0, $3); }
433 | LSTOPSUB startanonsub block /* sub f(&@); f { foo } ... */
434 { $3 = newANONATTRSUB($2, 0, Nullop, $3); }
435 listexpr %prec LSTOP /* ... @bar */
436 { $$ = newUNOP(OP_ENTERSUB, OPf_STACKED,
438 prepend_elem(OP_LIST, $3, $5), $1)); }
441 /* Names of methods. May use $object->$methodname */
446 /* Some kind of subscripted expression */
447 subscripted: star '{' expr ';' '}' /* *main::{something} */
448 /* In this and all the hash accessors, ';' is
449 * provided by the tokeniser */
450 { $$ = newBINOP(OP_GELEM, 0, $1, scalar($3));
451 PL_expect = XOPERATOR; }
452 | scalar '[' expr ']' /* $array[$element] */
453 { $$ = newBINOP(OP_AELEM, 0, oopsAV($1), scalar($3)); }
454 | term ARROW '[' expr ']' /* somearef->[$element] */
455 { $$ = newBINOP(OP_AELEM, 0,
456 ref(newAVREF($1),OP_RV2AV),
458 | subscripted '[' expr ']' /* $foo->[$bar]->[$baz] */
459 { $$ = newBINOP(OP_AELEM, 0,
460 ref(newAVREF($1),OP_RV2AV),
462 | scalar '{' expr ';' '}' /* $foo->{bar();} */
463 { $$ = newBINOP(OP_HELEM, 0, oopsHV($1), jmaybe($3));
464 PL_expect = XOPERATOR; }
465 | term ARROW '{' expr ';' '}' /* somehref->{bar();} */
466 { $$ = newBINOP(OP_HELEM, 0,
467 ref(newHVREF($1),OP_RV2HV),
469 PL_expect = XOPERATOR; }
470 | subscripted '{' expr ';' '}' /* $foo->[bar]->{baz;} */
471 { $$ = newBINOP(OP_HELEM, 0,
472 ref(newHVREF($1),OP_RV2HV),
474 PL_expect = XOPERATOR; }
475 | term ARROW '(' ')' /* $subref->() */
476 { $$ = newUNOP(OP_ENTERSUB, OPf_STACKED,
477 newCVREF(0, scalar($1))); }
478 | term ARROW '(' expr ')' /* $subref->(@args) */
479 { $$ = newUNOP(OP_ENTERSUB, OPf_STACKED,
480 append_elem(OP_LIST, $4,
481 newCVREF(0, scalar($1)))); }
483 | subscripted '(' expr ')' /* $foo->{bar}->(@args) */
484 { $$ = newUNOP(OP_ENTERSUB, OPf_STACKED,
485 append_elem(OP_LIST, $3,
486 newCVREF(0, scalar($1)))); }
487 | subscripted '(' ')' /* $foo->{bar}->() */
488 { $$ = newUNOP(OP_ENTERSUB, OPf_STACKED,
489 newCVREF(0, scalar($1))); }
490 | '(' expr ')' '[' expr ']' /* list slice */
491 { $$ = newSLICEOP(0, $5, $2); }
492 | '(' ')' '[' expr ']' /* empty list slice! */
493 { $$ = newSLICEOP(0, $4, Nullop); }
496 /* Binary operators between terms */
497 termbinop : term ASSIGNOP term /* $x = $y */
498 { $$ = newASSIGNOP(OPf_STACKED, $1, $2, $3); }
499 | term POWOP term /* $x ** $y */
500 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
501 | term MULOP term /* $x * $y, $x x $y */
502 { if ($2 != OP_REPEAT)
504 $$ = newBINOP($2, 0, $1, scalar($3)); }
505 | term ADDOP term /* $x + $y */
506 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
507 | term SHIFTOP term /* $x >> $y, $x << $y */
508 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
509 | term RELOP term /* $x > $y, etc. */
510 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
511 | term EQOP term /* $x == $y, $x eq $y */
512 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
513 | term BITANDOP term /* $x & $y */
514 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
515 | term BITOROP term /* $x | $y */
516 { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }
517 | term DOTDOT term /* $x..$y, $x...$y */
518 { $$ = newRANGE($2, scalar($1), scalar($3));}
519 | term ANDAND term /* $x && $y */
520 { $$ = newLOGOP(OP_AND, 0, $1, $3); }
521 | term OROR term /* $x || $y */
522 { $$ = newLOGOP(OP_OR, 0, $1, $3); }
523 | term DORDOR term /* $x // $y */
524 { $$ = newLOGOP(OP_DOR, 0, $1, $3); }
525 | term MATCHOP term /* $x =~ /$y/ */
526 { $$ = bind_match($2, $1, $3); }
529 /* Unary operators and terms */
530 termunop : '-' term %prec UMINUS /* -$x */
531 { $$ = newUNOP(OP_NEGATE, 0, scalar($2)); }
532 | '+' term %prec UMINUS /* +$x */
535 { $$ = newUNOP(OP_NOT, 0, scalar($2)); }
537 { $$ = newUNOP(OP_COMPLEMENT, 0, scalar($2));}
538 | term POSTINC /* $x++ */
539 { $$ = newUNOP(OP_POSTINC, 0,
540 mod(scalar($1), OP_POSTINC)); }
541 | term POSTDEC /* $x-- */
542 { $$ = newUNOP(OP_POSTDEC, 0,
543 mod(scalar($1), OP_POSTDEC)); }
544 | PREINC term /* ++$x */
545 { $$ = newUNOP(OP_PREINC, 0,
546 mod(scalar($2), OP_PREINC)); }
547 | PREDEC term /* --$x */
548 { $$ = newUNOP(OP_PREDEC, 0,
549 mod(scalar($2), OP_PREDEC)); }
553 /* Constructors for anonymous data */
554 anonymous: '[' expr ']'
555 { $$ = newANONLIST($2); }
557 { $$ = newANONLIST(Nullop); }
558 | HASHBRACK expr ';' '}' %prec '(' /* { foo => "Bar" } */
559 { $$ = newANONHASH($2); }
560 | HASHBRACK ';' '}' %prec '(' /* { } (';' by tokener) */
561 { $$ = newANONHASH(Nullop); }
562 | ANONSUB startanonsub proto subattrlist block %prec '('
563 { $$ = newANONATTRSUB($2, $3, $4, $5); }
567 /* Things called with "do" */
568 termdo : DO term %prec UNIOP /* do $filename */
569 { $$ = dofile($2, $1); }
570 | DO block %prec '(' /* do { code */
571 { $$ = newUNOP(OP_NULL, OPf_SPECIAL, scope($2)); }
572 | DO WORD '(' ')' /* do somesub() */
573 { $$ = newUNOP(OP_ENTERSUB,
574 OPf_SPECIAL|OPf_STACKED,
575 prepend_elem(OP_LIST,
577 (OPpENTERSUB_AMPER<<8),
580 | DO WORD '(' expr ')' /* do somesub(@args) */
581 { $$ = newUNOP(OP_ENTERSUB,
582 OPf_SPECIAL|OPf_STACKED,
586 (OPpENTERSUB_AMPER<<8),
589 | DO scalar '(' ')' /* do $subref () */
590 { $$ = newUNOP(OP_ENTERSUB, OPf_SPECIAL|OPf_STACKED,
591 prepend_elem(OP_LIST,
592 scalar(newCVREF(0,scalar($2))), Nullop)); dep();}
593 | DO scalar '(' expr ')' /* do $subref (@args) */
594 { $$ = newUNOP(OP_ENTERSUB, OPf_SPECIAL|OPf_STACKED,
595 prepend_elem(OP_LIST,
597 scalar(newCVREF(0,scalar($2))))); dep();}
605 | term '?' term ':' term
606 { $$ = newCONDOP(0, $1, $3, $5); }
607 | REFGEN term /* \$x, \@y, \%z */
608 { $$ = newUNOP(OP_REFGEN, 0, mod($2,OP_REFGEN)); }
609 | myattrterm %prec UNIOP
611 | LOCAL term %prec UNIOP
612 { $$ = localize($2,$1); }
614 { $$ = sawparens($2); }
616 { $$ = sawparens(newNULLLIST()); }
625 | arylen %prec '(' /* $#x, $#{ something } */
626 { $$ = newUNOP(OP_AV2ARYLEN, 0, ref($1, OP_AV2ARYLEN));}
629 | ary '[' expr ']' /* array slice */
630 { $$ = prepend_elem(OP_ASLICE,
631 newOP(OP_PUSHMARK, 0),
632 newLISTOP(OP_ASLICE, 0,
634 ref($1, OP_ASLICE))); }
635 | ary '{' expr ';' '}' /* @hash{@keys} */
636 { $$ = prepend_elem(OP_HSLICE,
637 newOP(OP_PUSHMARK, 0),
638 newLISTOP(OP_HSLICE, 0,
640 ref(oopsHV($1), OP_HSLICE)));
641 PL_expect = XOPERATOR; }
645 { $$ = newUNOP(OP_ENTERSUB, 0, scalar($1)); }
646 | amper '(' ')' /* &foo() */
647 { $$ = newUNOP(OP_ENTERSUB, OPf_STACKED, scalar($1)); }
648 | amper '(' expr ')' /* &foo(@args) */
649 { $$ = newUNOP(OP_ENTERSUB, OPf_STACKED,
650 append_elem(OP_LIST, $3, scalar($1))); }
651 | NOAMP WORD listexpr /* foo(@args) */
652 { $$ = newUNOP(OP_ENTERSUB, OPf_STACKED,
653 append_elem(OP_LIST, $3, scalar($2))); }
654 | LOOPEX /* loop exiting command (goto, last, dump, etc) */
655 { $$ = newOP($1, OPf_SPECIAL);
656 PL_hints |= HINT_BLOCK_SCOPE; }
658 { $$ = newLOOPEX($1,$2); }
659 | NOTOP argexpr /* not $foo */
660 { $$ = newUNOP(OP_NOT, 0, scalar($2)); }
661 | UNIOP /* Unary op, $_ implied */
662 { $$ = newOP($1, 0); }
663 | UNIOP block /* eval { foo } */
664 { $$ = newUNOP($1, 0, $2); }
665 | UNIOP term /* Unary op */
666 { $$ = newUNOP($1, 0, $2); }
667 | REQUIRE /* require, $_ implied */
668 { $$ = newOP(OP_REQUIRE, $1 ? OPf_SPECIAL : 0); }
669 | REQUIRE term /* require Foo */
670 { $$ = newUNOP(OP_REQUIRE, $1 ? OPf_SPECIAL : 0, $2); }
671 | UNIOPSUB term /* Sub treated as unop */
672 { $$ = newUNOP(OP_ENTERSUB, OPf_STACKED,
673 append_elem(OP_LIST, $2, scalar($1))); }
674 | FUNC0 /* Nullary operator */
675 { $$ = newOP($1, 0); }
677 { $$ = newOP($1, 0); }
678 | FUNC0SUB /* Sub treated as nullop */
679 { $$ = newUNOP(OP_ENTERSUB, OPf_STACKED,
681 | FUNC1 '(' ')' /* not () */
682 { $$ = $1 == OP_NOT ? newUNOP($1, 0, newSVOP(OP_CONST, 0, newSViv(0)))
683 : newOP($1, OPf_SPECIAL); }
684 | FUNC1 '(' expr ')' /* not($foo) */
685 { $$ = newUNOP($1, 0, $3); }
686 | PMFUNC '(' argexpr ')' /* m//, s///, tr/// */
687 { $$ = pmruntime($1, $3, 1); }
692 /* "my" declarations, with optional attributes */
693 myattrterm: MY myterm myattrlist
694 { $$ = my_attrs($2,$3); }
696 { $$ = localize($2,$1); }
699 /* Things that can be "my"'d */
700 myterm : '(' expr ')'
701 { $$ = sawparens($2); }
703 { $$ = sawparens(newNULLLIST()); }
712 /* Basic list expressions */
713 listexpr: /* NULL */ %prec PREC_LOW
715 | argexpr %prec PREC_LOW
719 listexprcom: /* NULL */
727 /* A little bit of trickery to make "for my $foo (@bar)" actually be
730 { PL_in_my = 0; $$ = my($1); }
734 { $$ = newCVREF($1,$2); }
738 { $$ = newSVREF($2); }
742 { $$ = newAVREF($2); }
746 { $$ = newHVREF($2); }
749 arylen : DOLSHARP indirob
750 { $$ = newAVREF($2); }
754 { $$ = newGVREF(0,$2); }
757 /* Indirect objects */
760 | scalar %prec PREC_LOW