1/* parser.h 2 * 3 * Copyright (c) 2006, 2007, 2009, 2010, 2011 Larry Wall and others 4 * 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. 7 * 8 * This file defines the layout of the parser object used by the parser 9 * and lexer (perly.c, toke.c). 10 */ 11 12#define YYEMPTY (-2) 13 14typedef struct { 15 YYSTYPE val; /* semantic value */ 16 short state; 17 I32 savestack_ix; /* size of savestack at this state */ 18 CV *compcv; /* value of PL_compcv when this value was created */ 19#ifdef DEBUGGING 20 const char *name; /* token/rule name for -Dpv */ 21#endif 22} yy_stack_frame; 23 24/* Fields that need to be shared with (i.e., visible to) inner lex- 25 ing scopes. */ 26typedef struct yy_lexshared { 27 struct yy_lexshared *ls_prev; 28 SV *ls_linestr; /* mirrors PL_parser->linestr */ 29 char *ls_bufptr; /* mirrors PL_parser->bufptr */ 30 char *re_eval_start; /* start of "(?{..." text */ 31 SV *re_eval_str; /* "(?{...})" text */ 32} LEXSHARED; 33 34typedef struct yy_parser { 35 36 /* parser state */ 37 38 struct yy_parser *old_parser; /* previous value of PL_parser */ 39 YYSTYPE yylval; /* value of lookahead symbol, set by yylex() */ 40 int yychar; /* The lookahead symbol. */ 41 42 /* Number of tokens to shift before error messages enabled. */ 43 int yyerrstatus; 44 45 yy_stack_frame *stack; /* base of stack */ 46 yy_stack_frame *stack_max1;/* (top-1)th element of allocated stack */ 47 yy_stack_frame *ps; /* current stack frame */ 48 int yylen; /* length of active reduction */ 49 50 /* lexer state */ 51 52 I32 lex_formbrack; /* bracket count at outer format level */ 53 I32 lex_brackets; /* square and curly bracket count */ 54 I32 lex_casemods; /* casemod count */ 55 char *lex_brackstack;/* what kind of brackets to pop */ 56 char *lex_casestack; /* what kind of case mods in effect */ 57 U8 lex_defer; /* state after determined token */ 58 U8 lex_dojoin; /* doing an array interpolation 59 1 = @{...} 2 = ->@ */ 60 U8 expect; /* how to interpret ambiguous tokens */ 61 bool preambled; 62 bool sub_no_recover; /* can't recover from a sublex error */ 63 U8 sub_error_count; /* the number of errors before sublexing */ 64 OP *lex_inpat; /* in pattern $) and $| are special */ 65 OP *lex_op; /* extra info to pass back on op */ 66 SV *lex_repl; /* runtime replacement from s/// */ 67 U16 lex_inwhat; /* what kind of quoting are we in */ 68 OPCODE last_lop_op; /* last named list or unary operator */ 69 I32 lex_starts; /* how many interps done on level */ 70 SV *lex_stuff; /* runtime pattern from m// or s/// */ 71 I32 multi_start; /* 1st line of multi-line string */ 72 I32 multi_end; /* last line of multi-line string */ 73 UV multi_open; /* delimiter code point of said string */ 74 UV multi_close; /* delimiter code point of said string */ 75 bool lex_re_reparsing; /* we're doing G_RE_REPARSING */ 76 U8 lex_super_state;/* lexer state to save */ 77 U16 lex_sub_inwhat; /* "lex_inwhat" to use in sublex_push */ 78 I32 lex_allbrackets;/* (), [], {}, ?: bracket count */ 79 OP *lex_sub_op; /* current op in y/// or pattern */ 80 SV *lex_sub_repl; /* repl of s/// used in sublex_push */ 81 LEXSHARED *lex_shared; 82 SV *linestr; /* current chunk of src text */ 83 char *bufptr; /* carries the cursor (current parsing 84 position) from one invocation of yylex 85 to the next */ 86 char *oldbufptr; /* in yylex, beginning of current token */ 87 char *oldoldbufptr; /* in yylex, beginning of previous token */ 88 char *bufend; 89 char *linestart; /* beginning of most recently read line */ 90 char *last_uni; /* position of last named-unary op */ 91 char *last_lop; /* position of last list operator */ 92 /* copline is used to pass a specific line number to newSTATEOP. It 93 is a one-time line number, as newSTATEOP invalidates it (sets it to 94 NOLINE) after using it. The purpose of this is to report line num- 95 bers in multiline constructs using the number of the first line. */ 96 line_t copline; 97 U16 in_my; /* we're compiling a "my"/"our" declaration */ 98 U8 lex_state; /* next token is determined */ 99 U8 error_count; /* how many compile errors so far, max 10 */ 100 HV *in_my_stash; /* declared class of this "my" declaration */ 101 PerlIO *rsfp; /* current source file pointer */ 102 AV *rsfp_filters; /* holds chain of active source filters */ 103 104 YYSTYPE nextval[5]; /* value of next token, if any */ 105 I32 nexttype[5]; /* type of next token */ 106 U8 nexttoke; 107 U8 form_lex_state; /* remember lex_state when parsing fmt */ 108 U8 lex_fakeeof; /* precedence at which to fake EOF */ 109 U8 lex_flags; 110 COP *saved_curcop; /* the previous PL_curcop */ 111 char tokenbuf[256]; 112 line_t herelines; /* number of lines in here-doc */ 113 line_t preambling; /* line # when processing $ENV{PERL5DB} */ 114 115 /* these are valid while parsing a subroutine signature */ 116 UV sig_elems; /* number of signature elements seen so far */ 117 UV sig_optelems; /* number of optional signature elems seen */ 118 char sig_slurpy; /* the sigil of the slurpy var (or null) */ 119 bool sig_seen; /* the currently parsing sub has a signature */ 120 121 bool recheck_utf8_validity; 122 123 PERL_BITFIELD16 in_pod:1; /* lexer is within a =pod section */ 124 PERL_BITFIELD16 filtered:1; /* source filters in evalbytes */ 125 PERL_BITFIELD16 saw_infix_sigil:1; /* saw & or * or % operator */ 126 PERL_BITFIELD16 parsed_sub:1; /* last thing parsed was a sub */ 127} yy_parser; 128 129/* flags for lexer API */ 130#define LEX_STUFF_UTF8 0x00000001 131#define LEX_KEEP_PREVIOUS 0x00000002 132 133#ifdef PERL_CORE 134# define LEX_START_SAME_FILTER 0x00000001 135# define LEX_IGNORE_UTF8_HINTS 0x00000002 136# define LEX_EVALBYTES 0x00000004 137# define LEX_START_COPIED 0x00000008 138# define LEX_DONT_CLOSE_RSFP 0x00000010 139# define LEX_START_FLAGS \ 140 (LEX_START_SAME_FILTER|LEX_START_COPIED \ 141 |LEX_IGNORE_UTF8_HINTS|LEX_EVALBYTES|LEX_DONT_CLOSE_RSFP) 142#endif 143 144/* flags for parser API */ 145#define PARSE_OPTIONAL 0x00000001 146 147/* values for lex_fakeeof */ 148enum { 149 LEX_FAKEEOF_NEVER, /* don't fake EOF */ 150 LEX_FAKEEOF_CLOSING, /* fake EOF at unmatched closing punctuation */ 151 LEX_FAKEEOF_NONEXPR, /* ... and at token that can't be in expression */ 152 LEX_FAKEEOF_LOWLOGIC, /* ... and at low-precedence logic operator */ 153 LEX_FAKEEOF_COMMA, /* ... and at comma */ 154 LEX_FAKEEOF_ASSIGN, /* ... and at assignment operator */ 155 LEX_FAKEEOF_IFELSE, /* ... and at ?: operator */ 156 LEX_FAKEEOF_RANGE, /* ... and at range operator */ 157 LEX_FAKEEOF_LOGIC, /* ... and at logic operator */ 158 LEX_FAKEEOF_BITWISE, /* ... and at bitwise operator */ 159 LEX_FAKEEOF_COMPARE, /* ... and at comparison operator */ 160 LEX_FAKEEOF_MAX 161}; 162 163/* 164 * ex: set ts=8 sts=4 sw=4 et: 165 */ 166