1 /* regexp.h 2 * 3 * Copyright (c) 1997-2002, Larry Wall 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 */ 9 10 /* 11 * Definitions etc. for regexp(3) routines. 12 * 13 * Caveat: this is V8 regexp(3) [actually, a reimplementation thereof], 14 * not the System V one. 15 */ 16 17 18 struct regnode { 19 U8 flags; 20 U8 type; 21 U16 next_off; 22 }; 23 24 typedef struct regnode regnode; 25 26 struct reg_substr_data; 27 28 struct reg_data; 29 30 typedef struct regexp { 31 I32 *startp; 32 I32 *endp; 33 regnode *regstclass; 34 struct reg_substr_data *substrs; 35 char *precomp; /* pre-compilation regular expression */ 36 struct reg_data *data; /* Additional data. */ 37 char *subbeg; /* saved or original string 38 so \digit works forever. */ 39 U32 *offsets; /* offset annotations 20001228 MJD */ 40 I32 sublen; /* Length of string pointed by subbeg */ 41 I32 refcnt; 42 I32 minlen; /* mininum possible length of $& */ 43 I32 prelen; /* length of precomp */ 44 U32 nparens; /* number of parentheses */ 45 U32 lastparen; /* last paren matched */ 46 U32 lastcloseparen; /* last paren matched */ 47 U32 reganch; /* Internal use only + 48 Tainted information used by regexec? */ 49 regnode program[1]; /* Unwarranted chumminess with compiler. */ 50 } regexp; 51 52 #define ROPT_ANCH (ROPT_ANCH_BOL|ROPT_ANCH_MBOL|ROPT_ANCH_GPOS|ROPT_ANCH_SBOL) 53 #define ROPT_ANCH_SINGLE (ROPT_ANCH_SBOL|ROPT_ANCH_GPOS) 54 #define ROPT_ANCH_BOL 0x00001 55 #define ROPT_ANCH_MBOL 0x00002 56 #define ROPT_ANCH_SBOL 0x00004 57 #define ROPT_ANCH_GPOS 0x00008 58 #define ROPT_SKIP 0x00010 59 #define ROPT_IMPLICIT 0x00020 /* Converted .* to ^.* */ 60 #define ROPT_NOSCAN 0x00040 /* Check-string always at start. */ 61 #define ROPT_GPOS_SEEN 0x00080 62 #define ROPT_CHECK_ALL 0x00100 63 #define ROPT_LOOKBEHIND_SEEN 0x00200 64 #define ROPT_EVAL_SEEN 0x00400 65 #define ROPT_CANY_SEEN 0x00800 66 #define ROPT_SANY_SEEN ROPT_CANY_SEEN /* src bckwrd cmpt */ 67 68 /* 0xf800 of reganch is used by PMf_COMPILETIME */ 69 70 #define ROPT_UTF8 0x10000 71 #define ROPT_NAUGHTY 0x20000 /* how exponential is this pattern? */ 72 #define ROPT_COPY_DONE 0x40000 /* subbeg is a copy of the string */ 73 #define ROPT_TAINTED_SEEN 0x80000 74 75 #define RE_USE_INTUIT_NOML 0x0100000 /* Best to intuit before matching */ 76 #define RE_USE_INTUIT_ML 0x0200000 77 #define REINT_AUTORITATIVE_NOML 0x0400000 /* Can trust a positive answer */ 78 #define REINT_AUTORITATIVE_ML 0x0800000 79 #define REINT_ONCE_NOML 0x1000000 /* Intuit can succed once only. */ 80 #define REINT_ONCE_ML 0x2000000 81 #define RE_INTUIT_ONECHAR 0x4000000 82 #define RE_INTUIT_TAIL 0x8000000 83 84 #define RE_USE_INTUIT (RE_USE_INTUIT_NOML|RE_USE_INTUIT_ML) 85 #define REINT_AUTORITATIVE (REINT_AUTORITATIVE_NOML|REINT_AUTORITATIVE_ML) 86 #define REINT_ONCE (REINT_ONCE_NOML|REINT_ONCE_ML) 87 88 #define RX_MATCH_TAINTED(prog) ((prog)->reganch & ROPT_TAINTED_SEEN) 89 #define RX_MATCH_TAINTED_on(prog) ((prog)->reganch |= ROPT_TAINTED_SEEN) 90 #define RX_MATCH_TAINTED_off(prog) ((prog)->reganch &= ~ROPT_TAINTED_SEEN) 91 #define RX_MATCH_TAINTED_set(prog, t) ((t) \ 92 ? RX_MATCH_TAINTED_on(prog) \ 93 : RX_MATCH_TAINTED_off(prog)) 94 95 #define RX_MATCH_COPIED(prog) ((prog)->reganch & ROPT_COPY_DONE) 96 #define RX_MATCH_COPIED_on(prog) ((prog)->reganch |= ROPT_COPY_DONE) 97 #define RX_MATCH_COPIED_off(prog) ((prog)->reganch &= ~ROPT_COPY_DONE) 98 #define RX_MATCH_COPIED_set(prog,t) ((t) \ 99 ? RX_MATCH_COPIED_on(prog) \ 100 : RX_MATCH_COPIED_off(prog)) 101 102 #define REXEC_COPY_STR 0x01 /* Need to copy the string. */ 103 #define REXEC_CHECKED 0x02 /* check_substr already checked. */ 104 #define REXEC_SCREAM 0x04 /* use scream table. */ 105 #define REXEC_IGNOREPOS 0x08 /* \G matches at start. */ 106 #define REXEC_NOT_FIRST 0x10 /* This is another iteration of //g. */ 107 #define REXEC_ML 0x20 /* $* was set. */ 108 109 #define ReREFCNT_inc(re) ((void)(re && re->refcnt++), re) 110 #define ReREFCNT_dec(re) CALLREGFREE(aTHX_ re) 111 112 #define FBMcf_TAIL_DOLLAR 1 113 #define FBMcf_TAIL_DOLLARM 2 114 #define FBMcf_TAIL_Z 4 115 #define FBMcf_TAIL_z 8 116 #define FBMcf_TAIL (FBMcf_TAIL_DOLLAR|FBMcf_TAIL_DOLLARM|FBMcf_TAIL_Z|FBMcf_TAIL_z) 117 118 #define FBMrf_MULTILINE 1 119 120 struct re_scream_pos_data_s; 121