diff options
| author | Yves Orton <demerphq@gmail.com> | 2006-10-06 21:16:01 +0200 |
|---|---|---|
| committer | Rafael Garcia-Suarez <rgarciasuarez@gmail.com> | 2006-10-07 14:30:32 +0000 |
| commit | 81714fb9c03d91d66b66cab6e899e81bf64a2ca7 (patch) | |
| tree | 40861dec0355f417fff2a7ff3c393082960066cc /regnodes.h | |
| parent | f5def3a2a0d8913110936f9f4e13e37835754c28 (diff) | |
| download | perl-81714fb9c03d91d66b66cab6e899e81bf64a2ca7.tar.gz | |
Re: [PATCH] Initial attempt at named captures for perls regexp engine
Message-ID: <9b18b3110610061016x5ddce965u30d9a821f632d450@mail.gmail.com>
p4raw-id: //depot/perl@28957
Diffstat (limited to 'regnodes.h')
| -rw-r--r-- | regnodes.h | 147 |
1 files changed, 81 insertions, 66 deletions
diff --git a/regnodes.h b/regnodes.h index 78db033020..3030e04684 100644 --- a/regnodes.h +++ b/regnodes.h @@ -6,8 +6,8 @@ /* Regops and State definitions */ -#define REGNODE_MAX 68 -#define REGMATCH_STATE_MAX 98 +#define REGNODE_MAX 71 +#define REGMATCH_STATE_MAX 101 #define END 0 /* 0000 End of program. */ #define SUCCEED 1 /* 0x01 Return from a subroutine, basically. */ @@ -76,41 +76,44 @@ #define AHOCORASICKC 64 /* 0x40 Same as AHOCORASICK, but with embedded charclass data */ #define RECURSE 65 /* 0x41 recurse to paren arg1 at (signed) ofs arg2 */ #define SRECURSE 66 /* 0x42 recurse to start of pattern */ -#define OPTIMIZED 67 /* 0x43 Placeholder for dump. */ -#define PSEUDO 68 /* 0x44 Pseudo opcode for internal use. */ +#define NREF 67 /* 0x43 Match some already matched string */ +#define NREFF 68 /* 0x44 Match already matched string, folded */ +#define NREFFL 69 /* 0x45 Match already matched string, folded in loc. */ +#define OPTIMIZED 70 /* 0x46 Placeholder for dump. */ +#define PSEUDO 71 /* 0x47 Pseudo opcode for internal use. */ /* ------------ States ------------- */ -#define TRIE_next 69 /* 0x45 Regmatch state for TRIE */ -#define TRIE_next_fail 70 /* 0x46 Regmatch state for TRIE */ -#define EVAL_AB 71 /* 0x47 Regmatch state for EVAL */ -#define EVAL_AB_fail 72 /* 0x48 Regmatch state for EVAL */ -#define CURLYX_end 73 /* 0x49 Regmatch state for CURLYX */ -#define CURLYX_end_fail 74 /* 0x4a Regmatch state for CURLYX */ -#define WHILEM_A_pre 75 /* 0x4b Regmatch state for WHILEM */ -#define WHILEM_A_pre_fail 76 /* 0x4c Regmatch state for WHILEM */ -#define WHILEM_A_min 77 /* 0x4d Regmatch state for WHILEM */ -#define WHILEM_A_min_fail 78 /* 0x4e Regmatch state for WHILEM */ -#define WHILEM_A_max 79 /* 0x4f Regmatch state for WHILEM */ -#define WHILEM_A_max_fail 80 /* 0x50 Regmatch state for WHILEM */ -#define WHILEM_B_min 81 /* 0x51 Regmatch state for WHILEM */ -#define WHILEM_B_min_fail 82 /* 0x52 Regmatch state for WHILEM */ -#define WHILEM_B_max 83 /* 0x53 Regmatch state for WHILEM */ -#define WHILEM_B_max_fail 84 /* 0x54 Regmatch state for WHILEM */ -#define BRANCH_next 85 /* 0x55 Regmatch state for BRANCH */ -#define BRANCH_next_fail 86 /* 0x56 Regmatch state for BRANCH */ -#define CURLYM_A 87 /* 0x57 Regmatch state for CURLYM */ -#define CURLYM_A_fail 88 /* 0x58 Regmatch state for CURLYM */ -#define CURLYM_B 89 /* 0x59 Regmatch state for CURLYM */ -#define CURLYM_B_fail 90 /* 0x5a Regmatch state for CURLYM */ -#define IFMATCH_A 91 /* 0x5b Regmatch state for IFMATCH */ -#define IFMATCH_A_fail 92 /* 0x5c Regmatch state for IFMATCH */ -#define CURLY_B_min_known 93 /* 0x5d Regmatch state for CURLY */ -#define CURLY_B_min_known_fail 94 /* 0x5e Regmatch state for CURLY */ -#define CURLY_B_min 95 /* 0x5f Regmatch state for CURLY */ -#define CURLY_B_min_fail 96 /* 0x60 Regmatch state for CURLY */ -#define CURLY_B_max 97 /* 0x61 Regmatch state for CURLY */ -#define CURLY_B_max_fail 98 /* 0x62 Regmatch state for CURLY */ +#define TRIE_next 72 /* 0x48 Regmatch state for TRIE */ +#define TRIE_next_fail 73 /* 0x49 Regmatch state for TRIE */ +#define EVAL_AB 74 /* 0x4a Regmatch state for EVAL */ +#define EVAL_AB_fail 75 /* 0x4b Regmatch state for EVAL */ +#define CURLYX_end 76 /* 0x4c Regmatch state for CURLYX */ +#define CURLYX_end_fail 77 /* 0x4d Regmatch state for CURLYX */ +#define WHILEM_A_pre 78 /* 0x4e Regmatch state for WHILEM */ +#define WHILEM_A_pre_fail 79 /* 0x4f Regmatch state for WHILEM */ +#define WHILEM_A_min 80 /* 0x50 Regmatch state for WHILEM */ +#define WHILEM_A_min_fail 81 /* 0x51 Regmatch state for WHILEM */ +#define WHILEM_A_max 82 /* 0x52 Regmatch state for WHILEM */ +#define WHILEM_A_max_fail 83 /* 0x53 Regmatch state for WHILEM */ +#define WHILEM_B_min 84 /* 0x54 Regmatch state for WHILEM */ +#define WHILEM_B_min_fail 85 /* 0x55 Regmatch state for WHILEM */ +#define WHILEM_B_max 86 /* 0x56 Regmatch state for WHILEM */ +#define WHILEM_B_max_fail 87 /* 0x57 Regmatch state for WHILEM */ +#define BRANCH_next 88 /* 0x58 Regmatch state for BRANCH */ +#define BRANCH_next_fail 89 /* 0x59 Regmatch state for BRANCH */ +#define CURLYM_A 90 /* 0x5a Regmatch state for CURLYM */ +#define CURLYM_A_fail 91 /* 0x5b Regmatch state for CURLYM */ +#define CURLYM_B 92 /* 0x5c Regmatch state for CURLYM */ +#define CURLYM_B_fail 93 /* 0x5d Regmatch state for CURLYM */ +#define IFMATCH_A 94 /* 0x5e Regmatch state for IFMATCH */ +#define IFMATCH_A_fail 95 /* 0x5f Regmatch state for IFMATCH */ +#define CURLY_B_min_known 96 /* 0x60 Regmatch state for CURLY */ +#define CURLY_B_min_known_fail 97 /* 0x61 Regmatch state for CURLY */ +#define CURLY_B_min 98 /* 0x62 Regmatch state for CURLY */ +#define CURLY_B_min_fail 99 /* 0x63 Regmatch state for CURLY */ +#define CURLY_B_max 100 /* 0x64 Regmatch state for CURLY */ +#define CURLY_B_max_fail 101 /* 0x65 Regmatch state for CURLY */ /* PL_regkind[] What type of regop or state is this. */ @@ -185,6 +188,9 @@ EXTCONST U8 PL_regkind[] = { TRIE, /* AHOCORASICKC */ RECURSE, /* RECURSE */ RECURSE, /* SRECURSE */ + NREF, /* NREF */ + NREF, /* NREFF */ + NREF, /* NREFFL */ NOTHING, /* OPTIMIZED */ PSEUDO, /* PSEUDO */ /* ------------ States ------------- */ @@ -292,6 +298,9 @@ static const U8 regarglen[] = { EXTRA_SIZE(struct regnode_charclass), /* AHOCORASICKC */ EXTRA_SIZE(struct regnode_2L), /* RECURSE */ 0, /* SRECURSE */ + EXTRA_SIZE(struct regnode_1), /* NREF */ + EXTRA_SIZE(struct regnode_1), /* NREFF */ + EXTRA_SIZE(struct regnode_1), /* NREFFL */ 0, /* OPTIMIZED */ 0, /* PSEUDO */ }; @@ -366,6 +375,9 @@ static const char reg_off_by_arg[] = { 0, /* AHOCORASICKC */ 0, /* RECURSE */ 0, /* SRECURSE */ + 0, /* NREF */ + 0, /* NREFF */ + 0, /* NREFFL */ 0, /* OPTIMIZED */ 0, /* PSEUDO */ }; @@ -441,39 +453,42 @@ const char * reg_name[] = { "AHOCORASICKC", /* 0x40 */ "RECURSE", /* 0x41 */ "SRECURSE", /* 0x42 */ - "OPTIMIZED", /* 0x43 */ - "PSEUDO", /* 0x44 */ + "NREF", /* 0x43 */ + "NREFF", /* 0x44 */ + "NREFFL", /* 0x45 */ + "OPTIMIZED", /* 0x46 */ + "PSEUDO", /* 0x47 */ /* ------------ States ------------- */ - "TRIE_next", /* 0x45 */ - "TRIE_next_fail", /* 0x46 */ - "EVAL_AB", /* 0x47 */ - "EVAL_AB_fail", /* 0x48 */ - "CURLYX_end", /* 0x49 */ - "CURLYX_end_fail", /* 0x4a */ - "WHILEM_A_pre", /* 0x4b */ - "WHILEM_A_pre_fail", /* 0x4c */ - "WHILEM_A_min", /* 0x4d */ - "WHILEM_A_min_fail", /* 0x4e */ - "WHILEM_A_max", /* 0x4f */ - "WHILEM_A_max_fail", /* 0x50 */ - "WHILEM_B_min", /* 0x51 */ - "WHILEM_B_min_fail", /* 0x52 */ - "WHILEM_B_max", /* 0x53 */ - "WHILEM_B_max_fail", /* 0x54 */ - "BRANCH_next", /* 0x55 */ - "BRANCH_next_fail", /* 0x56 */ - "CURLYM_A", /* 0x57 */ - "CURLYM_A_fail", /* 0x58 */ - "CURLYM_B", /* 0x59 */ - "CURLYM_B_fail", /* 0x5a */ - "IFMATCH_A", /* 0x5b */ - "IFMATCH_A_fail", /* 0x5c */ - "CURLY_B_min_known", /* 0x5d */ - "CURLY_B_min_known_fail", /* 0x5e */ - "CURLY_B_min", /* 0x5f */ - "CURLY_B_min_fail", /* 0x60 */ - "CURLY_B_max", /* 0x61 */ - "CURLY_B_max_fail", /* 0x62 */ + "TRIE_next", /* 0x48 */ + "TRIE_next_fail", /* 0x49 */ + "EVAL_AB", /* 0x4a */ + "EVAL_AB_fail", /* 0x4b */ + "CURLYX_end", /* 0x4c */ + "CURLYX_end_fail", /* 0x4d */ + "WHILEM_A_pre", /* 0x4e */ + "WHILEM_A_pre_fail", /* 0x4f */ + "WHILEM_A_min", /* 0x50 */ + "WHILEM_A_min_fail", /* 0x51 */ + "WHILEM_A_max", /* 0x52 */ + "WHILEM_A_max_fail", /* 0x53 */ + "WHILEM_B_min", /* 0x54 */ + "WHILEM_B_min_fail", /* 0x55 */ + "WHILEM_B_max", /* 0x56 */ + "WHILEM_B_max_fail", /* 0x57 */ + "BRANCH_next", /* 0x58 */ + "BRANCH_next_fail", /* 0x59 */ + "CURLYM_A", /* 0x5a */ + "CURLYM_A_fail", /* 0x5b */ + "CURLYM_B", /* 0x5c */ + "CURLYM_B_fail", /* 0x5d */ + "IFMATCH_A", /* 0x5e */ + "IFMATCH_A_fail", /* 0x5f */ + "CURLY_B_min_known", /* 0x60 */ + "CURLY_B_min_known_fail", /* 0x61 */ + "CURLY_B_min", /* 0x62 */ + "CURLY_B_min_fail", /* 0x63 */ + "CURLY_B_max", /* 0x64 */ + "CURLY_B_max_fail", /* 0x65 */ }; #endif /* DEBUGGING */ #else |