+++ /dev/null
-#ifndef REGPARSE_H\r
-#define REGPARSE_H\r
-/**********************************************************************\r
- regparse.h - Oniguruma (regular expression library)\r
-**********************************************************************/\r
-/*-\r
- * Copyright (c) 2002-2019 K.Kosako <sndgk393 AT ybb DOT ne DOT jp>\r
- * All rights reserved.\r
- *\r
- * Redistribution and use in source and binary forms, with or without\r
- * modification, are permitted provided that the following conditions\r
- * are met:\r
- * 1. Redistributions of source code must retain the above copyright\r
- * notice, this list of conditions and the following disclaimer.\r
- * 2. Redistributions in binary form must reproduce the above copyright\r
- * notice, this list of conditions and the following disclaimer in the\r
- * documentation and/or other materials provided with the distribution.\r
- *\r
- * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND\r
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE\r
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE\r
- * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE\r
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL\r
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS\r
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)\r
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT\r
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY\r
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF\r
- * SUCH DAMAGE.\r
- */\r
-\r
-#include "regint.h"\r
-\r
-#define NODE_STRING_MARGIN 16\r
-#define NODE_STRING_BUF_SIZE 24 /* sizeof(CClassNode) - sizeof(int)*4 */\r
-#define NODE_BACKREFS_SIZE 6\r
-\r
-/* node type */\r
-typedef enum {\r
- NODE_STRING = 0,\r
- NODE_CCLASS = 1,\r
- NODE_CTYPE = 2,\r
- NODE_BACKREF = 3,\r
- NODE_QUANT = 4,\r
- NODE_BAG = 5,\r
- NODE_ANCHOR = 6,\r
- NODE_LIST = 7,\r
- NODE_ALT = 8,\r
- NODE_CALL = 9,\r
- NODE_GIMMICK = 10\r
-} NodeType;\r
-\r
-enum BagType {\r
- BAG_MEMORY = 0,\r
- BAG_OPTION = 1,\r
- BAG_STOP_BACKTRACK = 2,\r
- BAG_IF_ELSE = 3,\r
-};\r
-\r
-enum GimmickType {\r
- GIMMICK_FAIL = 0,\r
- GIMMICK_SAVE = 1,\r
- GIMMICK_UPDATE_VAR = 2,\r
-#ifdef USE_CALLOUT\r
- GIMMICK_CALLOUT = 3,\r
-#endif\r
-};\r
-\r
-enum BodyEmptyType {\r
- BODY_IS_NOT_EMPTY = 0,\r
- BODY_IS_EMPTY_POSSIBILITY = 1,\r
- BODY_IS_EMPTY_POSSIBILITY_MEM = 2,\r
- BODY_IS_EMPTY_POSSIBILITY_REC = 3\r
-};\r
-\r
-typedef struct {\r
- NodeType node_type;\r
- int status;\r
-\r
- UChar* s;\r
- UChar* end;\r
- unsigned int flag;\r
- int capacity; /* (allocated size - 1) or 0: use buf[] */\r
- UChar buf[NODE_STRING_BUF_SIZE];\r
-} StrNode;\r
-\r
-typedef struct {\r
- NodeType node_type;\r
- int status;\r
-\r
- unsigned int flags;\r
- BitSet bs;\r
- BBuf* mbuf; /* multi-byte info or NULL */\r
-} CClassNode;\r
-\r
-typedef struct {\r
- NodeType node_type;\r
- int status;\r
- struct _Node* body;\r
-\r
- int lower;\r
- int upper;\r
- int greedy;\r
- enum BodyEmptyType emptiness;\r
- struct _Node* head_exact;\r
- struct _Node* next_head_exact;\r
- int is_refered; /* include called node. don't eliminate even if {0} */\r
-} QuantNode;\r
-\r
-typedef struct {\r
- NodeType node_type;\r
- int status;\r
- struct _Node* body;\r
-\r
- enum BagType type;\r
- union {\r
- struct {\r
- int regnum;\r
- AbsAddrType called_addr;\r
- int entry_count;\r
- int called_state;\r
- } m;\r
- struct {\r
- OnigOptionType options;\r
- } o;\r
- struct {\r
- /* body is condition */\r
- struct _Node* Then;\r
- struct _Node* Else;\r
- } te;\r
- };\r
- /* for multiple call reference */\r
- OnigLen min_len; /* min length (byte) */\r
- OnigLen max_len; /* max length (byte) */\r
- int char_len; /* character length */\r
- int opt_count; /* referenced count in optimize_nodes() */\r
-} BagNode;\r
-\r
-#ifdef USE_CALL\r
-\r
-typedef struct {\r
- int offset;\r
- struct _Node* target;\r
-} UnsetAddr;\r
-\r
-typedef struct {\r
- int num;\r
- int alloc;\r
- UnsetAddr* us;\r
-} UnsetAddrList;\r
-\r
-typedef struct {\r
- NodeType node_type;\r
- int status;\r
- struct _Node* body; /* to BagNode : BAG_MEMORY */\r
-\r
- int by_number;\r
- int group_num;\r
- UChar* name;\r
- UChar* name_end;\r
- int entry_count;\r
-} CallNode;\r
-\r
-#endif\r
-\r
-typedef struct {\r
- NodeType node_type;\r
- int status;\r
-\r
- int back_num;\r
- int back_static[NODE_BACKREFS_SIZE];\r
- int* back_dynamic;\r
- int nest_level;\r
-} BackRefNode;\r
-\r
-typedef struct {\r
- NodeType node_type;\r
- int status;\r
- struct _Node* body;\r
-\r
- int type;\r
- int char_len;\r
- int ascii_mode;\r
-} AnchorNode;\r
-\r
-typedef struct {\r
- NodeType node_type;\r
- int status;\r
-\r
- struct _Node* car;\r
- struct _Node* cdr;\r
-} ConsAltNode;\r
-\r
-typedef struct {\r
- NodeType node_type;\r
- int status;\r
-\r
- int ctype;\r
- int not;\r
- OnigOptionType options;\r
- int ascii_mode;\r
-} CtypeNode;\r
-\r
-typedef struct {\r
- NodeType node_type;\r
- int status;\r
-\r
- enum GimmickType type;\r
- int detail_type;\r
- int num;\r
- int id;\r
-} GimmickNode;\r
-\r
-typedef struct _Node {\r
- union {\r
- struct {\r
- NodeType node_type;\r
- int status;\r
- struct _Node* body;\r
- } base;\r
-\r
- StrNode str;\r
- CClassNode cclass;\r
- QuantNode quant;\r
- BagNode bag;\r
- BackRefNode backref;\r
- AnchorNode anchor;\r
- ConsAltNode cons;\r
- CtypeNode ctype;\r
-#ifdef USE_CALL\r
- CallNode call;\r
-#endif\r
- GimmickNode gimmick;\r
- } u;\r
-} Node;\r
-\r
-#define NULL_NODE ((Node* )0)\r
-\r
-\r
-/* node type bit */\r
-#define NODE_TYPE2BIT(type) (1<<(type))\r
-\r
-#define NODE_BIT_STRING NODE_TYPE2BIT(NODE_STRING)\r
-#define NODE_BIT_CCLASS NODE_TYPE2BIT(NODE_CCLASS)\r
-#define NODE_BIT_CTYPE NODE_TYPE2BIT(NODE_CTYPE)\r
-#define NODE_BIT_BACKREF NODE_TYPE2BIT(NODE_BACKREF)\r
-#define NODE_BIT_QUANT NODE_TYPE2BIT(NODE_QUANT)\r
-#define NODE_BIT_BAG NODE_TYPE2BIT(NODE_BAG)\r
-#define NODE_BIT_ANCHOR NODE_TYPE2BIT(NODE_ANCHOR)\r
-#define NODE_BIT_LIST NODE_TYPE2BIT(NODE_LIST)\r
-#define NODE_BIT_ALT NODE_TYPE2BIT(NODE_ALT)\r
-#define NODE_BIT_CALL NODE_TYPE2BIT(NODE_CALL)\r
-#define NODE_BIT_GIMMICK NODE_TYPE2BIT(NODE_GIMMICK)\r
-\r
-#define NODE_TYPE(node) ((node)->u.base.node_type)\r
-#define NODE_SET_TYPE(node, ntype) (node)->u.base.node_type = (ntype)\r
-\r
-#define STR_(node) (&((node)->u.str))\r
-#define CCLASS_(node) (&((node)->u.cclass))\r
-#define CTYPE_(node) (&((node)->u.ctype))\r
-#define BACKREF_(node) (&((node)->u.backref))\r
-#define QUANT_(node) (&((node)->u.quant))\r
-#define BAG_(node) (&((node)->u.bag))\r
-#define ANCHOR_(node) (&((node)->u.anchor))\r
-#define CONS_(node) (&((node)->u.cons))\r
-#define CALL_(node) (&((node)->u.call))\r
-#define GIMMICK_(node) (&((node)->u.gimmick))\r
-\r
-#define NODE_CAR(node) (CONS_(node)->car)\r
-#define NODE_CDR(node) (CONS_(node)->cdr)\r
-\r
-#define CTYPE_ANYCHAR -1\r
-#define NODE_IS_ANYCHAR(node) \\r
- (NODE_TYPE(node) == NODE_CTYPE && CTYPE_(node)->ctype == CTYPE_ANYCHAR)\r
-\r
-#define CTYPE_OPTION(node, reg) \\r
- (NODE_IS_FIXED_OPTION(node) ? CTYPE_(node)->options : reg->options)\r
-\r
-\r
-#define ANCR_ANYCHAR_INF_MASK (ANCR_ANYCHAR_INF | ANCR_ANYCHAR_INF_ML)\r
-#define ANCR_END_BUF_MASK (ANCR_END_BUF | ANCR_SEMI_END_BUF)\r
-\r
-#define NODE_STRING_RAW (1<<0) /* by backslashed number */\r
-#define NODE_STRING_AMBIG (1<<1)\r
-#define NODE_STRING_GOOD_AMBIG (1<<2)\r
-#define NODE_STRING_DONT_GET_OPT_INFO (1<<3)\r
-\r
-#define NODE_STRING_LEN(node) (int )((node)->u.str.end - (node)->u.str.s)\r
-#define NODE_STRING_SET_RAW(node) (node)->u.str.flag |= NODE_STRING_RAW\r
-#define NODE_STRING_CLEAR_RAW(node) (node)->u.str.flag &= ~NODE_STRING_RAW\r
-#define NODE_STRING_SET_AMBIG(node) (node)->u.str.flag |= NODE_STRING_AMBIG\r
-#define NODE_STRING_SET_GOOD_AMBIG(node) (node)->u.str.flag |= NODE_STRING_GOOD_AMBIG\r
-#define NODE_STRING_SET_DONT_GET_OPT_INFO(node) \\r
- (node)->u.str.flag |= NODE_STRING_DONT_GET_OPT_INFO\r
-#define NODE_STRING_IS_RAW(node) \\r
- (((node)->u.str.flag & NODE_STRING_RAW) != 0)\r
-#define NODE_STRING_IS_AMBIG(node) \\r
- (((node)->u.str.flag & NODE_STRING_AMBIG) != 0)\r
-#define NODE_STRING_IS_GOOD_AMBIG(node) \\r
- (((node)->u.str.flag & NODE_STRING_GOOD_AMBIG) != 0)\r
-#define NODE_STRING_IS_DONT_GET_OPT_INFO(node) \\r
- (((node)->u.str.flag & NODE_STRING_DONT_GET_OPT_INFO) != 0)\r
-\r
-#define BACKREFS_P(br) \\r
- (IS_NOT_NULL((br)->back_dynamic) ? (br)->back_dynamic : (br)->back_static)\r
-\r
-/* node status bits */\r
-#define NODE_ST_MIN_FIXED (1<<0)\r
-#define NODE_ST_MAX_FIXED (1<<1)\r
-#define NODE_ST_CLEN_FIXED (1<<2)\r
-#define NODE_ST_MARK1 (1<<3)\r
-#define NODE_ST_MARK2 (1<<4)\r
-#define NODE_ST_STRICT_REAL_REPEAT (1<<5)\r
-#define NODE_ST_RECURSION (1<<6)\r
-#define NODE_ST_CALLED (1<<7)\r
-#define NODE_ST_ADDR_FIXED (1<<8)\r
-#define NODE_ST_NAMED_GROUP (1<<9)\r
-#define NODE_ST_IN_REAL_REPEAT (1<<10) /* STK_REPEAT is nested in stack. */\r
-#define NODE_ST_IN_ZERO_REPEAT (1<<11) /* (....){0} */\r
-#define NODE_ST_IN_MULTI_ENTRY (1<<12)\r
-#define NODE_ST_NEST_LEVEL (1<<13)\r
-#define NODE_ST_BY_NUMBER (1<<14) /* {n,m} */\r
-#define NODE_ST_BY_NAME (1<<15) /* backref by name */\r
-#define NODE_ST_BACKREF (1<<16)\r
-#define NODE_ST_CHECKER (1<<17)\r
-#define NODE_ST_FIXED_OPTION (1<<18)\r
-#define NODE_ST_PROHIBIT_RECURSION (1<<19)\r
-#define NODE_ST_SUPER (1<<20)\r
-\r
-\r
-#define NODE_STATUS(node) (((Node* )node)->u.base.status)\r
-#define NODE_STATUS_ADD(node,f) (NODE_STATUS(node) |= (NODE_ST_ ## f))\r
-#define NODE_STATUS_REMOVE(node,f) (NODE_STATUS(node) &= ~(NODE_ST_ ## f))\r
-\r
-#define NODE_IS_BY_NUMBER(node) ((NODE_STATUS(node) & NODE_ST_BY_NUMBER) != 0)\r
-#define NODE_IS_IN_REAL_REPEAT(node) ((NODE_STATUS(node) & NODE_ST_IN_REAL_REPEAT) != 0)\r
-#define NODE_IS_CALLED(node) ((NODE_STATUS(node) & NODE_ST_CALLED) != 0)\r
-#define NODE_IS_IN_MULTI_ENTRY(node) ((NODE_STATUS(node) & NODE_ST_IN_MULTI_ENTRY) != 0)\r
-#define NODE_IS_RECURSION(node) ((NODE_STATUS(node) & NODE_ST_RECURSION) != 0)\r
-#define NODE_IS_IN_ZERO_REPEAT(node) ((NODE_STATUS(node) & NODE_ST_IN_ZERO_REPEAT) != 0)\r
-#define NODE_IS_NAMED_GROUP(node) ((NODE_STATUS(node) & NODE_ST_NAMED_GROUP) != 0)\r
-#define NODE_IS_ADDR_FIXED(node) ((NODE_STATUS(node) & NODE_ST_ADDR_FIXED) != 0)\r
-#define NODE_IS_CLEN_FIXED(node) ((NODE_STATUS(node) & NODE_ST_CLEN_FIXED) != 0)\r
-#define NODE_IS_MIN_FIXED(node) ((NODE_STATUS(node) & NODE_ST_MIN_FIXED) != 0)\r
-#define NODE_IS_MAX_FIXED(node) ((NODE_STATUS(node) & NODE_ST_MAX_FIXED) != 0)\r
-#define NODE_IS_MARK1(node) ((NODE_STATUS(node) & NODE_ST_MARK1) != 0)\r
-#define NODE_IS_MARK2(node) ((NODE_STATUS(node) & NODE_ST_MARK2) != 0)\r
-#define NODE_IS_NEST_LEVEL(node) ((NODE_STATUS(node) & NODE_ST_NEST_LEVEL) != 0)\r
-#define NODE_IS_BY_NAME(node) ((NODE_STATUS(node) & NODE_ST_BY_NAME) != 0)\r
-#define NODE_IS_BACKREF(node) ((NODE_STATUS(node) & NODE_ST_BACKREF) != 0)\r
-#define NODE_IS_CHECKER(node) ((NODE_STATUS(node) & NODE_ST_CHECKER) != 0)\r
-#define NODE_IS_FIXED_OPTION(node) ((NODE_STATUS(node) & NODE_ST_FIXED_OPTION) != 0)\r
-#define NODE_IS_SUPER(node) ((NODE_STATUS(node) & NODE_ST_SUPER) != 0)\r
-#define NODE_IS_PROHIBIT_RECURSION(node) \\r
- ((NODE_STATUS(node) & NODE_ST_PROHIBIT_RECURSION) != 0)\r
-#define NODE_IS_STRICT_REAL_REPEAT(node) \\r
- ((NODE_STATUS(node) & NODE_ST_STRICT_REAL_REPEAT) != 0)\r
-\r
-#define NODE_BODY(node) ((node)->u.base.body)\r
-#define NODE_QUANT_BODY(node) ((node)->body)\r
-#define NODE_BAG_BODY(node) ((node)->body)\r
-#define NODE_CALL_BODY(node) ((node)->body)\r
-#define NODE_ANCHOR_BODY(node) ((node)->body)\r
-\r
-#define SCANENV_MEMENV_SIZE 8\r
-#define SCANENV_MEMENV(senv) \\r
- (IS_NOT_NULL((senv)->mem_env_dynamic) ? \\r
- (senv)->mem_env_dynamic : (senv)->mem_env_static)\r
-\r
-typedef struct {\r
- Node* node;\r
-#if 0\r
- int in;\r
- int recursion;\r
-#endif\r
-} MemEnv;\r
-\r
-typedef struct {\r
- enum SaveType type;\r
-} SaveItem;\r
-\r
-typedef struct {\r
- OnigOptionType options;\r
- OnigCaseFoldType case_fold_flag;\r
- OnigEncoding enc;\r
- OnigSyntaxType* syntax;\r
- MemStatusType capture_history;\r
- MemStatusType bt_mem_start;\r
- MemStatusType bt_mem_end;\r
- MemStatusType backrefed_mem;\r
- UChar* pattern;\r
- UChar* pattern_end;\r
- UChar* error;\r
- UChar* error_end;\r
- regex_t* reg; /* for reg->names only */\r
- int num_call;\r
-#ifdef USE_CALL\r
- UnsetAddrList* unset_addr_list;\r
- int has_call_zero;\r
-#endif\r
- int num_mem;\r
- int num_named;\r
- int mem_alloc;\r
- MemEnv mem_env_static[SCANENV_MEMENV_SIZE];\r
- MemEnv* mem_env_dynamic;\r
- unsigned int parse_depth;\r
-\r
- int keep_num;\r
- int save_num;\r
- int save_alloc_num;\r
- SaveItem* saves;\r
-} ScanEnv;\r
-\r
-\r
-#define IS_SYNTAX_OP(syn, opm) (((syn)->op & (opm)) != 0)\r
-#define IS_SYNTAX_OP2(syn, opm) (((syn)->op2 & (opm)) != 0)\r
-#define IS_SYNTAX_BV(syn, bvm) (((syn)->behavior & (bvm)) != 0)\r
-\r
-typedef struct {\r
- int new_val;\r
-} GroupNumRemap;\r
-\r
-extern int onig_renumber_name_table P_((regex_t* reg, GroupNumRemap* map));\r
-\r
-extern int onig_strncmp P_((const UChar* s1, const UChar* s2, int n));\r
-extern void onig_strcpy P_((UChar* dest, const UChar* src, const UChar* end));\r
-extern void onig_scan_env_set_error_string P_((ScanEnv* env, int ecode, UChar* arg, UChar* arg_end));\r
-extern int onig_scan_unsigned_number P_((UChar** src, const UChar* end, OnigEncoding enc));\r
-extern void onig_reduce_nested_quantifier P_((Node* pnode, Node* cnode));\r
-extern void onig_node_conv_to_str_node P_((Node* node, int raw));\r
-extern int onig_node_str_cat P_((Node* node, const UChar* s, const UChar* end));\r
-extern int onig_node_str_set P_((Node* node, const UChar* s, const UChar* end));\r
-extern void onig_node_free P_((Node* node));\r
-extern Node* onig_node_new_bag P_((enum BagType type));\r
-extern Node* onig_node_new_anchor P_((int type, int ascii_mode));\r
-extern Node* onig_node_new_str P_((const UChar* s, const UChar* end));\r
-extern Node* onig_node_new_list P_((Node* left, Node* right));\r
-extern Node* onig_node_list_add P_((Node* list, Node* x));\r
-extern Node* onig_node_new_alt P_((Node* left, Node* right));\r
-extern void onig_node_str_clear P_((Node* node));\r
-extern int onig_names_free P_((regex_t* reg));\r
-extern int onig_parse_tree P_((Node** root, const UChar* pattern, const UChar* end, regex_t* reg, ScanEnv* env));\r
-extern int onig_free_shared_cclass_table P_((void));\r
-extern int onig_is_code_in_cc P_((OnigEncoding enc, OnigCodePoint code, CClassNode* cc));\r
-extern OnigLen onig_get_tiny_min_len(Node* node, unsigned int inhibit_node_types, int* invalid_node);\r
-\r
-#ifdef USE_CALLOUT\r
-extern int onig_global_callout_names_free(void);\r
-#endif\r
-\r
-#ifdef ONIG_DEBUG\r
-extern int onig_print_names(FILE*, regex_t*);\r
-#endif\r
-\r
-#endif /* REGPARSE_H */\r