regparse.h

   1 #ifndef ONIGURUMA_REGPARSE_H
   2 #define ONIGURUMA_REGPARSE_H
   3 /**********************************************************************
   4   regparse.h -  Oniguruma (regular expression library)
   5 **********************************************************************/
   6 /*-
   7  * Copyright (c) 2002-2007  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>
   8  * All rights reserved.
   9  *
  10  * Redistribution and use in source and binary forms, with or without
  11  * modification, are permitted provided that the following conditions
  12  * are met:
  13  * 1. Redistributions of source code must retain the above copyright
  14  *    notice, this list of conditions and the following disclaimer.
  15  * 2. Redistributions in binary form must reproduce the above copyright
  16  *    notice, this list of conditions and the following disclaimer in the
  17  *    documentation and/or other materials provided with the distribution.
  18  *
  19  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
  20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
  23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  29  * SUCH DAMAGE.
  30  */
  31
  32 #include "regint.h"
  33
  34 /* node type */
  35 #define NT_STR         0
  36 #define NT_CCLASS      1
  37 #define NT_CTYPE       2
  38 #define NT_CANY        3
  39 #define NT_BREF        4
  40 #define NT_QTFR        5
  41 #define NT_ENCLOSE     6
  42 #define NT_ANCHOR      7
  43 #define NT_LIST        8
  44 #define NT_ALT         9
  45 #define NT_CALL       10
  46
  47 /* node type bit */
  48 #define NTYPE2BIT(type)      (1<<(type))
  49
  50 #define BIT_NT_STR        NTYPE2BIT(NT_STR)
  51 #define BIT_NT_CCLASS     NTYPE2BIT(NT_CCLASS)
  52 #define BIT_NT_CTYPE      NTYPE2BIT(NT_CTYPE)
  53 #define BIT_NT_CANY       NTYPE2BIT(NT_CANY)
  54 #define BIT_NT_BREF       NTYPE2BIT(NT_BREF)
  55 #define BIT_NT_QTFR       NTYPE2BIT(NT_QTFR)
  56 #define BIT_NT_ENCLOSE    NTYPE2BIT(NT_ENCLOSE)
  57 #define BIT_NT_ANCHOR     NTYPE2BIT(NT_ANCHOR)
  58 #define BIT_NT_LIST       NTYPE2BIT(NT_LIST)
  59 #define BIT_NT_ALT        NTYPE2BIT(NT_ALT)
  60 #define BIT_NT_CALL       NTYPE2BIT(NT_CALL)
  61
  62 #define IS_NODE_TYPE_SIMPLE(type) \
  63   ((NTYPE2BIT(type) & (BIT_NT_STR | BIT_NT_CCLASS | BIT_NT_CTYPE |\
  64                        BIT_NT_CANY | BIT_NT_BREF)) != 0)
  65
  66 #define NTYPE(node)             ((node)->u.base.type)
  67 #define SET_NTYPE(node, ntype)   (node)->u.base.type = (ntype)
  68
  69 #define NSTR(node)         (&((node)->u.str))
  70 #define NCCLASS(node)      (&((node)->u.cclass))
  71 #define NCTYPE(node)       (&((node)->u.ctype))
  72 #define NBREF(node)        (&((node)->u.bref))
  73 #define NQTFR(node)        (&((node)->u.qtfr))
  74 #define NENCLOSE(node)     (&((node)->u.enclose))
  75 #define NANCHOR(node)      (&((node)->u.anchor))
  76 #define NCONS(node)        (&((node)->u.cons))
  77 #define NCALL(node)        (&((node)->u.call))
  78
  79 #define NCAR(node)         (NCONS(node)->car)
  80 #define NCDR(node)         (NCONS(node)->cdr)
  81
  82
  83
  84 #define ANCHOR_ANYCHAR_STAR_MASK (ANCHOR_ANYCHAR_STAR | ANCHOR_ANYCHAR_STAR_ML)
  85 #define ANCHOR_END_BUF_MASK      (ANCHOR_END_BUF | ANCHOR_SEMI_END_BUF)
  86
  87 #define ENCLOSE_MEMORY           (1<<0)
  88 #define ENCLOSE_OPTION           (1<<1)
  89 #define ENCLOSE_STOP_BACKTRACK   (1<<2)
  90
  91 #define NODE_STR_MARGIN         16
  92 #define NODE_STR_BUF_SIZE       24  /* sizeof(CClassNode) - sizeof(int)*4 */
  93 #define NODE_BACKREFS_SIZE       6
  94
  95 #define NSTR_RAW                (1<<0) /* by backslashed number */
  96 #define NSTR_AMBIG              (1<<1)
  97 #define NSTR_DONT_GET_OPT_INFO  (1<<2)
  98
  99 #define NSTRING_LEN(node)             ((node)->u.str.end - (node)->u.str.s)
 100 #define NSTRING_SET_RAW(node)          (node)->u.str.flag |= NSTR_RAW
 101 #define NSTRING_CLEAR_RAW(node)        (node)->u.str.flag &= ~NSTR_RAW
 102 #define NSTRING_SET_AMBIG(node)        (node)->u.str.flag |= NSTR_AMBIG
 103 #define NSTRING_SET_DONT_GET_OPT_INFO(node) \
 104   (node)->u.str.flag |= NSTR_DONT_GET_OPT_INFO
 105 #define NSTRING_IS_RAW(node)          (((node)->u.str.flag & NSTR_RAW)   != 0)
 106 #define NSTRING_IS_AMBIG(node)        (((node)->u.str.flag & NSTR_AMBIG) != 0)
 107 #define NSTRING_IS_DONT_GET_OPT_INFO(node) \
 108   (((node)->u.str.flag & NSTR_DONT_GET_OPT_INFO) != 0)
 109
 110 #define BACKREFS_P(br) \
 111   (IS_NOT_NULL((br)->back_dynamic) ? (br)->back_dynamic : (br)->back_static);
 112
 113 #define NQ_TARGET_ISNOT_EMPTY     0
 114 #define NQ_TARGET_IS_EMPTY        1
 115 #define NQ_TARGET_IS_EMPTY_MEM    2
 116 #define NQ_TARGET_IS_EMPTY_REC    3
 117
 118 /* status bits */
 119 #define NST_MIN_FIXED             (1<<0)
 120 #define NST_MAX_FIXED             (1<<1)
 121 #define NST_CLEN_FIXED            (1<<2)
 122 #define NST_MARK1                 (1<<3)
 123 #define NST_MARK2                 (1<<4)
 124 #define NST_MEM_BACKREFED         (1<<5)
 125 #define NST_STOP_BT_SIMPLE_REPEAT (1<<6)
 126 #define NST_RECURSION             (1<<7)
 127 #define NST_CALLED                (1<<8)
 128 #define NST_ADDR_FIXED            (1<<9)
 129 #define NST_NAMED_GROUP           (1<<10)
 130 #define NST_NAME_REF              (1<<11)
 131 #define NST_IN_REPEAT             (1<<12) /* STK_REPEAT is nested in stack. */
 132 #define NST_NEST_LEVEL            (1<<13)
 133 #define NST_BY_NUMBER             (1<<14) /* {n,m} */
 134
 135 #define SET_ENCLOSE_STATUS(node,f)      (node)->u.enclose.state |=  (f)
 136 #define CLEAR_ENCLOSE_STATUS(node,f)    (node)->u.enclose.state &= ~(f)
 137
 138 #define IS_ENCLOSE_CALLED(en)          (((en)->state & NST_CALLED)        != 0)
 139 #define IS_ENCLOSE_ADDR_FIXED(en)      (((en)->state & NST_ADDR_FIXED)    != 0)
 140 #define IS_ENCLOSE_RECURSION(en)       (((en)->state & NST_RECURSION)     != 0)
 141 #define IS_ENCLOSE_MARK1(en)           (((en)->state & NST_MARK1)         != 0)
 142 #define IS_ENCLOSE_MARK2(en)           (((en)->state & NST_MARK2)         != 0)
 143 #define IS_ENCLOSE_MIN_FIXED(en)       (((en)->state & NST_MIN_FIXED)     != 0)
 144 #define IS_ENCLOSE_MAX_FIXED(en)       (((en)->state & NST_MAX_FIXED)     != 0)
 145 #define IS_ENCLOSE_CLEN_FIXED(en)      (((en)->state & NST_CLEN_FIXED)    != 0)
 146 #define IS_ENCLOSE_STOP_BT_SIMPLE_REPEAT(en) \
 147     (((en)->state & NST_STOP_BT_SIMPLE_REPEAT) != 0)
 148 #define IS_ENCLOSE_NAMED_GROUP(en)     (((en)->state & NST_NAMED_GROUP)   != 0)
 149
 150 #define SET_CALL_RECURSION(node)       (node)->u.call.state |= NST_RECURSION
 151 #define IS_CALL_RECURSION(cn)          (((cn)->state & NST_RECURSION)  != 0)
 152 #define IS_CALL_NAME_REF(cn)           (((cn)->state & NST_NAME_REF)   != 0)
 153 #define IS_BACKREF_NAME_REF(bn)        (((bn)->state & NST_NAME_REF)   != 0)
 154 #define IS_BACKREF_NEST_LEVEL(bn)      (((bn)->state & NST_NEST_LEVEL) != 0)
 155 #define IS_QUANTIFIER_IN_REPEAT(qn)    (((qn)->state & NST_IN_REPEAT)  != 0)
 156 #define IS_QUANTIFIER_BY_NUMBER(qn)    (((qn)->state & NST_BY_NUMBER)  != 0)
 157
 158 #define CALLNODE_REFNUM_UNDEF  -1
 159
 160 typedef struct {
 161   NodeBase base;
 162   UChar* s;
 163   UChar* end;
 164   unsigned int flag;
 165   int    capa;    /* (allocated size - 1) or 0: use buf[] */
 166   UChar  buf[NODE_STR_BUF_SIZE];
 167 } StrNode;
 168
 169 typedef struct {
 170   NodeBase base;
 171   int state;
 172   struct _Node* target;
 173   int lower;
 174   int upper;
 175   int greedy;
 176   int target_empty_info;
 177   struct _Node* head_exact;
 178   struct _Node* next_head_exact;
 179   int is_refered;     /* include called node. don't eliminate even if {0} */
 180 #ifdef USE_COMBINATION_EXPLOSION_CHECK
 181   int comb_exp_check_num;  /* 1,2,3...: check,  0: no check  */
 182 #endif
 183 } QtfrNode;
 184
 185 typedef struct {
 186   NodeBase base;
 187   int state;
 188   int type;
 189   int regnum;
 190   OnigOptionType option;
 191   struct _Node*  target;
 192   AbsAddrType    call_addr;
 193   /* for multiple call reference */
 194   OnigDistance min_len; /* min length (byte) */
 195   OnigDistance max_len; /* max length (byte) */
 196   int char_len;         /* character length  */
 197   int opt_count;        /* referenced count in optimize_node_left() */
 198 } EncloseNode;
 199
 200 #ifdef USE_SUBEXP_CALL
 201
 202 typedef struct {
 203   int           offset;
 204   struct _Node* target;
 205 } UnsetAddr;
 206
 207 typedef struct {
 208   int        num;
 209   int        alloc;
 210   UnsetAddr* us;
 211 } UnsetAddrList;
 212
 213 typedef struct {
 214   NodeBase base;
 215   int     state;
 216   int     group_num;
 217   UChar*  name;
 218   UChar*  name_end;
 219   struct _Node*  target;  /* EncloseNode : ENCLOSE_MEMORY */
 220   UnsetAddrList* unset_addr_list;
 221 } CallNode;
 222
 223 #endif
 224
 225 typedef struct {
 226   NodeBase base;
 227   int  state;
 228   int  back_num;
 229   int  back_static[NODE_BACKREFS_SIZE];
 230   int* back_dynamic;
 231   int  nest_level;
 232 } BRefNode;
 233
 234 typedef struct {
 235   NodeBase base;
 236   int type;
 237   struct _Node* target;
 238   int char_len;
 239 } AnchorNode;
 240
 241 typedef struct {
 242   NodeBase base;
 243   struct _Node* car;
 244   struct _Node* cdr;
 245 } ConsAltNode;
 246
 247 typedef struct {
 248   NodeBase base;
 249   int ctype;
 250   int not;
 251 } CtypeNode;
 252
 253 typedef struct _Node {
 254   union {
 255     NodeBase     base;
 256     StrNode      str;
 257     CClassNode   cclass;
 258     QtfrNode     qtfr;
 259     EncloseNode  enclose;
 260     BRefNode     bref;
 261     AnchorNode   anchor;
 262     ConsAltNode  cons;
 263     CtypeNode    ctype;
 264 #ifdef USE_SUBEXP_CALL
 265     CallNode     call;
 266 #endif
 267   } u;
 268 } Node;
 269
 270
 271 #define NULL_NODE  ((Node* )0)
 272
 273 #define SCANENV_MEMNODES_SIZE               8
 274 #define SCANENV_MEM_NODES(senv)   \
 275  (IS_NOT_NULL((senv)->mem_nodes_dynamic) ? \
 276     (senv)->mem_nodes_dynamic : (senv)->mem_nodes_static)
 277
 278 typedef struct {
 279   OnigOptionType   option;
 280   OnigCaseFoldType case_fold_flag;
 281   OnigEncoding     enc;
 282   OnigSyntaxType*  syntax;
 283   BitStatusType    capture_history;
 284   BitStatusType    bt_mem_start;
 285   BitStatusType    bt_mem_end;
 286   BitStatusType    backrefed_mem;
 287   UChar*           pattern;
 288   UChar*           pattern_end;
 289   UChar*           error;
 290   UChar*           error_end;
 291   regex_t*         reg;       /* for reg->names only */
 292   int              num_call;
 293 #ifdef USE_SUBEXP_CALL
 294   UnsetAddrList*   unset_addr_list;
 295 #endif
 296   int              num_mem;
 297 #ifdef USE_NAMED_GROUP
 298   int              num_named;
 299 #endif
 300   int              mem_alloc;
 301   Node*            mem_nodes_static[SCANENV_MEMNODES_SIZE];
 302   Node**           mem_nodes_dynamic;
 303 #ifdef USE_COMBINATION_EXPLOSION_CHECK
 304   int num_comb_exp_check;
 305   int comb_exp_max_regnum;
 306   int curr_max_regnum;
 307   int has_recursion;
 308 #endif
 309 } ScanEnv;
 310
 311
 312 #define IS_SYNTAX_OP(syn, opm)    (((syn)->op  & (opm)) != 0)
 313 #define IS_SYNTAX_OP2(syn, opm)   (((syn)->op2 & (opm)) != 0)
 314 #define IS_SYNTAX_BV(syn, bvm)    (((syn)->behavior & (bvm)) != 0)
 315
 316 #ifdef USE_NAMED_GROUP
 317 typedef struct {
 318   int new_val;
 319 } GroupNumRemap;
 320
 321 extern int    onig_renumber_name_table P_((regex_t* reg, GroupNumRemap* map));
 322 #endif
 323
 324 extern int    onig_strncmp P_((const UChar* s1, const UChar* s2, int n));
 325 extern void   onig_strcpy P_((UChar* dest, const UChar* src, const UChar* end));
 326 extern void   onig_scan_env_set_error_string P_((ScanEnv* env, int ecode, UChar* arg, UChar* arg_end));
 327 extern int    onig_scan_unsigned_number P_((UChar** src, const UChar* end, OnigEncoding enc));
 328 extern void   onig_reduce_nested_quantifier P_((Node* pnode, Node* cnode));
 329 extern void   onig_node_conv_to_str_node P_((Node* node, int raw));
 330 extern int    onig_node_str_cat P_((Node* node, const UChar* s, const UChar* end));
 331 extern int    onig_node_str_set P_((Node* node, const UChar* s, const UChar* end));
 332 extern void   onig_node_free P_((Node* node));
 333 extern Node*  onig_node_new_enclose P_((int type));
 334 extern Node*  onig_node_new_anchor P_((int type));
 335 extern Node*  onig_node_new_str P_((const UChar* s, const UChar* end));
 336 extern Node*  onig_node_new_list P_((Node* left, Node* right));
 337 extern Node*  onig_node_list_add P_((Node* list, Node* x));
 338 extern Node*  onig_node_new_alt P_((Node* left, Node* right));
 339 extern void   onig_node_str_clear P_((Node* node));
 340 extern int    onig_free_node_list P_((void));
 341 extern int    onig_names_free P_((regex_t* reg));
 342 extern int    onig_parse_make_tree P_((Node** root, const UChar* pattern, const UChar* end, regex_t* reg, ScanEnv* env));
 343 extern int    onig_free_shared_cclass_table P_((void));
 344
 345 #ifdef ONIG_DEBUG
 346 #ifdef USE_NAMED_GROUP
 347 extern int onig_print_names(FILE*, regex_t*);
 348 #endif
 349 #endif
 350
 351 #endif /* ONIGURUMA_REGPARSE_H */