regparse.h

   1 #ifndef ONIGMO_REGPARSE_H
   2 #define ONIGMO_REGPARSE_H
   3 /**********************************************************************
   4   regparse.h -  Onigmo (Oniguruma-mod) (regular expression library)
   5 **********************************************************************/
   6 /*-
   7  * Copyright (c) 2002-2007  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>
   8  * Copyright (c) 2011-2016  K.Takata  <kentkt AT csc DOT jp>
   9  * All rights reserved.
  10  *
  11  * Redistribution and use in source and binary forms, with or without
  12  * modification, are permitted provided that the following conditions
  13  * are met:
  14  * 1. Redistributions of source code must retain the above copyright
  15  *    notice, this list of conditions and the following disclaimer.
  16  * 2. Redistributions in binary form must reproduce the above copyright
  17  *    notice, this list of conditions and the following disclaimer in the
  18  *    documentation and/or other materials provided with the distribution.
  19  *
  20  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
  21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
  24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  30  * SUCH DAMAGE.
  31  */
  32
  33 #include "regint.h"
  34
  35 RUBY_SYMBOL_EXPORT_BEGIN
  36
  37 /* node type */
  38 #define NT_STR         0
  39 #define NT_CCLASS      1
  40 #define NT_CTYPE       2
  41 #define NT_CANY        3
  42 #define NT_BREF        4
  43 #define NT_QTFR        5
  44 #define NT_ENCLOSE     6
  45 #define NT_ANCHOR      7
  46 #define NT_LIST        8
  47 #define NT_ALT         9
  48 #define NT_CALL       10
  49
  50 /* node type bit */
  51 #define NTYPE2BIT(type)      (1<<(type))
  52
  53 #define BIT_NT_STR        NTYPE2BIT(NT_STR)
  54 #define BIT_NT_CCLASS     NTYPE2BIT(NT_CCLASS)
  55 #define BIT_NT_CTYPE      NTYPE2BIT(NT_CTYPE)
  56 #define BIT_NT_CANY       NTYPE2BIT(NT_CANY)
  57 #define BIT_NT_BREF       NTYPE2BIT(NT_BREF)
  58 #define BIT_NT_QTFR       NTYPE2BIT(NT_QTFR)
  59 #define BIT_NT_ENCLOSE    NTYPE2BIT(NT_ENCLOSE)
  60 #define BIT_NT_ANCHOR     NTYPE2BIT(NT_ANCHOR)
  61 #define BIT_NT_LIST       NTYPE2BIT(NT_LIST)
  62 #define BIT_NT_ALT        NTYPE2BIT(NT_ALT)
  63 #define BIT_NT_CALL       NTYPE2BIT(NT_CALL)
  64
  65 #define IS_NODE_TYPE_SIMPLE(type) \
  66   ((NTYPE2BIT(type) & (BIT_NT_STR | BIT_NT_CCLASS | BIT_NT_CTYPE |\
  67                        BIT_NT_CANY | BIT_NT_BREF)) != 0)
  68
  69 #define NTYPE(node)             ((node)->u.base.type)
  70 #define SET_NTYPE(node, ntype) \
  71     do { \
  72         int value = ntype; \
  73         memcpy(&((node)->u.base.type), &value, sizeof(int)); \
  74     } while (0)
  75
  76 #define NSTR(node)         (&((node)->u.str))
  77 #define NCCLASS(node)      (&((node)->u.cclass))
  78 #define NCTYPE(node)       (&((node)->u.ctype))
  79 #define NBREF(node)        (&((node)->u.bref))
  80 #define NQTFR(node)        (&((node)->u.qtfr))
  81 #define NENCLOSE(node)     (&((node)->u.enclose))
  82 #define NANCHOR(node)      (&((node)->u.anchor))
  83 #define NCONS(node)        (&((node)->u.cons))
  84 #define NCALL(node)        (&((node)->u.call))
  85
  86 #define NCAR(node)         (NCONS(node)->car)
  87 #define NCDR(node)         (NCONS(node)->cdr)
  88
  89
  90
  91 #define ANCHOR_ANYCHAR_STAR_MASK (ANCHOR_ANYCHAR_STAR | ANCHOR_ANYCHAR_STAR_ML)
  92 #define ANCHOR_END_BUF_MASK      (ANCHOR_END_BUF | ANCHOR_SEMI_END_BUF)
  93
  94 #define ENCLOSE_MEMORY           (1<<0)
  95 #define ENCLOSE_OPTION           (1<<1)
  96 #define ENCLOSE_STOP_BACKTRACK   (1<<2)
  97 #define ENCLOSE_CONDITION        (1<<3)
  98 #define ENCLOSE_ABSENT           (1<<4)
  99
 100 #define NODE_STR_MARGIN         16
 101 #define NODE_STR_BUF_SIZE       24  /* sizeof(CClassNode) - sizeof(int)*4 */
 102 #define NODE_BACKREFS_SIZE       6
 103
 104 #define NSTR_RAW                (1<<0) /* by backslashed number */
 105 #define NSTR_AMBIG              (1<<1)
 106 #define NSTR_DONT_GET_OPT_INFO  (1<<2)
 107
 108 #define NSTRING_LEN(node) (OnigDistance )((node)->u.str.end - (node)->u.str.s)
 109 #define NSTRING_SET_RAW(node)          (node)->u.str.flag |= NSTR_RAW
 110 #define NSTRING_CLEAR_RAW(node)        (node)->u.str.flag &= ~NSTR_RAW
 111 #define NSTRING_SET_AMBIG(node)        (node)->u.str.flag |= NSTR_AMBIG
 112 #define NSTRING_SET_DONT_GET_OPT_INFO(node) \
 113   (node)->u.str.flag |= NSTR_DONT_GET_OPT_INFO
 114 #define NSTRING_IS_RAW(node)          (((node)->u.str.flag & NSTR_RAW)   != 0)
 115 #define NSTRING_IS_AMBIG(node)        (((node)->u.str.flag & NSTR_AMBIG) != 0)
 116 #define NSTRING_IS_DONT_GET_OPT_INFO(node) \
 117   (((node)->u.str.flag & NSTR_DONT_GET_OPT_INFO) != 0)
 118
 119 #define BACKREFS_P(br) \
 120   (IS_NOT_NULL((br)->back_dynamic) ? (br)->back_dynamic : (br)->back_static);
 121
 122 #define NQ_TARGET_ISNOT_EMPTY     0
 123 #define NQ_TARGET_IS_EMPTY        1
 124 #define NQ_TARGET_IS_EMPTY_MEM    2
 125 #define NQ_TARGET_IS_EMPTY_REC    3
 126
 127 /* status bits */
 128 #define NST_MIN_FIXED             (1<<0)
 129 #define NST_MAX_FIXED             (1<<1)
 130 #define NST_CLEN_FIXED            (1<<2)
 131 #define NST_MARK1                 (1<<3)
 132 #define NST_MARK2                 (1<<4)
 133 #define NST_MEM_BACKREFED         (1<<5)
 134 #define NST_STOP_BT_SIMPLE_REPEAT (1<<6)
 135 #define NST_RECURSION             (1<<7)
 136 #define NST_CALLED                (1<<8)
 137 #define NST_ADDR_FIXED            (1<<9)
 138 #define NST_NAMED_GROUP           (1<<10)
 139 #define NST_NAME_REF              (1<<11)
 140 #define NST_IN_REPEAT             (1<<12) /* STK_REPEAT is nested in stack. */
 141 #define NST_NEST_LEVEL            (1<<13)
 142 #define NST_BY_NUMBER             (1<<14) /* {n,m} */
 143
 144 #define SET_ENCLOSE_STATUS(node,f)      (node)->u.enclose.state |=  (f)
 145 #define CLEAR_ENCLOSE_STATUS(node,f)    (node)->u.enclose.state &= ~(f)
 146
 147 #define IS_ENCLOSE_CALLED(en)          (((en)->state & NST_CALLED)        != 0)
 148 #define IS_ENCLOSE_ADDR_FIXED(en)      (((en)->state & NST_ADDR_FIXED)    != 0)
 149 #define IS_ENCLOSE_RECURSION(en)       (((en)->state & NST_RECURSION)     != 0)
 150 #define IS_ENCLOSE_MARK1(en)           (((en)->state & NST_MARK1)         != 0)
 151 #define IS_ENCLOSE_MARK2(en)           (((en)->state & NST_MARK2)         != 0)
 152 #define IS_ENCLOSE_MIN_FIXED(en)       (((en)->state & NST_MIN_FIXED)     != 0)
 153 #define IS_ENCLOSE_MAX_FIXED(en)       (((en)->state & NST_MAX_FIXED)     != 0)
 154 #define IS_ENCLOSE_CLEN_FIXED(en)      (((en)->state & NST_CLEN_FIXED)    != 0)
 155 #define IS_ENCLOSE_STOP_BT_SIMPLE_REPEAT(en) \
 156     (((en)->state & NST_STOP_BT_SIMPLE_REPEAT) != 0)
 157 #define IS_ENCLOSE_NAMED_GROUP(en)     (((en)->state & NST_NAMED_GROUP)   != 0)
 158 #define IS_ENCLOSE_NAME_REF(en)        (((en)->state & NST_NAME_REF)      != 0)
 159
 160 #define SET_CALL_RECURSION(node)       (node)->u.call.state |= NST_RECURSION
 161 #define IS_CALL_RECURSION(cn)          (((cn)->state & NST_RECURSION)  != 0)
 162 #define IS_CALL_NAME_REF(cn)           (((cn)->state & NST_NAME_REF)   != 0)
 163 #define IS_BACKREF_NAME_REF(bn)        (((bn)->state & NST_NAME_REF)   != 0)
 164 #define IS_BACKREF_NEST_LEVEL(bn)      (((bn)->state & NST_NEST_LEVEL) != 0)
 165 #define IS_QUANTIFIER_IN_REPEAT(qn)    (((qn)->state & NST_IN_REPEAT)  != 0)
 166 #define IS_QUANTIFIER_BY_NUMBER(qn)    (((qn)->state & NST_BY_NUMBER)  != 0)
 167
 168 #define CALLNODE_REFNUM_UNDEF  -1
 169
 170 typedef struct {
 171   NodeBase base;
 172   UChar* s;
 173   UChar* end;
 174   unsigned int flag;
 175   int    capa;    /* (allocated size - 1) or 0: use buf[] */
 176   UChar  buf[NODE_STR_BUF_SIZE];
 177 } StrNode;
 178
 179 typedef struct {
 180   NodeBase base;
 181   int state;
 182   struct _Node* target;
 183   int lower;
 184   int upper;
 185   int greedy;
 186   int target_empty_info;
 187   struct _Node* head_exact;
 188   struct _Node* next_head_exact;
 189   int is_referred;     /* include called node. don't eliminate even if {0} */
 190 #ifdef USE_COMBINATION_EXPLOSION_CHECK
 191   int comb_exp_check_num;  /* 1,2,3...: check,  0: no check  */
 192 #endif
 193 } QtfrNode;
 194
 195 typedef struct {
 196   NodeBase base;
 197   int state;
 198   int type;
 199   int regnum;
 200   OnigOptionType option;
 201   AbsAddrType    call_addr;
 202   struct _Node*  target;
 203   /* for multiple call reference */
 204   OnigDistance min_len; /* min length (byte) */
 205   OnigDistance max_len; /* max length (byte) */
 206   int char_len;         /* character length  */
 207   int opt_count;        /* referenced count in optimize_node_left() */
 208 } EncloseNode;
 209
 210 #ifdef USE_SUBEXP_CALL
 211
 212 typedef struct {
 213   int           offset;
 214   struct _Node* target;
 215 } UnsetAddr;
 216
 217 typedef struct {
 218   int        num;
 219   int        alloc;
 220   UnsetAddr* us;
 221 } UnsetAddrList;
 222
 223 typedef struct {
 224   NodeBase base;
 225   int     state;
 226   int     group_num;
 227   UChar*  name;
 228   UChar*  name_end;
 229   struct _Node*  target;  /* EncloseNode : ENCLOSE_MEMORY */
 230   UnsetAddrList* unset_addr_list;
 231 } CallNode;
 232
 233 #endif
 234
 235 typedef struct {
 236   NodeBase base;
 237   int  state;
 238   int  back_num;
 239   int  back_static[NODE_BACKREFS_SIZE];
 240   int* back_dynamic;
 241   int  nest_level;
 242 } BRefNode;
 243
 244 typedef struct {
 245   NodeBase base;
 246   int type;
 247   struct _Node* target;
 248   int char_len;
 249   int ascii_range;
 250 } AnchorNode;
 251
 252 typedef struct {
 253   NodeBase base;
 254   struct _Node* car;
 255   struct _Node* cdr;
 256 } ConsAltNode;
 257
 258 typedef struct {
 259   NodeBase base;
 260   int ctype;
 261   int not;
 262   int ascii_range;
 263 } CtypeNode;
 264
 265 typedef struct _Node {
 266   union {
 267     NodeBase     base;
 268     StrNode      str;
 269     CClassNode   cclass;
 270     QtfrNode     qtfr;
 271     EncloseNode  enclose;
 272     BRefNode     bref;
 273     AnchorNode   anchor;
 274     ConsAltNode  cons;
 275     CtypeNode    ctype;
 276 #ifdef USE_SUBEXP_CALL
 277     CallNode     call;
 278 #endif
 279   } u;
 280 } Node;
 281
 282
 283 #define NULL_NODE  ((Node* )0)
 284
 285 #define SCANENV_MEMNODES_SIZE               8
 286 #define SCANENV_MEM_NODES(senv)   \
 287  (IS_NOT_NULL((senv)->mem_nodes_dynamic) ? \
 288     (senv)->mem_nodes_dynamic : (senv)->mem_nodes_static)
 289
 290 typedef struct {
 291   OnigOptionType   option;
 292   OnigCaseFoldType case_fold_flag;
 293   OnigEncoding     enc;
 294   const OnigSyntaxType* syntax;
 295   BitStatusType    capture_history;
 296   BitStatusType    bt_mem_start;
 297   BitStatusType    bt_mem_end;
 298   BitStatusType    backrefed_mem;
 299   UChar*           pattern;
 300   UChar*           pattern_end;
 301   UChar*           error;
 302   UChar*           error_end;
 303   regex_t*         reg;       /* for reg->names only */
 304 #ifdef USE_SUBEXP_CALL
 305   UnsetAddrList*   unset_addr_list;
 306 #endif
 307   int              num_call;
 308   int              num_mem;
 309 #ifdef USE_NAMED_GROUP
 310   int              num_named;
 311 #endif
 312   int              mem_alloc;
 313   Node*            mem_nodes_static[SCANENV_MEMNODES_SIZE];
 314   Node**           mem_nodes_dynamic;
 315 #ifdef USE_COMBINATION_EXPLOSION_CHECK
 316   int num_comb_exp_check;
 317   int comb_exp_max_regnum;
 318   int curr_max_regnum;
 319   int has_recursion;
 320 #endif
 321   unsigned int parse_depth;
 322   int warnings_flag;
 323 #ifdef RUBY
 324   const char* sourcefile;
 325   int sourceline;
 326 #endif
 327 } ScanEnv;
 328
 329
 330 #define IS_SYNTAX_OP(syn, opm)    (((syn)->op  & (opm)) != 0)
 331 #define IS_SYNTAX_OP2(syn, opm)   (((syn)->op2 & (opm)) != 0)
 332 #define IS_SYNTAX_BV(syn, bvm)    (((syn)->behavior & (bvm)) != 0)
 333
 334 #ifdef USE_NAMED_GROUP
 335 typedef struct {
 336   int new_val;
 337 } GroupNumRemap;
 338
 339 extern int    onig_renumber_name_table(regex_t* reg, GroupNumRemap* map);
 340 #endif
 341
 342 extern int    onig_strncmp(const UChar* s1, const UChar* s2, int n);
 343 extern void   onig_strcpy(UChar* dest, const UChar* src, const UChar* end);
 344 extern void   onig_scan_env_set_error_string(ScanEnv* env, int ecode, UChar* arg, UChar* arg_end);
 345 extern int    onig_scan_unsigned_number(UChar** src, const UChar* end, OnigEncoding enc);
 346 extern void   onig_reduce_nested_quantifier(Node* pnode, Node* cnode);
 347 extern void   onig_node_conv_to_str_node(Node* node, int raw);
 348 extern int    onig_node_str_cat(Node* node, const UChar* s, const UChar* end);
 349 extern int    onig_node_str_set(Node* node, const UChar* s, const UChar* end);
 350 extern void   onig_node_free(Node* node);
 351 extern Node*  onig_node_new_enclose(int type);
 352 extern Node*  onig_node_new_anchor(int type);
 353 extern Node*  onig_node_new_str(const UChar* s, const UChar* end);
 354 extern Node*  onig_node_new_list(Node* left, Node* right);
 355 extern Node*  onig_node_list_add(Node* list, Node* x);
 356 extern Node*  onig_node_new_alt(Node* left, Node* right);
 357 extern void   onig_node_str_clear(Node* node);
 358 extern int    onig_names_free(regex_t* reg);
 359 extern int    onig_parse_make_tree(Node** root, const UChar* pattern, const UChar* end, regex_t* reg, ScanEnv* env);
 360 extern int    onig_free_shared_cclass_table(void);
 361
 362 #ifdef ONIG_DEBUG
 363 # ifdef USE_NAMED_GROUP
 364 extern int onig_print_names(FILE*, regex_t*);
 365 # endif
 366 #endif
 367
 368 RUBY_SYMBOL_EXPORT_END
 369
 370 #endif /* ONIGMO_REGPARSE_H */