]>
Commit | Line | Data |
---|---|---|
14b0e578 CS |
1 | #ifndef REGPARSE_H\r |
2 | #define REGPARSE_H\r | |
3 | /**********************************************************************\r | |
4 | regparse.h - Oniguruma (regular expression library)\r | |
5 | **********************************************************************/\r | |
6 | /*-\r | |
7 | * Copyright (c) 2002-2007 K.Kosako <sndgk393 AT ybb DOT ne DOT jp>\r | |
8 | * All rights reserved.\r | |
9 | * \r | |
0af8e57c | 10 | * (C) Copyright 2015 Hewlett Packard Enterprise Development LP<BR>\r |
14b0e578 CS |
11 | *\r |
12 | * Redistribution and use in source and binary forms, with or without\r | |
13 | * modification, are permitted provided that the following conditions\r | |
14 | * are met:\r | |
15 | * 1. Redistributions of source code must retain the above copyright\r | |
16 | * notice, this list of conditions and the following disclaimer.\r | |
17 | * 2. Redistributions in binary form must reproduce the above copyright\r | |
18 | * notice, this list of conditions and the following disclaimer in the\r | |
19 | * documentation and/or other materials provided with the distribution.\r | |
20 | *\r | |
21 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND\r | |
22 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE\r | |
23 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE\r | |
24 | * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE\r | |
25 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL\r | |
26 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS\r | |
27 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)\r | |
28 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT\r | |
29 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY\r | |
30 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF\r | |
31 | * SUCH DAMAGE.\r | |
32 | */\r | |
33 | \r | |
34 | #include "regint.h"\r | |
35 | \r | |
36 | /* node type */\r | |
37 | #define NT_STR 0\r | |
38 | #define NT_CCLASS 1\r | |
39 | #define NT_CTYPE 2\r | |
40 | #define NT_CANY 3\r | |
41 | #define NT_BREF 4\r | |
42 | #define NT_QTFR 5\r | |
43 | #define NT_ENCLOSE 6\r | |
44 | #define NT_ANCHOR 7\r | |
45 | #define NT_LIST 8\r | |
46 | #define NT_ALT 9\r | |
47 | #define NT_CALL 10\r | |
48 | \r | |
49 | /* node type bit */\r | |
50 | #define NTYPE2BIT(type) (1<<(type))\r | |
51 | \r | |
52 | #define BIT_NT_STR NTYPE2BIT(NT_STR)\r | |
53 | #define BIT_NT_CCLASS NTYPE2BIT(NT_CCLASS)\r | |
54 | #define BIT_NT_CTYPE NTYPE2BIT(NT_CTYPE)\r | |
55 | #define BIT_NT_CANY NTYPE2BIT(NT_CANY)\r | |
56 | #define BIT_NT_BREF NTYPE2BIT(NT_BREF)\r | |
57 | #define BIT_NT_QTFR NTYPE2BIT(NT_QTFR)\r | |
58 | #define BIT_NT_ENCLOSE NTYPE2BIT(NT_ENCLOSE)\r | |
59 | #define BIT_NT_ANCHOR NTYPE2BIT(NT_ANCHOR)\r | |
60 | #define BIT_NT_LIST NTYPE2BIT(NT_LIST)\r | |
61 | #define BIT_NT_ALT NTYPE2BIT(NT_ALT)\r | |
62 | #define BIT_NT_CALL NTYPE2BIT(NT_CALL)\r | |
63 | \r | |
64 | #define IS_NODE_TYPE_SIMPLE(type) \\r | |
65 | ((NTYPE2BIT(type) & (BIT_NT_STR | BIT_NT_CCLASS | BIT_NT_CTYPE |\\r | |
66 | BIT_NT_CANY | BIT_NT_BREF)) != 0)\r | |
67 | \r | |
68 | #define NTYPE(node) ((node)->u.base.type)\r | |
69 | #define SET_NTYPE(node, ntype) (node)->u.base.type = (ntype)\r | |
70 | \r | |
71 | #define NSTR(node) (&((node)->u.str))\r | |
72 | #define NCCLASS(node) (&((node)->u.cclass))\r | |
73 | #define NCTYPE(node) (&((node)->u.ctype))\r | |
74 | #define NBREF(node) (&((node)->u.bref))\r | |
75 | #define NQTFR(node) (&((node)->u.qtfr))\r | |
76 | #define NENCLOSE(node) (&((node)->u.enclose))\r | |
77 | #define NANCHOR(node) (&((node)->u.anchor))\r | |
78 | #define NCONS(node) (&((node)->u.cons))\r | |
79 | #define NCALL(node) (&((node)->u.call))\r | |
80 | \r | |
81 | #define NCAR(node) (NCONS(node)->car)\r | |
82 | #define NCDR(node) (NCONS(node)->cdr)\r | |
83 | \r | |
84 | \r | |
85 | \r | |
86 | #define ANCHOR_ANYCHAR_STAR_MASK (ANCHOR_ANYCHAR_STAR | ANCHOR_ANYCHAR_STAR_ML)\r | |
87 | #define ANCHOR_END_BUF_MASK (ANCHOR_END_BUF | ANCHOR_SEMI_END_BUF)\r | |
88 | \r | |
89 | #define ENCLOSE_MEMORY (1<<0)\r | |
90 | #define ENCLOSE_OPTION (1<<1)\r | |
91 | #define ENCLOSE_STOP_BACKTRACK (1<<2)\r | |
92 | \r | |
93 | #define NODE_STR_MARGIN 16\r | |
94 | #define NODE_STR_BUF_SIZE 24 /* sizeof(CClassNode) - sizeof(int)*4 */\r | |
95 | #define NODE_BACKREFS_SIZE 6\r | |
96 | \r | |
97 | #define NSTR_RAW (1<<0) /* by backslashed number */\r | |
98 | #define NSTR_AMBIG (1<<1)\r | |
99 | #define NSTR_DONT_GET_OPT_INFO (1<<2)\r | |
100 | \r | |
101 | #define NSTRING_LEN(node) ((int)((node)->u.str.end - (node)->u.str.s))\r | |
102 | #define NSTRING_SET_RAW(node) (node)->u.str.flag |= NSTR_RAW\r | |
103 | #define NSTRING_CLEAR_RAW(node) (node)->u.str.flag &= ~NSTR_RAW\r | |
104 | #define NSTRING_SET_AMBIG(node) (node)->u.str.flag |= NSTR_AMBIG\r | |
105 | #define NSTRING_SET_DONT_GET_OPT_INFO(node) \\r | |
106 | (node)->u.str.flag |= NSTR_DONT_GET_OPT_INFO\r | |
107 | #define NSTRING_IS_RAW(node) (((node)->u.str.flag & NSTR_RAW) != 0)\r | |
108 | #define NSTRING_IS_AMBIG(node) (((node)->u.str.flag & NSTR_AMBIG) != 0)\r | |
109 | #define NSTRING_IS_DONT_GET_OPT_INFO(node) \\r | |
110 | (((node)->u.str.flag & NSTR_DONT_GET_OPT_INFO) != 0)\r | |
111 | \r | |
112 | #define BACKREFS_P(br) \\r | |
113 | (IS_NOT_NULL((br)->back_dynamic) ? (br)->back_dynamic : (br)->back_static);\r | |
114 | \r | |
115 | #define NQ_TARGET_ISNOT_EMPTY 0\r | |
116 | #define NQ_TARGET_IS_EMPTY 1\r | |
117 | #define NQ_TARGET_IS_EMPTY_MEM 2\r | |
118 | #define NQ_TARGET_IS_EMPTY_REC 3\r | |
119 | \r | |
120 | /* status bits */\r | |
121 | #define NST_MIN_FIXED (1<<0)\r | |
122 | #define NST_MAX_FIXED (1<<1)\r | |
123 | #define NST_CLEN_FIXED (1<<2)\r | |
124 | #define NST_MARK1 (1<<3)\r | |
125 | #define NST_MARK2 (1<<4)\r | |
126 | #define NST_MEM_BACKREFED (1<<5)\r | |
127 | #define NST_STOP_BT_SIMPLE_REPEAT (1<<6)\r | |
128 | #define NST_RECURSION (1<<7)\r | |
129 | #define NST_CALLED (1<<8)\r | |
130 | #define NST_ADDR_FIXED (1<<9)\r | |
131 | #define NST_NAMED_GROUP (1<<10)\r | |
132 | #define NST_NAME_REF (1<<11)\r | |
133 | #define NST_IN_REPEAT (1<<12) /* STK_REPEAT is nested in stack. */\r | |
134 | #define NST_NEST_LEVEL (1<<13)\r | |
135 | #define NST_BY_NUMBER (1<<14) /* {n,m} */\r | |
136 | \r | |
137 | #define SET_ENCLOSE_STATUS(node,f) (node)->u.enclose.state |= (f)\r | |
138 | #define CLEAR_ENCLOSE_STATUS(node,f) (node)->u.enclose.state &= ~(f)\r | |
139 | \r | |
140 | #define IS_ENCLOSE_CALLED(en) (((en)->state & NST_CALLED) != 0)\r | |
141 | #define IS_ENCLOSE_ADDR_FIXED(en) (((en)->state & NST_ADDR_FIXED) != 0)\r | |
142 | #define IS_ENCLOSE_RECURSION(en) (((en)->state & NST_RECURSION) != 0)\r | |
143 | #define IS_ENCLOSE_MARK1(en) (((en)->state & NST_MARK1) != 0)\r | |
144 | #define IS_ENCLOSE_MARK2(en) (((en)->state & NST_MARK2) != 0)\r | |
145 | #define IS_ENCLOSE_MIN_FIXED(en) (((en)->state & NST_MIN_FIXED) != 0)\r | |
146 | #define IS_ENCLOSE_MAX_FIXED(en) (((en)->state & NST_MAX_FIXED) != 0)\r | |
147 | #define IS_ENCLOSE_CLEN_FIXED(en) (((en)->state & NST_CLEN_FIXED) != 0)\r | |
148 | #define IS_ENCLOSE_STOP_BT_SIMPLE_REPEAT(en) \\r | |
149 | (((en)->state & NST_STOP_BT_SIMPLE_REPEAT) != 0)\r | |
150 | #define IS_ENCLOSE_NAMED_GROUP(en) (((en)->state & NST_NAMED_GROUP) != 0)\r | |
151 | \r | |
152 | #define SET_CALL_RECURSION(node) (node)->u.call.state |= NST_RECURSION\r | |
153 | #define IS_CALL_RECURSION(cn) (((cn)->state & NST_RECURSION) != 0)\r | |
154 | #define IS_CALL_NAME_REF(cn) (((cn)->state & NST_NAME_REF) != 0)\r | |
155 | #define IS_BACKREF_NAME_REF(bn) (((bn)->state & NST_NAME_REF) != 0)\r | |
156 | #define IS_BACKREF_NEST_LEVEL(bn) (((bn)->state & NST_NEST_LEVEL) != 0)\r | |
157 | #define IS_QUANTIFIER_IN_REPEAT(qn) (((qn)->state & NST_IN_REPEAT) != 0)\r | |
158 | #define IS_QUANTIFIER_BY_NUMBER(qn) (((qn)->state & NST_BY_NUMBER) != 0)\r | |
159 | \r | |
160 | #define CALLNODE_REFNUM_UNDEF -1\r | |
161 | \r | |
162 | typedef struct {\r | |
163 | NodeBase base;\r | |
164 | UChar* s;\r | |
165 | UChar* end;\r | |
166 | unsigned int flag;\r | |
167 | int capa; /* (allocated size - 1) or 0: use buf[] */\r | |
168 | UChar buf[NODE_STR_BUF_SIZE];\r | |
169 | } StrNode;\r | |
170 | \r | |
171 | typedef struct {\r | |
172 | NodeBase base;\r | |
173 | int state;\r | |
174 | struct _Node* target;\r | |
175 | int lower;\r | |
176 | int upper;\r | |
177 | int greedy;\r | |
178 | int target_empty_info;\r | |
179 | struct _Node* head_exact;\r | |
180 | struct _Node* next_head_exact;\r | |
181 | int is_refered; /* include called node. don't eliminate even if {0} */\r | |
182 | #ifdef USE_COMBINATION_EXPLOSION_CHECK\r | |
183 | int comb_exp_check_num; /* 1,2,3...: check, 0: no check */\r | |
184 | #endif\r | |
185 | } QtfrNode;\r | |
186 | \r | |
187 | typedef struct {\r | |
188 | NodeBase base;\r | |
189 | int state;\r | |
190 | int type;\r | |
191 | int regnum;\r | |
192 | OnigOptionType option;\r | |
193 | struct _Node* target;\r | |
194 | AbsAddrType call_addr;\r | |
195 | /* for multiple call reference */\r | |
196 | OnigDistance min_len; /* min length (byte) */\r | |
197 | OnigDistance max_len; /* max length (byte) */ \r | |
198 | int char_len; /* character length */\r | |
199 | int opt_count; /* referenced count in optimize_node_left() */\r | |
200 | } EncloseNode;\r | |
201 | \r | |
202 | #ifdef USE_SUBEXP_CALL\r | |
203 | \r | |
204 | typedef struct {\r | |
205 | int offset;\r | |
206 | struct _Node* target;\r | |
207 | } UnsetAddr;\r | |
208 | \r | |
209 | typedef struct {\r | |
210 | int num;\r | |
211 | int alloc;\r | |
212 | UnsetAddr* us;\r | |
213 | } UnsetAddrList;\r | |
214 | \r | |
215 | typedef struct {\r | |
216 | NodeBase base;\r | |
217 | int state;\r | |
218 | int group_num;\r | |
219 | UChar* name;\r | |
220 | UChar* name_end;\r | |
221 | struct _Node* target; /* EncloseNode : ENCLOSE_MEMORY */\r | |
222 | UnsetAddrList* unset_addr_list;\r | |
223 | } CallNode;\r | |
224 | \r | |
225 | #endif\r | |
226 | \r | |
227 | typedef struct {\r | |
228 | NodeBase base;\r | |
229 | int state;\r | |
230 | int back_num;\r | |
231 | int back_static[NODE_BACKREFS_SIZE];\r | |
232 | int* back_dynamic;\r | |
233 | int nest_level;\r | |
234 | } BRefNode;\r | |
235 | \r | |
236 | typedef struct {\r | |
237 | NodeBase base;\r | |
238 | int type;\r | |
239 | struct _Node* target;\r | |
240 | int char_len;\r | |
241 | } AnchorNode;\r | |
242 | \r | |
243 | typedef struct {\r | |
244 | NodeBase base;\r | |
245 | struct _Node* car;\r | |
246 | struct _Node* cdr;\r | |
247 | } ConsAltNode;\r | |
248 | \r | |
249 | typedef struct {\r | |
250 | NodeBase base;\r | |
251 | int ctype;\r | |
252 | int not;\r | |
253 | } CtypeNode;\r | |
254 | \r | |
255 | typedef struct _Node {\r | |
256 | union {\r | |
257 | NodeBase base;\r | |
258 | StrNode str;\r | |
259 | CClassNode cclass;\r | |
260 | QtfrNode qtfr;\r | |
261 | EncloseNode enclose;\r | |
262 | BRefNode bref;\r | |
263 | AnchorNode anchor;\r | |
264 | ConsAltNode cons;\r | |
265 | CtypeNode ctype;\r | |
266 | #ifdef USE_SUBEXP_CALL\r | |
267 | CallNode call;\r | |
268 | #endif\r | |
269 | } u;\r | |
270 | } Node;\r | |
271 | \r | |
272 | \r | |
273 | #define NULL_NODE ((Node* )0)\r | |
274 | \r | |
275 | #define SCANENV_MEMNODES_SIZE 8\r | |
276 | #define SCANENV_MEM_NODES(senv) \\r | |
277 | (IS_NOT_NULL((senv)->mem_nodes_dynamic) ? \\r | |
278 | (senv)->mem_nodes_dynamic : (senv)->mem_nodes_static)\r | |
279 | \r | |
280 | typedef struct {\r | |
281 | OnigOptionType option;\r | |
282 | OnigCaseFoldType case_fold_flag;\r | |
283 | OnigEncoding enc;\r | |
284 | OnigSyntaxType* syntax;\r | |
285 | BitStatusType capture_history;\r | |
286 | BitStatusType bt_mem_start;\r | |
287 | BitStatusType bt_mem_end;\r | |
288 | BitStatusType backrefed_mem;\r | |
289 | UChar* pattern;\r | |
290 | UChar* pattern_end;\r | |
291 | UChar* error;\r | |
292 | UChar* error_end;\r | |
293 | regex_t* reg; /* for reg->names only */\r | |
294 | int num_call;\r | |
295 | #ifdef USE_SUBEXP_CALL\r | |
296 | UnsetAddrList* unset_addr_list;\r | |
297 | #endif\r | |
298 | int num_mem;\r | |
299 | #ifdef USE_NAMED_GROUP\r | |
300 | int num_named;\r | |
301 | #endif\r | |
302 | int mem_alloc;\r | |
303 | Node* mem_nodes_static[SCANENV_MEMNODES_SIZE];\r | |
304 | Node** mem_nodes_dynamic;\r | |
305 | #ifdef USE_COMBINATION_EXPLOSION_CHECK\r | |
306 | int num_comb_exp_check;\r | |
307 | int comb_exp_max_regnum;\r | |
308 | int curr_max_regnum;\r | |
309 | int has_recursion;\r | |
310 | #endif\r | |
311 | } ScanEnv;\r | |
312 | \r | |
313 | \r | |
314 | #define IS_SYNTAX_OP(syn, opm) (((syn)->op & (opm)) != 0)\r | |
315 | #define IS_SYNTAX_OP2(syn, opm) (((syn)->op2 & (opm)) != 0)\r | |
316 | #define IS_SYNTAX_BV(syn, bvm) (((syn)->behavior & (bvm)) != 0)\r | |
317 | \r | |
318 | #ifdef USE_NAMED_GROUP\r | |
319 | typedef struct {\r | |
320 | int new_val;\r | |
321 | } GroupNumRemap;\r | |
322 | \r | |
323 | extern int onig_renumber_name_table P_((regex_t* reg, GroupNumRemap* map));\r | |
324 | #endif\r | |
325 | \r | |
326 | extern int onig_strncmp P_((const UChar* s1, const UChar* s2, int n));\r | |
327 | extern void onig_strcpy P_((UChar* dest, const UChar* src, const UChar* end));\r | |
328 | extern void onig_scan_env_set_error_string P_((ScanEnv* env, int ecode, UChar* arg, UChar* arg_end));\r | |
329 | extern int onig_scan_unsigned_number P_((UChar** src, const UChar* end, OnigEncoding enc));\r | |
330 | extern void onig_reduce_nested_quantifier P_((Node* pnode, Node* cnode));\r | |
331 | extern void onig_node_conv_to_str_node P_((Node* node, int raw));\r | |
332 | extern int onig_node_str_cat P_((Node* node, const UChar* s, const UChar* end));\r | |
333 | extern int onig_node_str_set P_((Node* node, const UChar* s, const UChar* end));\r | |
334 | extern void onig_node_free P_((Node* node));\r | |
335 | extern Node* onig_node_new_enclose P_((int type));\r | |
336 | extern Node* onig_node_new_anchor P_((int type));\r | |
337 | extern Node* onig_node_new_str P_((const UChar* s, const UChar* end));\r | |
338 | extern Node* onig_node_new_list P_((Node* left, Node* right));\r | |
339 | extern Node* onig_node_list_add P_((Node* list, Node* x));\r | |
340 | extern Node* onig_node_new_alt P_((Node* left, Node* right));\r | |
341 | extern void onig_node_str_clear P_((Node* node));\r | |
342 | extern int onig_free_node_list P_((void));\r | |
343 | extern int onig_names_free P_((regex_t* reg));\r | |
344 | extern int onig_parse_make_tree P_((Node** root, const UChar* pattern, const UChar* end, regex_t* reg, ScanEnv* env));\r | |
345 | extern int onig_free_shared_cclass_table P_((void));\r | |
346 | \r | |
347 | #ifdef ONIG_DEBUG\r | |
348 | #ifdef USE_NAMED_GROUP\r | |
349 | extern int onig_print_names(FILE*, regex_t*);\r | |
350 | #endif\r | |
351 | #endif\r | |
352 | \r | |
353 | #endif /* REGPARSE_H */\r |