xref: /PHP-5.6/ext/mbstring/oniguruma/regsyntax.c (revision fe92d64a)
1 /**********************************************************************
2   regsyntax.c -  Oniguruma (regular expression library)
3 **********************************************************************/
4 /*-
5  * Copyright (c) 2002-2006  K.Kosako  <sndgk393 AT ybb DOT ne DOT jp>
6  * All rights reserved.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  *
17  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
18  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
21  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27  * SUCH DAMAGE.
28  */
29 
30 #include "regint.h"
31 
32 OnigSyntaxType OnigSyntaxASIS = {
33     0
34   , ONIG_SYN_OP2_INEFFECTIVE_ESCAPE
35   , 0
36   , ONIG_OPTION_NONE
37   ,
38   {
39       (OnigCodePoint )'\\'                       /* esc */
40     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
41     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
42     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
43     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
44     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
45   }
46 };
47 
48 OnigSyntaxType OnigSyntaxPosixBasic = {
49   ( SYN_POSIX_COMMON_OP | ONIG_SYN_OP_ESC_LPAREN_SUBEXP |
50     ONIG_SYN_OP_ESC_BRACE_INTERVAL )
51   , 0
52   , 0
53   , ( ONIG_OPTION_SINGLELINE | ONIG_OPTION_MULTILINE )
54   ,
55   {
56       (OnigCodePoint )'\\'                       /* esc */
57     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
58     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
59     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
60     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
61     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
62   }
63 };
64 
65 OnigSyntaxType OnigSyntaxPosixExtended = {
66   ( SYN_POSIX_COMMON_OP | ONIG_SYN_OP_LPAREN_SUBEXP |
67     ONIG_SYN_OP_BRACE_INTERVAL |
68     ONIG_SYN_OP_PLUS_ONE_INF | ONIG_SYN_OP_QMARK_ZERO_ONE | ONIG_SYN_OP_VBAR_ALT )
69   , 0
70   , ( ONIG_SYN_CONTEXT_INDEP_ANCHORS |
71       ONIG_SYN_CONTEXT_INDEP_REPEAT_OPS | ONIG_SYN_CONTEXT_INVALID_REPEAT_OPS |
72       ONIG_SYN_ALLOW_UNMATCHED_CLOSE_SUBEXP |
73       ONIG_SYN_ALLOW_DOUBLE_RANGE_OP_IN_CC )
74   , ( ONIG_OPTION_SINGLELINE | ONIG_OPTION_MULTILINE )
75   ,
76   {
77       (OnigCodePoint )'\\'                       /* esc */
78     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
79     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
80     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
81     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
82     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
83   }
84 };
85 
86 OnigSyntaxType OnigSyntaxEmacs = {
87   ( ONIG_SYN_OP_DOT_ANYCHAR | ONIG_SYN_OP_BRACKET_CC |
88     ONIG_SYN_OP_ESC_BRACE_INTERVAL |
89     ONIG_SYN_OP_ESC_LPAREN_SUBEXP | ONIG_SYN_OP_ESC_VBAR_ALT |
90     ONIG_SYN_OP_ASTERISK_ZERO_INF | ONIG_SYN_OP_PLUS_ONE_INF |
91     ONIG_SYN_OP_QMARK_ZERO_ONE | ONIG_SYN_OP_DECIMAL_BACKREF |
92     ONIG_SYN_OP_LINE_ANCHOR | ONIG_SYN_OP_ESC_CONTROL_CHARS )
93   , ONIG_SYN_OP2_ESC_GNU_BUF_ANCHOR
94   , ONIG_SYN_ALLOW_EMPTY_RANGE_IN_CC
95   , ONIG_OPTION_NONE
96   ,
97   {
98       (OnigCodePoint )'\\'                       /* esc */
99     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
100     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
101     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
102     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
103     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
104   }
105 };
106 
107 OnigSyntaxType OnigSyntaxGrep = {
108   ( ONIG_SYN_OP_DOT_ANYCHAR | ONIG_SYN_OP_BRACKET_CC | ONIG_SYN_OP_POSIX_BRACKET |
109     ONIG_SYN_OP_ESC_BRACE_INTERVAL | ONIG_SYN_OP_ESC_LPAREN_SUBEXP |
110     ONIG_SYN_OP_ESC_VBAR_ALT |
111     ONIG_SYN_OP_ASTERISK_ZERO_INF | ONIG_SYN_OP_ESC_PLUS_ONE_INF |
112     ONIG_SYN_OP_ESC_QMARK_ZERO_ONE | ONIG_SYN_OP_LINE_ANCHOR |
113     ONIG_SYN_OP_ESC_W_WORD | ONIG_SYN_OP_ESC_B_WORD_BOUND |
114     ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END | ONIG_SYN_OP_DECIMAL_BACKREF )
115   , 0
116   , ( ONIG_SYN_ALLOW_EMPTY_RANGE_IN_CC | ONIG_SYN_NOT_NEWLINE_IN_NEGATIVE_CC )
117   , ONIG_OPTION_NONE
118   ,
119   {
120       (OnigCodePoint )'\\'                       /* esc */
121     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
122     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
123     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
124     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
125     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
126   }
127 };
128 
129 OnigSyntaxType OnigSyntaxGnuRegex = {
130   SYN_GNU_REGEX_OP
131   , 0
132   , SYN_GNU_REGEX_BV
133   , ONIG_OPTION_NONE
134   ,
135   {
136       (OnigCodePoint )'\\'                       /* esc */
137     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
138     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
139     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
140     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
141     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
142   }
143 };
144 
145 OnigSyntaxType OnigSyntaxJava = {
146   (( SYN_GNU_REGEX_OP | ONIG_SYN_OP_QMARK_NON_GREEDY |
147      ONIG_SYN_OP_ESC_CONTROL_CHARS | ONIG_SYN_OP_ESC_C_CONTROL |
148      ONIG_SYN_OP_ESC_OCTAL3 | ONIG_SYN_OP_ESC_X_HEX2 )
149    & ~ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END )
150   , ( ONIG_SYN_OP2_ESC_CAPITAL_Q_QUOTE | ONIG_SYN_OP2_QMARK_GROUP_EFFECT |
151       ONIG_SYN_OP2_OPTION_PERL | ONIG_SYN_OP2_PLUS_POSSESSIVE_REPEAT |
152       ONIG_SYN_OP2_PLUS_POSSESSIVE_INTERVAL | ONIG_SYN_OP2_CCLASS_SET_OP |
153       ONIG_SYN_OP2_ESC_V_VTAB | ONIG_SYN_OP2_ESC_U_HEX4 |
154       ONIG_SYN_OP2_ESC_P_BRACE_CHAR_PROPERTY )
155   , ( SYN_GNU_REGEX_BV | ONIG_SYN_DIFFERENT_LEN_ALT_LOOK_BEHIND )
156   , ONIG_OPTION_SINGLELINE
157   ,
158   {
159       (OnigCodePoint )'\\'                       /* esc */
160     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
161     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
162     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
163     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
164     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
165   }
166 };
167 
168 OnigSyntaxType OnigSyntaxPerl = {
169   (( SYN_GNU_REGEX_OP | ONIG_SYN_OP_QMARK_NON_GREEDY |
170      ONIG_SYN_OP_ESC_OCTAL3 | ONIG_SYN_OP_ESC_X_HEX2 |
171      ONIG_SYN_OP_ESC_X_BRACE_HEX8 | ONIG_SYN_OP_ESC_CONTROL_CHARS |
172      ONIG_SYN_OP_ESC_C_CONTROL )
173    & ~ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END )
174   , ( ONIG_SYN_OP2_ESC_CAPITAL_Q_QUOTE |
175       ONIG_SYN_OP2_QMARK_GROUP_EFFECT | ONIG_SYN_OP2_OPTION_PERL |
176       ONIG_SYN_OP2_ESC_P_BRACE_CHAR_PROPERTY |
177       ONIG_SYN_OP2_ESC_P_BRACE_CIRCUMFLEX_NOT )
178   , SYN_GNU_REGEX_BV
179   , ONIG_OPTION_SINGLELINE
180   ,
181   {
182       (OnigCodePoint )'\\'                       /* esc */
183     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
184     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
185     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
186     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
187     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
188   }
189 };
190 
191 /* Perl + named group */
192 OnigSyntaxType OnigSyntaxPerl_NG = {
193   (( SYN_GNU_REGEX_OP | ONIG_SYN_OP_QMARK_NON_GREEDY |
194      ONIG_SYN_OP_ESC_OCTAL3 | ONIG_SYN_OP_ESC_X_HEX2 |
195      ONIG_SYN_OP_ESC_X_BRACE_HEX8 | ONIG_SYN_OP_ESC_CONTROL_CHARS |
196      ONIG_SYN_OP_ESC_C_CONTROL )
197    & ~ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END )
198   , ( ONIG_SYN_OP2_ESC_CAPITAL_Q_QUOTE |
199       ONIG_SYN_OP2_QMARK_GROUP_EFFECT | ONIG_SYN_OP2_OPTION_PERL |
200       ONIG_SYN_OP2_ESC_P_BRACE_CHAR_PROPERTY  |
201       ONIG_SYN_OP2_ESC_P_BRACE_CIRCUMFLEX_NOT |
202       ONIG_SYN_OP2_QMARK_LT_NAMED_GROUP       |
203       ONIG_SYN_OP2_ESC_K_NAMED_BACKREF        |
204       ONIG_SYN_OP2_ESC_G_SUBEXP_CALL )
205   , ( SYN_GNU_REGEX_BV |
206       ONIG_SYN_CAPTURE_ONLY_NAMED_GROUP |
207       ONIG_SYN_ALLOW_MULTIPLEX_DEFINITION_NAME )
208   , ONIG_OPTION_SINGLELINE
209   ,
210   {
211       (OnigCodePoint )'\\'                       /* esc */
212     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
213     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
214     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
215     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
216     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
217   }
218 };
219 
220 
221 
222 extern int
onig_set_default_syntax(OnigSyntaxType * syntax)223 onig_set_default_syntax(OnigSyntaxType* syntax)
224 {
225   if (IS_NULL(syntax))
226     syntax = ONIG_SYNTAX_RUBY;
227 
228   OnigDefaultSyntax = syntax;
229   return 0;
230 }
231 
232 extern void
onig_copy_syntax(OnigSyntaxType * to,OnigSyntaxType * from)233 onig_copy_syntax(OnigSyntaxType* to, OnigSyntaxType* from)
234 {
235   *to = *from;
236 }
237 
238 extern void
onig_set_syntax_op(OnigSyntaxType * syntax,unsigned int op)239 onig_set_syntax_op(OnigSyntaxType* syntax, unsigned int op)
240 {
241   syntax->op = op;
242 }
243 
244 extern void
onig_set_syntax_op2(OnigSyntaxType * syntax,unsigned int op2)245 onig_set_syntax_op2(OnigSyntaxType* syntax, unsigned int op2)
246 {
247   syntax->op2 = op2;
248 }
249 
250 extern void
onig_set_syntax_behavior(OnigSyntaxType * syntax,unsigned int behavior)251 onig_set_syntax_behavior(OnigSyntaxType* syntax, unsigned int behavior)
252 {
253   syntax->behavior = behavior;
254 }
255 
256 extern void
onig_set_syntax_options(OnigSyntaxType * syntax,OnigOptionType options)257 onig_set_syntax_options(OnigSyntaxType* syntax, OnigOptionType options)
258 {
259   syntax->options = options;
260 }
261 
262 extern unsigned int
onig_get_syntax_op(OnigSyntaxType * syntax)263 onig_get_syntax_op(OnigSyntaxType* syntax)
264 {
265   return syntax->op;
266 }
267 
268 extern unsigned int
onig_get_syntax_op2(OnigSyntaxType * syntax)269 onig_get_syntax_op2(OnigSyntaxType* syntax)
270 {
271   return syntax->op2;
272 }
273 
274 extern unsigned int
onig_get_syntax_behavior(OnigSyntaxType * syntax)275 onig_get_syntax_behavior(OnigSyntaxType* syntax)
276 {
277   return syntax->behavior;
278 }
279 
280 extern OnigOptionType
onig_get_syntax_options(OnigSyntaxType * syntax)281 onig_get_syntax_options(OnigSyntaxType* syntax)
282 {
283   return syntax->options;
284 }
285 
286 #ifdef USE_VARIABLE_META_CHARS
onig_set_meta_char(OnigSyntaxType * enc,unsigned int what,OnigCodePoint code)287 extern int onig_set_meta_char(OnigSyntaxType* enc,
288                               unsigned int what, OnigCodePoint code)
289 {
290   switch (what) {
291   case ONIG_META_CHAR_ESCAPE:
292     enc->meta_char_table.esc = code;
293     break;
294   case ONIG_META_CHAR_ANYCHAR:
295     enc->meta_char_table.anychar = code;
296     break;
297   case ONIG_META_CHAR_ANYTIME:
298     enc->meta_char_table.anytime = code;
299     break;
300   case ONIG_META_CHAR_ZERO_OR_ONE_TIME:
301     enc->meta_char_table.zero_or_one_time = code;
302     break;
303   case ONIG_META_CHAR_ONE_OR_MORE_TIME:
304     enc->meta_char_table.one_or_more_time = code;
305     break;
306   case ONIG_META_CHAR_ANYCHAR_ANYTIME:
307     enc->meta_char_table.anychar_anytime = code;
308     break;
309   default:
310     return ONIGERR_INVALID_ARGUMENT;
311     break;
312   }
313   return 0;
314 }
315 #endif /* USE_VARIABLE_META_CHARS */
316