1 /**********************************************************************
2   regsyntax.c -  Oniguruma (regular expression library)
3 **********************************************************************/
4 /*-
5  * Copyright (c) 2002-2021  K.Kosako
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_ISOLATED_OPTION_CONTINUE_BRANCH |
156       ONIG_SYN_DIFFERENT_LEN_ALT_LOOK_BEHIND |
157       ONIG_SYN_VARIABLE_LEN_LOOK_BEHIND )
158   , ONIG_OPTION_SINGLELINE
159   ,
160   {
161       (OnigCodePoint )'\\'                       /* esc */
162     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
163     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
164     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
165     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
166     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
167   }
168 };
169 
170 OnigSyntaxType OnigSyntaxPerl = {
171   (( SYN_GNU_REGEX_OP | ONIG_SYN_OP_QMARK_NON_GREEDY |
172      ONIG_SYN_OP_ESC_OCTAL3 | ONIG_SYN_OP_ESC_X_HEX2 |
173      ONIG_SYN_OP_ESC_X_BRACE_HEX8 | ONIG_SYN_OP_ESC_O_BRACE_OCTAL |
174      ONIG_SYN_OP_ESC_CONTROL_CHARS |
175      ONIG_SYN_OP_ESC_C_CONTROL )
176    & ~ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END )
177   , ( ONIG_SYN_OP2_ESC_CAPITAL_Q_QUOTE |
178       ONIG_SYN_OP2_QMARK_GROUP_EFFECT | ONIG_SYN_OP2_OPTION_PERL |
179       ONIG_SYN_OP2_PLUS_POSSESSIVE_REPEAT | ONIG_SYN_OP2_PLUS_POSSESSIVE_INTERVAL |
180       ONIG_SYN_OP2_QMARK_LPAREN_IF_ELSE |
181       ONIG_SYN_OP2_QMARK_TILDE_ABSENT_GROUP |
182       ONIG_SYN_OP2_QMARK_BRACE_CALLOUT_CONTENTS |
183       ONIG_SYN_OP2_ASTERISK_CALLOUT_NAME    |
184       ONIG_SYN_OP2_ESC_X_Y_TEXT_SEGMENT |
185       ONIG_SYN_OP2_ESC_P_BRACE_CHAR_PROPERTY |
186       ONIG_SYN_OP2_ESC_P_BRACE_CIRCUMFLEX_NOT |
187       ONIG_SYN_OP2_ESC_CAPITAL_K_KEEP |
188       ONIG_SYN_OP2_ESC_CAPITAL_R_GENERAL_NEWLINE |
189       ONIG_SYN_OP2_ESC_CAPITAL_N_O_SUPER_DOT )
190   , SYN_GNU_REGEX_BV | ONIG_SYN_ISOLATED_OPTION_CONTINUE_BRANCH
191   , ONIG_OPTION_SINGLELINE
192   ,
193   {
194       (OnigCodePoint )'\\'                       /* esc */
195     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
196     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
197     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
198     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
199     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
200   }
201 };
202 
203 /* Perl + named group */
204 OnigSyntaxType OnigSyntaxPerl_NG = {
205   (( SYN_GNU_REGEX_OP | ONIG_SYN_OP_QMARK_NON_GREEDY |
206      ONIG_SYN_OP_ESC_OCTAL3 | ONIG_SYN_OP_ESC_X_HEX2 |
207      ONIG_SYN_OP_ESC_X_BRACE_HEX8 | ONIG_SYN_OP_ESC_O_BRACE_OCTAL |
208      ONIG_SYN_OP_ESC_CONTROL_CHARS |
209      ONIG_SYN_OP_ESC_C_CONTROL )
210    & ~ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END )
211   , ( ONIG_SYN_OP2_ESC_CAPITAL_Q_QUOTE |
212       ONIG_SYN_OP2_QMARK_GROUP_EFFECT | ONIG_SYN_OP2_OPTION_PERL |
213       ONIG_SYN_OP2_PLUS_POSSESSIVE_REPEAT | ONIG_SYN_OP2_PLUS_POSSESSIVE_INTERVAL |
214       ONIG_SYN_OP2_QMARK_LPAREN_IF_ELSE |
215       ONIG_SYN_OP2_QMARK_TILDE_ABSENT_GROUP |
216       ONIG_SYN_OP2_QMARK_BRACE_CALLOUT_CONTENTS |
217       ONIG_SYN_OP2_ASTERISK_CALLOUT_NAME    |
218       ONIG_SYN_OP2_ESC_X_Y_TEXT_SEGMENT |
219       ONIG_SYN_OP2_ESC_P_BRACE_CHAR_PROPERTY  |
220       ONIG_SYN_OP2_ESC_P_BRACE_CIRCUMFLEX_NOT |
221       ONIG_SYN_OP2_QMARK_LT_NAMED_GROUP       |
222       ONIG_SYN_OP2_ESC_K_NAMED_BACKREF        |
223       ONIG_SYN_OP2_ESC_G_SUBEXP_CALL |
224       ONIG_SYN_OP2_ESC_CAPITAL_K_KEEP |
225       ONIG_SYN_OP2_ESC_CAPITAL_R_GENERAL_NEWLINE |
226       ONIG_SYN_OP2_ESC_CAPITAL_N_O_SUPER_DOT |
227       ONIG_SYN_OP2_QMARK_PERL_SUBEXP_CALL )
228   , ( SYN_GNU_REGEX_BV | ONIG_SYN_ISOLATED_OPTION_CONTINUE_BRANCH |
229       ONIG_SYN_CAPTURE_ONLY_NAMED_GROUP |
230       ONIG_SYN_ALLOW_MULTIPLEX_DEFINITION_NAME )
231   , ONIG_OPTION_SINGLELINE
232   ,
233   {
234       (OnigCodePoint )'\\'                       /* esc */
235     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
236     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
237     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
238     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
239     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
240   }
241 };
242 
243 /* Python 3.9 */
244 OnigSyntaxType OnigSyntaxPython = {
245   (( SYN_GNU_REGEX_OP | ONIG_SYN_OP_QMARK_NON_GREEDY |
246      ONIG_SYN_OP_ESC_OCTAL3 | ONIG_SYN_OP_ESC_X_HEX2 |
247      ONIG_SYN_OP_ESC_CONTROL_CHARS |
248      ONIG_SYN_OP_ESC_C_CONTROL )
249    & ~ONIG_SYN_OP_ESC_LTGT_WORD_BEGIN_END )
250   , ( ONIG_SYN_OP2_QMARK_GROUP_EFFECT | ONIG_SYN_OP2_OPTION_PERL |
251       ONIG_SYN_OP2_QMARK_LPAREN_IF_ELSE |
252       ONIG_SYN_OP2_ASTERISK_CALLOUT_NAME    |
253       ONIG_SYN_OP2_ESC_P_BRACE_CHAR_PROPERTY  |
254       ONIG_SYN_OP2_ESC_P_BRACE_CIRCUMFLEX_NOT |
255       ONIG_SYN_OP2_QMARK_CAPITAL_P_NAME       |
256       ONIG_SYN_OP2_ESC_CAPITAL_K_KEEP |
257       ONIG_SYN_OP2_ESC_V_VTAB | ONIG_SYN_OP2_ESC_U_HEX4 )
258   , ( SYN_GNU_REGEX_BV | ONIG_SYN_ISOLATED_OPTION_CONTINUE_BRANCH |
259       ONIG_SYN_ALLOW_INTERVAL_LOW_ABBREV | ONIG_SYN_PYTHON )
260   , ONIG_OPTION_SINGLELINE
261   ,
262   {
263       (OnigCodePoint )'\\'                       /* esc */
264     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar '.'  */
265     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anytime '*'  */
266     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* zero or one time '?' */
267     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* one or more time '+' */
268     , (OnigCodePoint )ONIG_INEFFECTIVE_META_CHAR /* anychar anytime */
269   }
270 };
271 
272 
273 
274 extern int
onig_set_default_syntax(OnigSyntaxType * syntax)275 onig_set_default_syntax(OnigSyntaxType* syntax)
276 {
277   if (IS_NULL(syntax))
278     syntax = ONIG_SYNTAX_ONIGURUMA;
279 
280   OnigDefaultSyntax = syntax;
281   return 0;
282 }
283 
284 extern void
onig_copy_syntax(OnigSyntaxType * to,OnigSyntaxType * from)285 onig_copy_syntax(OnigSyntaxType* to, OnigSyntaxType* from)
286 {
287   *to = *from;
288 }
289 
290 extern void
onig_set_syntax_op(OnigSyntaxType * syntax,unsigned int op)291 onig_set_syntax_op(OnigSyntaxType* syntax, unsigned int op)
292 {
293   syntax->op = op;
294 }
295 
296 extern void
onig_set_syntax_op2(OnigSyntaxType * syntax,unsigned int op2)297 onig_set_syntax_op2(OnigSyntaxType* syntax, unsigned int op2)
298 {
299   syntax->op2 = op2;
300 }
301 
302 extern void
onig_set_syntax_behavior(OnigSyntaxType * syntax,unsigned int behavior)303 onig_set_syntax_behavior(OnigSyntaxType* syntax, unsigned int behavior)
304 {
305   syntax->behavior = behavior;
306 }
307 
308 extern void
onig_set_syntax_options(OnigSyntaxType * syntax,OnigOptionType options)309 onig_set_syntax_options(OnigSyntaxType* syntax, OnigOptionType options)
310 {
311   syntax->options = options;
312 }
313 
314 extern unsigned int
onig_get_syntax_op(OnigSyntaxType * syntax)315 onig_get_syntax_op(OnigSyntaxType* syntax)
316 {
317   return syntax->op;
318 }
319 
320 extern unsigned int
onig_get_syntax_op2(OnigSyntaxType * syntax)321 onig_get_syntax_op2(OnigSyntaxType* syntax)
322 {
323   return syntax->op2;
324 }
325 
326 extern unsigned int
onig_get_syntax_behavior(OnigSyntaxType * syntax)327 onig_get_syntax_behavior(OnigSyntaxType* syntax)
328 {
329   return syntax->behavior;
330 }
331 
332 extern OnigOptionType
onig_get_syntax_options(OnigSyntaxType * syntax)333 onig_get_syntax_options(OnigSyntaxType* syntax)
334 {
335   return syntax->options;
336 }
337 
338 #ifdef USE_VARIABLE_META_CHARS
onig_set_meta_char(OnigSyntaxType * enc,unsigned int what,OnigCodePoint code)339 extern int onig_set_meta_char(OnigSyntaxType* enc,
340                               unsigned int what, OnigCodePoint code)
341 {
342   switch (what) {
343   case ONIG_META_CHAR_ESCAPE:
344     enc->meta_char_table.esc = code;
345     break;
346   case ONIG_META_CHAR_ANYCHAR:
347     enc->meta_char_table.anychar = code;
348     break;
349   case ONIG_META_CHAR_ANYTIME:
350     enc->meta_char_table.anytime = code;
351     break;
352   case ONIG_META_CHAR_ZERO_OR_ONE_TIME:
353     enc->meta_char_table.zero_or_one_time = code;
354     break;
355   case ONIG_META_CHAR_ONE_OR_MORE_TIME:
356     enc->meta_char_table.one_or_more_time = code;
357     break;
358   case ONIG_META_CHAR_ANYCHAR_ANYTIME:
359     enc->meta_char_table.anychar_anytime = code;
360     break;
361   default:
362     return ONIGERR_INVALID_ARGUMENT;
363     break;
364   }
365   return 0;
366 }
367 #endif /* USE_VARIABLE_META_CHARS */
368