]>
Commit | Line | Data |
---|---|---|
7c673cae FG |
1 | /* |
2 | * | |
3 | * Copyright (c) 2004 | |
4 | * John Maddock | |
5 | * | |
6 | * Use, modification and distribution are subject to the | |
7 | * Boost Software License, Version 1.0. (See accompanying file | |
8 | * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) | |
9 | * | |
10 | */ | |
11 | ||
12 | /* | |
13 | * LOCATION: see http://www.boost.org for most recent version. | |
14 | * FILE basic_regex_creator.cpp | |
15 | * VERSION see <boost/version.hpp> | |
16 | * DESCRIPTION: Declares template class basic_regex_creator which fills in | |
17 | * the data members of a regex_data object. | |
18 | */ | |
19 | ||
20 | #ifndef BOOST_REGEX_V4_BASIC_REGEX_CREATOR_HPP | |
21 | #define BOOST_REGEX_V4_BASIC_REGEX_CREATOR_HPP | |
22 | ||
23 | #ifdef BOOST_MSVC | |
24 | #pragma warning(push) | |
25 | #pragma warning(disable: 4103) | |
26 | #endif | |
27 | #ifdef BOOST_HAS_ABI_HEADERS | |
28 | # include BOOST_ABI_PREFIX | |
29 | #endif | |
30 | #ifdef BOOST_MSVC | |
31 | #pragma warning(pop) | |
32 | #endif | |
33 | ||
34 | #ifdef BOOST_MSVC | |
35 | # pragma warning(push) | |
36 | # pragma warning(disable: 4800) | |
37 | #endif | |
38 | ||
39 | namespace boost{ | |
40 | ||
41 | namespace BOOST_REGEX_DETAIL_NS{ | |
42 | ||
43 | template <class charT> | |
44 | struct digraph : public std::pair<charT, charT> | |
45 | { | |
46 | digraph() : std::pair<charT, charT>(0, 0){} | |
47 | digraph(charT c1) : std::pair<charT, charT>(c1, 0){} | |
48 | digraph(charT c1, charT c2) : std::pair<charT, charT>(c1, c2) | |
49 | {} | |
50 | digraph(const digraph<charT>& d) : std::pair<charT, charT>(d.first, d.second){} | |
51 | template <class Seq> | |
52 | digraph(const Seq& s) : std::pair<charT, charT>() | |
53 | { | |
54 | BOOST_ASSERT(s.size() <= 2); | |
55 | BOOST_ASSERT(s.size()); | |
56 | this->first = s[0]; | |
57 | this->second = (s.size() > 1) ? s[1] : 0; | |
58 | } | |
59 | }; | |
60 | ||
61 | template <class charT, class traits> | |
62 | class basic_char_set | |
63 | { | |
64 | public: | |
65 | typedef digraph<charT> digraph_type; | |
66 | typedef typename traits::string_type string_type; | |
67 | typedef typename traits::char_class_type m_type; | |
68 | ||
69 | basic_char_set() | |
70 | { | |
71 | m_negate = false; | |
72 | m_has_digraphs = false; | |
73 | m_classes = 0; | |
74 | m_negated_classes = 0; | |
75 | m_empty = true; | |
76 | } | |
77 | ||
78 | void add_single(const digraph_type& s) | |
79 | { | |
80 | m_singles.insert(m_singles.end(), s); | |
81 | if(s.second) | |
82 | m_has_digraphs = true; | |
83 | m_empty = false; | |
84 | } | |
85 | void add_range(const digraph_type& first, const digraph_type& end) | |
86 | { | |
87 | m_ranges.insert(m_ranges.end(), first); | |
88 | m_ranges.insert(m_ranges.end(), end); | |
89 | if(first.second) | |
90 | { | |
91 | m_has_digraphs = true; | |
92 | add_single(first); | |
93 | } | |
94 | if(end.second) | |
95 | { | |
96 | m_has_digraphs = true; | |
97 | add_single(end); | |
98 | } | |
99 | m_empty = false; | |
100 | } | |
101 | void add_class(m_type m) | |
102 | { | |
103 | m_classes |= m; | |
104 | m_empty = false; | |
105 | } | |
106 | void add_negated_class(m_type m) | |
107 | { | |
108 | m_negated_classes |= m; | |
109 | m_empty = false; | |
110 | } | |
111 | void add_equivalent(const digraph_type& s) | |
112 | { | |
113 | m_equivalents.insert(m_equivalents.end(), s); | |
114 | if(s.second) | |
115 | { | |
116 | m_has_digraphs = true; | |
117 | add_single(s); | |
118 | } | |
119 | m_empty = false; | |
120 | } | |
121 | void negate() | |
122 | { | |
123 | m_negate = true; | |
124 | //m_empty = false; | |
125 | } | |
126 | ||
127 | // | |
128 | // accessor functions: | |
129 | // | |
130 | bool has_digraphs()const | |
131 | { | |
132 | return m_has_digraphs; | |
133 | } | |
134 | bool is_negated()const | |
135 | { | |
136 | return m_negate; | |
137 | } | |
138 | typedef typename std::vector<digraph_type>::const_iterator list_iterator; | |
139 | list_iterator singles_begin()const | |
140 | { | |
141 | return m_singles.begin(); | |
142 | } | |
143 | list_iterator singles_end()const | |
144 | { | |
145 | return m_singles.end(); | |
146 | } | |
147 | list_iterator ranges_begin()const | |
148 | { | |
149 | return m_ranges.begin(); | |
150 | } | |
151 | list_iterator ranges_end()const | |
152 | { | |
153 | return m_ranges.end(); | |
154 | } | |
155 | list_iterator equivalents_begin()const | |
156 | { | |
157 | return m_equivalents.begin(); | |
158 | } | |
159 | list_iterator equivalents_end()const | |
160 | { | |
161 | return m_equivalents.end(); | |
162 | } | |
163 | m_type classes()const | |
164 | { | |
165 | return m_classes; | |
166 | } | |
167 | m_type negated_classes()const | |
168 | { | |
169 | return m_negated_classes; | |
170 | } | |
171 | bool empty()const | |
172 | { | |
173 | return m_empty; | |
174 | } | |
175 | private: | |
176 | std::vector<digraph_type> m_singles; // a list of single characters to match | |
177 | std::vector<digraph_type> m_ranges; // a list of end points of our ranges | |
178 | bool m_negate; // true if the set is to be negated | |
179 | bool m_has_digraphs; // true if we have digraphs present | |
180 | m_type m_classes; // character classes to match | |
181 | m_type m_negated_classes; // negated character classes to match | |
182 | bool m_empty; // whether we've added anything yet | |
183 | std::vector<digraph_type> m_equivalents; // a list of equivalence classes | |
184 | }; | |
185 | ||
186 | template <class charT, class traits> | |
187 | class basic_regex_creator | |
188 | { | |
189 | public: | |
190 | basic_regex_creator(regex_data<charT, traits>* data); | |
191 | std::ptrdiff_t getoffset(void* addr) | |
192 | { | |
193 | return getoffset(addr, m_pdata->m_data.data()); | |
194 | } | |
195 | std::ptrdiff_t getoffset(const void* addr, const void* base) | |
196 | { | |
197 | return static_cast<const char*>(addr) - static_cast<const char*>(base); | |
198 | } | |
199 | re_syntax_base* getaddress(std::ptrdiff_t off) | |
200 | { | |
201 | return getaddress(off, m_pdata->m_data.data()); | |
202 | } | |
203 | re_syntax_base* getaddress(std::ptrdiff_t off, void* base) | |
204 | { | |
205 | return static_cast<re_syntax_base*>(static_cast<void*>(static_cast<char*>(base) + off)); | |
206 | } | |
207 | void init(unsigned l_flags) | |
208 | { | |
209 | m_pdata->m_flags = l_flags; | |
210 | m_icase = l_flags & regex_constants::icase; | |
211 | } | |
212 | regbase::flag_type flags() | |
213 | { | |
214 | return m_pdata->m_flags; | |
215 | } | |
216 | void flags(regbase::flag_type f) | |
217 | { | |
218 | m_pdata->m_flags = f; | |
219 | if(m_icase != static_cast<bool>(f & regbase::icase)) | |
220 | { | |
221 | m_icase = static_cast<bool>(f & regbase::icase); | |
222 | } | |
223 | } | |
224 | re_syntax_base* append_state(syntax_element_type t, std::size_t s = sizeof(re_syntax_base)); | |
225 | re_syntax_base* insert_state(std::ptrdiff_t pos, syntax_element_type t, std::size_t s = sizeof(re_syntax_base)); | |
226 | re_literal* append_literal(charT c); | |
227 | re_syntax_base* append_set(const basic_char_set<charT, traits>& char_set); | |
228 | re_syntax_base* append_set(const basic_char_set<charT, traits>& char_set, mpl::false_*); | |
229 | re_syntax_base* append_set(const basic_char_set<charT, traits>& char_set, mpl::true_*); | |
230 | void finalize(const charT* p1, const charT* p2); | |
231 | protected: | |
232 | regex_data<charT, traits>* m_pdata; // pointer to the basic_regex_data struct we are filling in | |
233 | const ::boost::regex_traits_wrapper<traits>& | |
234 | m_traits; // convenience reference to traits class | |
235 | re_syntax_base* m_last_state; // the last state we added | |
236 | bool m_icase; // true for case insensitive matches | |
237 | unsigned m_repeater_id; // the state_id of the next repeater | |
238 | bool m_has_backrefs; // true if there are actually any backrefs | |
239 | unsigned m_backrefs; // bitmask of permitted backrefs | |
240 | boost::uintmax_t m_bad_repeats; // bitmask of repeats we can't deduce a startmap for; | |
241 | bool m_has_recursions; // set when we have recursive expresisons to fixup | |
242 | std::vector<bool> m_recursion_checks; // notes which recursions we've followed while analysing this expression | |
243 | typename traits::char_class_type m_word_mask; // mask used to determine if a character is a word character | |
244 | typename traits::char_class_type m_mask_space; // mask used to determine if a character is a word character | |
245 | typename traits::char_class_type m_lower_mask; // mask used to determine if a character is a lowercase character | |
246 | typename traits::char_class_type m_upper_mask; // mask used to determine if a character is an uppercase character | |
247 | typename traits::char_class_type m_alpha_mask; // mask used to determine if a character is an alphabetic character | |
248 | private: | |
249 | basic_regex_creator& operator=(const basic_regex_creator&); | |
250 | basic_regex_creator(const basic_regex_creator&); | |
251 | ||
252 | void fixup_pointers(re_syntax_base* state); | |
253 | void fixup_recursions(re_syntax_base* state); | |
254 | void create_startmaps(re_syntax_base* state); | |
255 | int calculate_backstep(re_syntax_base* state); | |
256 | void create_startmap(re_syntax_base* state, unsigned char* l_map, unsigned int* pnull, unsigned char mask); | |
257 | unsigned get_restart_type(re_syntax_base* state); | |
258 | void set_all_masks(unsigned char* bits, unsigned char); | |
259 | bool is_bad_repeat(re_syntax_base* pt); | |
260 | void set_bad_repeat(re_syntax_base* pt); | |
261 | syntax_element_type get_repeat_type(re_syntax_base* state); | |
262 | void probe_leading_repeat(re_syntax_base* state); | |
263 | }; | |
264 | ||
265 | template <class charT, class traits> | |
266 | basic_regex_creator<charT, traits>::basic_regex_creator(regex_data<charT, traits>* data) | |
267 | : m_pdata(data), m_traits(*(data->m_ptraits)), m_last_state(0), m_repeater_id(0), m_has_backrefs(false), m_backrefs(0), m_has_recursions(false) | |
268 | { | |
269 | m_pdata->m_data.clear(); | |
270 | m_pdata->m_status = ::boost::regex_constants::error_ok; | |
271 | static const charT w = 'w'; | |
272 | static const charT s = 's'; | |
273 | static const charT l[5] = { 'l', 'o', 'w', 'e', 'r', }; | |
274 | static const charT u[5] = { 'u', 'p', 'p', 'e', 'r', }; | |
275 | static const charT a[5] = { 'a', 'l', 'p', 'h', 'a', }; | |
276 | m_word_mask = m_traits.lookup_classname(&w, &w +1); | |
277 | m_mask_space = m_traits.lookup_classname(&s, &s +1); | |
278 | m_lower_mask = m_traits.lookup_classname(l, l + 5); | |
279 | m_upper_mask = m_traits.lookup_classname(u, u + 5); | |
280 | m_alpha_mask = m_traits.lookup_classname(a, a + 5); | |
281 | m_pdata->m_word_mask = m_word_mask; | |
282 | BOOST_ASSERT(m_word_mask != 0); | |
283 | BOOST_ASSERT(m_mask_space != 0); | |
284 | BOOST_ASSERT(m_lower_mask != 0); | |
285 | BOOST_ASSERT(m_upper_mask != 0); | |
286 | BOOST_ASSERT(m_alpha_mask != 0); | |
287 | } | |
288 | ||
289 | template <class charT, class traits> | |
290 | re_syntax_base* basic_regex_creator<charT, traits>::append_state(syntax_element_type t, std::size_t s) | |
291 | { | |
292 | // if the state is a backref then make a note of it: | |
293 | if(t == syntax_element_backref) | |
294 | this->m_has_backrefs = true; | |
295 | // append a new state, start by aligning our last one: | |
296 | m_pdata->m_data.align(); | |
297 | // set the offset to the next state in our last one: | |
298 | if(m_last_state) | |
299 | m_last_state->next.i = m_pdata->m_data.size() - getoffset(m_last_state); | |
300 | // now actually extent our data: | |
301 | m_last_state = static_cast<re_syntax_base*>(m_pdata->m_data.extend(s)); | |
302 | // fill in boilerplate options in the new state: | |
303 | m_last_state->next.i = 0; | |
304 | m_last_state->type = t; | |
305 | return m_last_state; | |
306 | } | |
307 | ||
308 | template <class charT, class traits> | |
309 | re_syntax_base* basic_regex_creator<charT, traits>::insert_state(std::ptrdiff_t pos, syntax_element_type t, std::size_t s) | |
310 | { | |
311 | // append a new state, start by aligning our last one: | |
312 | m_pdata->m_data.align(); | |
313 | // set the offset to the next state in our last one: | |
314 | if(m_last_state) | |
315 | m_last_state->next.i = m_pdata->m_data.size() - getoffset(m_last_state); | |
316 | // remember the last state position: | |
317 | std::ptrdiff_t off = getoffset(m_last_state) + s; | |
318 | // now actually insert our data: | |
319 | re_syntax_base* new_state = static_cast<re_syntax_base*>(m_pdata->m_data.insert(pos, s)); | |
320 | // fill in boilerplate options in the new state: | |
321 | new_state->next.i = s; | |
322 | new_state->type = t; | |
323 | m_last_state = getaddress(off); | |
324 | return new_state; | |
325 | } | |
326 | ||
327 | template <class charT, class traits> | |
328 | re_literal* basic_regex_creator<charT, traits>::append_literal(charT c) | |
329 | { | |
330 | re_literal* result; | |
331 | // start by seeing if we have an existing re_literal we can extend: | |
332 | if((0 == m_last_state) || (m_last_state->type != syntax_element_literal)) | |
333 | { | |
334 | // no existing re_literal, create a new one: | |
335 | result = static_cast<re_literal*>(append_state(syntax_element_literal, sizeof(re_literal) + sizeof(charT))); | |
336 | result->length = 1; | |
337 | *static_cast<charT*>(static_cast<void*>(result+1)) = m_traits.translate(c, m_icase); | |
338 | } | |
339 | else | |
340 | { | |
341 | // we have an existing re_literal, extend it: | |
342 | std::ptrdiff_t off = getoffset(m_last_state); | |
343 | m_pdata->m_data.extend(sizeof(charT)); | |
344 | m_last_state = result = static_cast<re_literal*>(getaddress(off)); | |
345 | charT* characters = static_cast<charT*>(static_cast<void*>(result+1)); | |
346 | characters[result->length] = m_traits.translate(c, m_icase); | |
347 | result->length += 1; | |
348 | } | |
349 | return result; | |
350 | } | |
351 | ||
352 | template <class charT, class traits> | |
353 | inline re_syntax_base* basic_regex_creator<charT, traits>::append_set( | |
354 | const basic_char_set<charT, traits>& char_set) | |
355 | { | |
356 | typedef mpl::bool_< (sizeof(charT) == 1) > truth_type; | |
357 | return char_set.has_digraphs() | |
358 | ? append_set(char_set, static_cast<mpl::false_*>(0)) | |
359 | : append_set(char_set, static_cast<truth_type*>(0)); | |
360 | } | |
361 | ||
362 | template <class charT, class traits> | |
363 | re_syntax_base* basic_regex_creator<charT, traits>::append_set( | |
364 | const basic_char_set<charT, traits>& char_set, mpl::false_*) | |
365 | { | |
366 | typedef typename traits::string_type string_type; | |
367 | typedef typename basic_char_set<charT, traits>::list_iterator item_iterator; | |
368 | typedef typename traits::char_class_type m_type; | |
369 | ||
370 | re_set_long<m_type>* result = static_cast<re_set_long<m_type>*>(append_state(syntax_element_long_set, sizeof(re_set_long<m_type>))); | |
371 | // | |
372 | // fill in the basics: | |
373 | // | |
374 | result->csingles = static_cast<unsigned int>(::boost::BOOST_REGEX_DETAIL_NS::distance(char_set.singles_begin(), char_set.singles_end())); | |
375 | result->cranges = static_cast<unsigned int>(::boost::BOOST_REGEX_DETAIL_NS::distance(char_set.ranges_begin(), char_set.ranges_end())) / 2; | |
376 | result->cequivalents = static_cast<unsigned int>(::boost::BOOST_REGEX_DETAIL_NS::distance(char_set.equivalents_begin(), char_set.equivalents_end())); | |
377 | result->cclasses = char_set.classes(); | |
378 | result->cnclasses = char_set.negated_classes(); | |
379 | if(flags() & regbase::icase) | |
380 | { | |
381 | // adjust classes as needed: | |
382 | if(((result->cclasses & m_lower_mask) == m_lower_mask) || ((result->cclasses & m_upper_mask) == m_upper_mask)) | |
383 | result->cclasses |= m_alpha_mask; | |
384 | if(((result->cnclasses & m_lower_mask) == m_lower_mask) || ((result->cnclasses & m_upper_mask) == m_upper_mask)) | |
385 | result->cnclasses |= m_alpha_mask; | |
386 | } | |
387 | ||
388 | result->isnot = char_set.is_negated(); | |
389 | result->singleton = !char_set.has_digraphs(); | |
390 | // | |
391 | // remember where the state is for later: | |
392 | // | |
393 | std::ptrdiff_t offset = getoffset(result); | |
394 | // | |
395 | // now extend with all the singles: | |
396 | // | |
397 | item_iterator first, last; | |
398 | first = char_set.singles_begin(); | |
399 | last = char_set.singles_end(); | |
400 | while(first != last) | |
401 | { | |
402 | charT* p = static_cast<charT*>(this->m_pdata->m_data.extend(sizeof(charT) * (first->second ? 3 : 2))); | |
403 | p[0] = m_traits.translate(first->first, m_icase); | |
404 | if(first->second) | |
405 | { | |
406 | p[1] = m_traits.translate(first->second, m_icase); | |
407 | p[2] = 0; | |
408 | } | |
409 | else | |
410 | p[1] = 0; | |
411 | ++first; | |
412 | } | |
413 | // | |
414 | // now extend with all the ranges: | |
415 | // | |
416 | first = char_set.ranges_begin(); | |
417 | last = char_set.ranges_end(); | |
418 | while(first != last) | |
419 | { | |
420 | // first grab the endpoints of the range: | |
421 | digraph<charT> c1 = *first; | |
422 | c1.first = this->m_traits.translate(c1.first, this->m_icase); | |
423 | c1.second = this->m_traits.translate(c1.second, this->m_icase); | |
424 | ++first; | |
425 | digraph<charT> c2 = *first; | |
426 | c2.first = this->m_traits.translate(c2.first, this->m_icase); | |
427 | c2.second = this->m_traits.translate(c2.second, this->m_icase); | |
428 | ++first; | |
429 | string_type s1, s2; | |
430 | // different actions now depending upon whether collation is turned on: | |
431 | if(flags() & regex_constants::collate) | |
432 | { | |
433 | // we need to transform our range into sort keys: | |
434 | charT a1[3] = { c1.first, c1.second, charT(0), }; | |
435 | charT a2[3] = { c2.first, c2.second, charT(0), }; | |
436 | s1 = this->m_traits.transform(a1, (a1[1] ? a1+2 : a1+1)); | |
437 | s2 = this->m_traits.transform(a2, (a2[1] ? a2+2 : a2+1)); | |
438 | if(s1.size() == 0) | |
439 | s1 = string_type(1, charT(0)); | |
440 | if(s2.size() == 0) | |
441 | s2 = string_type(1, charT(0)); | |
442 | } | |
443 | else | |
444 | { | |
445 | if(c1.second) | |
446 | { | |
447 | s1.insert(s1.end(), c1.first); | |
448 | s1.insert(s1.end(), c1.second); | |
449 | } | |
450 | else | |
451 | s1 = string_type(1, c1.first); | |
452 | if(c2.second) | |
453 | { | |
454 | s2.insert(s2.end(), c2.first); | |
455 | s2.insert(s2.end(), c2.second); | |
456 | } | |
457 | else | |
458 | s2.insert(s2.end(), c2.first); | |
459 | } | |
460 | if(s1 > s2) | |
461 | { | |
462 | // Oops error: | |
463 | return 0; | |
464 | } | |
465 | charT* p = static_cast<charT*>(this->m_pdata->m_data.extend(sizeof(charT) * (s1.size() + s2.size() + 2) ) ); | |
466 | BOOST_REGEX_DETAIL_NS::copy(s1.begin(), s1.end(), p); | |
467 | p[s1.size()] = charT(0); | |
468 | p += s1.size() + 1; | |
469 | BOOST_REGEX_DETAIL_NS::copy(s2.begin(), s2.end(), p); | |
470 | p[s2.size()] = charT(0); | |
471 | } | |
472 | // | |
473 | // now process the equivalence classes: | |
474 | // | |
475 | first = char_set.equivalents_begin(); | |
476 | last = char_set.equivalents_end(); | |
477 | while(first != last) | |
478 | { | |
479 | string_type s; | |
480 | if(first->second) | |
481 | { | |
482 | charT cs[3] = { first->first, first->second, charT(0), }; | |
483 | s = m_traits.transform_primary(cs, cs+2); | |
484 | } | |
485 | else | |
486 | s = m_traits.transform_primary(&first->first, &first->first+1); | |
487 | if(s.empty()) | |
488 | return 0; // invalid or unsupported equivalence class | |
489 | charT* p = static_cast<charT*>(this->m_pdata->m_data.extend(sizeof(charT) * (s.size()+1) ) ); | |
490 | BOOST_REGEX_DETAIL_NS::copy(s.begin(), s.end(), p); | |
491 | p[s.size()] = charT(0); | |
492 | ++first; | |
493 | } | |
494 | // | |
495 | // finally reset the address of our last state: | |
496 | // | |
497 | m_last_state = result = static_cast<re_set_long<m_type>*>(getaddress(offset)); | |
498 | return result; | |
499 | } | |
500 | ||
501 | template<class T> | |
502 | inline bool char_less(T t1, T t2) | |
503 | { | |
504 | return t1 < t2; | |
505 | } | |
506 | inline bool char_less(char t1, char t2) | |
507 | { | |
508 | return static_cast<unsigned char>(t1) < static_cast<unsigned char>(t2); | |
509 | } | |
510 | inline bool char_less(signed char t1, signed char t2) | |
511 | { | |
512 | return static_cast<unsigned char>(t1) < static_cast<unsigned char>(t2); | |
513 | } | |
514 | ||
515 | template <class charT, class traits> | |
516 | re_syntax_base* basic_regex_creator<charT, traits>::append_set( | |
517 | const basic_char_set<charT, traits>& char_set, mpl::true_*) | |
518 | { | |
519 | typedef typename traits::string_type string_type; | |
520 | typedef typename basic_char_set<charT, traits>::list_iterator item_iterator; | |
521 | ||
522 | re_set* result = static_cast<re_set*>(append_state(syntax_element_set, sizeof(re_set))); | |
523 | bool negate = char_set.is_negated(); | |
524 | std::memset(result->_map, 0, sizeof(result->_map)); | |
525 | // | |
526 | // handle singles first: | |
527 | // | |
528 | item_iterator first, last; | |
529 | first = char_set.singles_begin(); | |
530 | last = char_set.singles_end(); | |
531 | while(first != last) | |
532 | { | |
533 | for(unsigned int i = 0; i < (1 << CHAR_BIT); ++i) | |
534 | { | |
535 | if(this->m_traits.translate(static_cast<charT>(i), this->m_icase) | |
536 | == this->m_traits.translate(first->first, this->m_icase)) | |
537 | result->_map[i] = true; | |
538 | } | |
539 | ++first; | |
540 | } | |
541 | // | |
542 | // OK now handle ranges: | |
543 | // | |
544 | first = char_set.ranges_begin(); | |
545 | last = char_set.ranges_end(); | |
546 | while(first != last) | |
547 | { | |
548 | // first grab the endpoints of the range: | |
549 | charT c1 = this->m_traits.translate(first->first, this->m_icase); | |
550 | ++first; | |
551 | charT c2 = this->m_traits.translate(first->first, this->m_icase); | |
552 | ++first; | |
553 | // different actions now depending upon whether collation is turned on: | |
554 | if(flags() & regex_constants::collate) | |
555 | { | |
556 | // we need to transform our range into sort keys: | |
557 | charT c3[2] = { c1, charT(0), }; | |
558 | string_type s1 = this->m_traits.transform(c3, c3+1); | |
559 | c3[0] = c2; | |
560 | string_type s2 = this->m_traits.transform(c3, c3+1); | |
561 | if(s1 > s2) | |
562 | { | |
563 | // Oops error: | |
564 | return 0; | |
565 | } | |
566 | BOOST_ASSERT(c3[1] == charT(0)); | |
567 | for(unsigned i = 0; i < (1u << CHAR_BIT); ++i) | |
568 | { | |
569 | c3[0] = static_cast<charT>(i); | |
570 | string_type s3 = this->m_traits.transform(c3, c3 +1); | |
571 | if((s1 <= s3) && (s3 <= s2)) | |
572 | result->_map[i] = true; | |
573 | } | |
574 | } | |
575 | else | |
576 | { | |
577 | if(char_less(c2, c1)) | |
578 | { | |
579 | // Oops error: | |
580 | return 0; | |
581 | } | |
582 | // everything in range matches: | |
583 | std::memset(result->_map + static_cast<unsigned char>(c1), true, 1 + static_cast<unsigned char>(c2) - static_cast<unsigned char>(c1)); | |
584 | } | |
585 | } | |
586 | // | |
587 | // and now the classes: | |
588 | // | |
589 | typedef typename traits::char_class_type m_type; | |
590 | m_type m = char_set.classes(); | |
591 | if(flags() & regbase::icase) | |
592 | { | |
593 | // adjust m as needed: | |
594 | if(((m & m_lower_mask) == m_lower_mask) || ((m & m_upper_mask) == m_upper_mask)) | |
595 | m |= m_alpha_mask; | |
596 | } | |
597 | if(m != 0) | |
598 | { | |
599 | for(unsigned i = 0; i < (1u << CHAR_BIT); ++i) | |
600 | { | |
601 | if(this->m_traits.isctype(static_cast<charT>(i), m)) | |
602 | result->_map[i] = true; | |
603 | } | |
604 | } | |
605 | // | |
606 | // and now the negated classes: | |
607 | // | |
608 | m = char_set.negated_classes(); | |
609 | if(flags() & regbase::icase) | |
610 | { | |
611 | // adjust m as needed: | |
612 | if(((m & m_lower_mask) == m_lower_mask) || ((m & m_upper_mask) == m_upper_mask)) | |
613 | m |= m_alpha_mask; | |
614 | } | |
615 | if(m != 0) | |
616 | { | |
617 | for(unsigned i = 0; i < (1u << CHAR_BIT); ++i) | |
618 | { | |
619 | if(0 == this->m_traits.isctype(static_cast<charT>(i), m)) | |
620 | result->_map[i] = true; | |
621 | } | |
622 | } | |
623 | // | |
624 | // now process the equivalence classes: | |
625 | // | |
626 | first = char_set.equivalents_begin(); | |
627 | last = char_set.equivalents_end(); | |
628 | while(first != last) | |
629 | { | |
630 | string_type s; | |
631 | BOOST_ASSERT(static_cast<charT>(0) == first->second); | |
632 | s = m_traits.transform_primary(&first->first, &first->first+1); | |
633 | if(s.empty()) | |
634 | return 0; // invalid or unsupported equivalence class | |
635 | for(unsigned i = 0; i < (1u << CHAR_BIT); ++i) | |
636 | { | |
637 | charT c[2] = { (static_cast<charT>(i)), charT(0), }; | |
638 | string_type s2 = this->m_traits.transform_primary(c, c+1); | |
639 | if(s == s2) | |
640 | result->_map[i] = true; | |
641 | } | |
642 | ++first; | |
643 | } | |
644 | if(negate) | |
645 | { | |
646 | for(unsigned i = 0; i < (1u << CHAR_BIT); ++i) | |
647 | { | |
648 | result->_map[i] = !(result->_map[i]); | |
649 | } | |
650 | } | |
651 | return result; | |
652 | } | |
653 | ||
654 | template <class charT, class traits> | |
655 | void basic_regex_creator<charT, traits>::finalize(const charT* p1, const charT* p2) | |
656 | { | |
657 | if(this->m_pdata->m_status) | |
658 | return; | |
659 | // we've added all the states we need, now finish things off. | |
660 | // start by adding a terminating state: | |
661 | append_state(syntax_element_match); | |
662 | // extend storage to store original expression: | |
663 | std::ptrdiff_t len = p2 - p1; | |
664 | m_pdata->m_expression_len = len; | |
665 | charT* ps = static_cast<charT*>(m_pdata->m_data.extend(sizeof(charT) * (1 + (p2 - p1)))); | |
666 | m_pdata->m_expression = ps; | |
667 | BOOST_REGEX_DETAIL_NS::copy(p1, p2, ps); | |
668 | ps[p2 - p1] = 0; | |
669 | // fill in our other data... | |
670 | // successful parsing implies a zero status: | |
671 | m_pdata->m_status = 0; | |
672 | // get the first state of the machine: | |
673 | m_pdata->m_first_state = static_cast<re_syntax_base*>(m_pdata->m_data.data()); | |
674 | // fixup pointers in the machine: | |
675 | fixup_pointers(m_pdata->m_first_state); | |
676 | if(m_has_recursions) | |
677 | { | |
678 | m_pdata->m_has_recursions = true; | |
679 | fixup_recursions(m_pdata->m_first_state); | |
680 | if(this->m_pdata->m_status) | |
681 | return; | |
682 | } | |
683 | else | |
684 | m_pdata->m_has_recursions = false; | |
685 | // create nested startmaps: | |
686 | create_startmaps(m_pdata->m_first_state); | |
687 | // create main startmap: | |
688 | std::memset(m_pdata->m_startmap, 0, sizeof(m_pdata->m_startmap)); | |
689 | m_pdata->m_can_be_null = 0; | |
690 | ||
691 | m_bad_repeats = 0; | |
692 | if(m_has_recursions) | |
693 | m_recursion_checks.assign(1 + m_pdata->m_mark_count, false); | |
694 | create_startmap(m_pdata->m_first_state, m_pdata->m_startmap, &(m_pdata->m_can_be_null), mask_all); | |
695 | // get the restart type: | |
696 | m_pdata->m_restart_type = get_restart_type(m_pdata->m_first_state); | |
697 | // optimise a leading repeat if there is one: | |
698 | probe_leading_repeat(m_pdata->m_first_state); | |
699 | } | |
700 | ||
701 | template <class charT, class traits> | |
702 | void basic_regex_creator<charT, traits>::fixup_pointers(re_syntax_base* state) | |
703 | { | |
704 | while(state) | |
705 | { | |
706 | switch(state->type) | |
707 | { | |
708 | case syntax_element_recurse: | |
709 | m_has_recursions = true; | |
710 | if(state->next.i) | |
711 | state->next.p = getaddress(state->next.i, state); | |
712 | else | |
713 | state->next.p = 0; | |
714 | break; | |
715 | case syntax_element_rep: | |
716 | case syntax_element_dot_rep: | |
717 | case syntax_element_char_rep: | |
718 | case syntax_element_short_set_rep: | |
719 | case syntax_element_long_set_rep: | |
720 | // set the state_id of this repeat: | |
721 | static_cast<re_repeat*>(state)->state_id = m_repeater_id++; | |
722 | BOOST_FALLTHROUGH; | |
723 | case syntax_element_alt: | |
724 | std::memset(static_cast<re_alt*>(state)->_map, 0, sizeof(static_cast<re_alt*>(state)->_map)); | |
725 | static_cast<re_alt*>(state)->can_be_null = 0; | |
726 | BOOST_FALLTHROUGH; | |
727 | case syntax_element_jump: | |
728 | static_cast<re_jump*>(state)->alt.p = getaddress(static_cast<re_jump*>(state)->alt.i, state); | |
729 | BOOST_FALLTHROUGH; | |
730 | default: | |
731 | if(state->next.i) | |
732 | state->next.p = getaddress(state->next.i, state); | |
733 | else | |
734 | state->next.p = 0; | |
735 | } | |
736 | state = state->next.p; | |
737 | } | |
738 | } | |
739 | ||
740 | template <class charT, class traits> | |
741 | void basic_regex_creator<charT, traits>::fixup_recursions(re_syntax_base* state) | |
742 | { | |
743 | re_syntax_base* base = state; | |
744 | while(state) | |
745 | { | |
746 | switch(state->type) | |
747 | { | |
748 | case syntax_element_assert_backref: | |
749 | { | |
750 | // just check that the index is valid: | |
751 | int idx = static_cast<const re_brace*>(state)->index; | |
752 | if(idx < 0) | |
753 | { | |
754 | idx = -idx-1; | |
755 | if(idx >= 10000) | |
756 | { | |
757 | idx = m_pdata->get_id(idx); | |
758 | if(idx <= 0) | |
759 | { | |
760 | // check of sub-expression that doesn't exist: | |
761 | if(0 == this->m_pdata->m_status) // update the error code if not already set | |
762 | this->m_pdata->m_status = boost::regex_constants::error_bad_pattern; | |
763 | // | |
764 | // clear the expression, we should be empty: | |
765 | // | |
766 | this->m_pdata->m_expression = 0; | |
767 | this->m_pdata->m_expression_len = 0; | |
768 | // | |
769 | // and throw if required: | |
770 | // | |
771 | if(0 == (this->flags() & regex_constants::no_except)) | |
772 | { | |
773 | std::string message = "Encountered a forward reference to a marked sub-expression that does not exist."; | |
774 | boost::regex_error e(message, boost::regex_constants::error_bad_pattern, 0); | |
775 | e.raise(); | |
776 | } | |
777 | } | |
778 | } | |
779 | } | |
780 | } | |
781 | break; | |
782 | case syntax_element_recurse: | |
783 | { | |
784 | bool ok = false; | |
785 | re_syntax_base* p = base; | |
786 | std::ptrdiff_t idx = static_cast<re_jump*>(state)->alt.i; | |
787 | if(idx > 10000) | |
788 | { | |
789 | // | |
790 | // There may be more than one capture group with this hash, just do what Perl | |
791 | // does and recurse to the leftmost: | |
792 | // | |
793 | idx = m_pdata->get_id(static_cast<int>(idx)); | |
794 | } | |
795 | while(p) | |
796 | { | |
797 | if((p->type == syntax_element_startmark) && (static_cast<re_brace*>(p)->index == idx)) | |
798 | { | |
799 | // | |
800 | // We've found the target of the recursion, set the jump target: | |
801 | // | |
802 | static_cast<re_jump*>(state)->alt.p = p; | |
803 | ok = true; | |
804 | // | |
805 | // Now scan the target for nested repeats: | |
806 | // | |
807 | p = p->next.p; | |
808 | int next_rep_id = 0; | |
809 | while(p) | |
810 | { | |
811 | switch(p->type) | |
812 | { | |
813 | case syntax_element_rep: | |
814 | case syntax_element_dot_rep: | |
815 | case syntax_element_char_rep: | |
816 | case syntax_element_short_set_rep: | |
817 | case syntax_element_long_set_rep: | |
818 | next_rep_id = static_cast<re_repeat*>(p)->state_id; | |
819 | break; | |
820 | case syntax_element_endmark: | |
821 | if(static_cast<const re_brace*>(p)->index == idx) | |
822 | next_rep_id = -1; | |
823 | break; | |
824 | default: | |
825 | break; | |
826 | } | |
827 | if(next_rep_id) | |
828 | break; | |
829 | p = p->next.p; | |
830 | } | |
831 | if(next_rep_id > 0) | |
832 | { | |
833 | static_cast<re_recurse*>(state)->state_id = next_rep_id - 1; | |
834 | } | |
835 | ||
836 | break; | |
837 | } | |
838 | p = p->next.p; | |
839 | } | |
840 | if(!ok) | |
841 | { | |
842 | // recursion to sub-expression that doesn't exist: | |
843 | if(0 == this->m_pdata->m_status) // update the error code if not already set | |
844 | this->m_pdata->m_status = boost::regex_constants::error_bad_pattern; | |
845 | // | |
846 | // clear the expression, we should be empty: | |
847 | // | |
848 | this->m_pdata->m_expression = 0; | |
849 | this->m_pdata->m_expression_len = 0; | |
850 | // | |
851 | // and throw if required: | |
852 | // | |
853 | if(0 == (this->flags() & regex_constants::no_except)) | |
854 | { | |
855 | std::string message = "Encountered a forward reference to a recursive sub-expression that does not exist."; | |
856 | boost::regex_error e(message, boost::regex_constants::error_bad_pattern, 0); | |
857 | e.raise(); | |
858 | } | |
859 | } | |
860 | } | |
861 | break; | |
862 | default: | |
863 | break; | |
864 | } | |
865 | state = state->next.p; | |
866 | } | |
867 | } | |
868 | ||
869 | template <class charT, class traits> | |
870 | void basic_regex_creator<charT, traits>::create_startmaps(re_syntax_base* state) | |
871 | { | |
872 | // non-recursive implementation: | |
873 | // create the last map in the machine first, so that earlier maps | |
874 | // can make use of the result... | |
875 | // | |
876 | // This was originally a recursive implementation, but that caused stack | |
877 | // overflows with complex expressions on small stacks (think COM+). | |
878 | ||
879 | // start by saving the case setting: | |
880 | bool l_icase = m_icase; | |
881 | std::vector<std::pair<bool, re_syntax_base*> > v; | |
882 | ||
883 | while(state) | |
884 | { | |
885 | switch(state->type) | |
886 | { | |
887 | case syntax_element_toggle_case: | |
888 | // we need to track case changes here: | |
889 | m_icase = static_cast<re_case*>(state)->icase; | |
890 | state = state->next.p; | |
891 | continue; | |
892 | case syntax_element_alt: | |
893 | case syntax_element_rep: | |
894 | case syntax_element_dot_rep: | |
895 | case syntax_element_char_rep: | |
896 | case syntax_element_short_set_rep: | |
897 | case syntax_element_long_set_rep: | |
898 | // just push the state onto our stack for now: | |
899 | v.push_back(std::pair<bool, re_syntax_base*>(m_icase, state)); | |
900 | state = state->next.p; | |
901 | break; | |
902 | case syntax_element_backstep: | |
903 | // we need to calculate how big the backstep is: | |
904 | static_cast<re_brace*>(state)->index | |
905 | = this->calculate_backstep(state->next.p); | |
906 | if(static_cast<re_brace*>(state)->index < 0) | |
907 | { | |
908 | // Oops error: | |
909 | if(0 == this->m_pdata->m_status) // update the error code if not already set | |
910 | this->m_pdata->m_status = boost::regex_constants::error_bad_pattern; | |
911 | // | |
912 | // clear the expression, we should be empty: | |
913 | // | |
914 | this->m_pdata->m_expression = 0; | |
915 | this->m_pdata->m_expression_len = 0; | |
916 | // | |
917 | // and throw if required: | |
918 | // | |
919 | if(0 == (this->flags() & regex_constants::no_except)) | |
920 | { | |
921 | std::string message = "Invalid lookbehind assertion encountered in the regular expression."; | |
922 | boost::regex_error e(message, boost::regex_constants::error_bad_pattern, 0); | |
923 | e.raise(); | |
924 | } | |
925 | } | |
926 | BOOST_FALLTHROUGH; | |
927 | default: | |
928 | state = state->next.p; | |
929 | } | |
930 | } | |
931 | ||
932 | // now work through our list, building all the maps as we go: | |
933 | while(v.size()) | |
934 | { | |
935 | // Initialize m_recursion_checks if we need it: | |
936 | if(m_has_recursions) | |
937 | m_recursion_checks.assign(1 + m_pdata->m_mark_count, false); | |
938 | ||
939 | const std::pair<bool, re_syntax_base*>& p = v.back(); | |
940 | m_icase = p.first; | |
941 | state = p.second; | |
942 | v.pop_back(); | |
943 | ||
944 | // Build maps: | |
945 | m_bad_repeats = 0; | |
946 | create_startmap(state->next.p, static_cast<re_alt*>(state)->_map, &static_cast<re_alt*>(state)->can_be_null, mask_take); | |
947 | m_bad_repeats = 0; | |
948 | ||
949 | if(m_has_recursions) | |
950 | m_recursion_checks.assign(1 + m_pdata->m_mark_count, false); | |
951 | create_startmap(static_cast<re_alt*>(state)->alt.p, static_cast<re_alt*>(state)->_map, &static_cast<re_alt*>(state)->can_be_null, mask_skip); | |
952 | // adjust the type of the state to allow for faster matching: | |
953 | state->type = this->get_repeat_type(state); | |
954 | } | |
955 | // restore case sensitivity: | |
956 | m_icase = l_icase; | |
957 | } | |
958 | ||
959 | template <class charT, class traits> | |
960 | int basic_regex_creator<charT, traits>::calculate_backstep(re_syntax_base* state) | |
961 | { | |
962 | typedef typename traits::char_class_type m_type; | |
963 | int result = 0; | |
964 | while(state) | |
965 | { | |
966 | switch(state->type) | |
967 | { | |
968 | case syntax_element_startmark: | |
969 | if((static_cast<re_brace*>(state)->index == -1) | |
970 | || (static_cast<re_brace*>(state)->index == -2)) | |
971 | { | |
972 | state = static_cast<re_jump*>(state->next.p)->alt.p->next.p; | |
973 | continue; | |
974 | } | |
975 | else if(static_cast<re_brace*>(state)->index == -3) | |
976 | { | |
977 | state = state->next.p->next.p; | |
978 | continue; | |
979 | } | |
980 | break; | |
981 | case syntax_element_endmark: | |
982 | if((static_cast<re_brace*>(state)->index == -1) | |
983 | || (static_cast<re_brace*>(state)->index == -2)) | |
984 | return result; | |
985 | break; | |
986 | case syntax_element_literal: | |
987 | result += static_cast<re_literal*>(state)->length; | |
988 | break; | |
989 | case syntax_element_wild: | |
990 | case syntax_element_set: | |
991 | result += 1; | |
992 | break; | |
993 | case syntax_element_dot_rep: | |
994 | case syntax_element_char_rep: | |
995 | case syntax_element_short_set_rep: | |
996 | case syntax_element_backref: | |
997 | case syntax_element_rep: | |
998 | case syntax_element_combining: | |
999 | case syntax_element_long_set_rep: | |
1000 | case syntax_element_backstep: | |
1001 | { | |
1002 | re_repeat* rep = static_cast<re_repeat *>(state); | |
1003 | // adjust the type of the state to allow for faster matching: | |
1004 | state->type = this->get_repeat_type(state); | |
1005 | if((state->type == syntax_element_dot_rep) | |
1006 | || (state->type == syntax_element_char_rep) | |
1007 | || (state->type == syntax_element_short_set_rep)) | |
1008 | { | |
1009 | if(rep->max != rep->min) | |
1010 | return -1; | |
1011 | result += static_cast<int>(rep->min); | |
1012 | state = rep->alt.p; | |
1013 | continue; | |
1014 | } | |
1015 | else if(state->type == syntax_element_long_set_rep) | |
1016 | { | |
1017 | BOOST_ASSERT(rep->next.p->type == syntax_element_long_set); | |
1018 | if(static_cast<re_set_long<m_type>*>(rep->next.p)->singleton == 0) | |
1019 | return -1; | |
1020 | if(rep->max != rep->min) | |
1021 | return -1; | |
1022 | result += static_cast<int>(rep->min); | |
1023 | state = rep->alt.p; | |
1024 | continue; | |
1025 | } | |
1026 | } | |
1027 | return -1; | |
1028 | case syntax_element_long_set: | |
1029 | if(static_cast<re_set_long<m_type>*>(state)->singleton == 0) | |
1030 | return -1; | |
1031 | result += 1; | |
1032 | break; | |
1033 | case syntax_element_jump: | |
1034 | state = static_cast<re_jump*>(state)->alt.p; | |
1035 | continue; | |
1036 | case syntax_element_alt: | |
1037 | { | |
1038 | int r1 = calculate_backstep(state->next.p); | |
1039 | int r2 = calculate_backstep(static_cast<re_alt*>(state)->alt.p); | |
1040 | if((r1 < 0) || (r1 != r2)) | |
1041 | return -1; | |
1042 | return result + r1; | |
1043 | } | |
1044 | default: | |
1045 | break; | |
1046 | } | |
1047 | state = state->next.p; | |
1048 | } | |
1049 | return -1; | |
1050 | } | |
1051 | ||
1052 | template <class charT, class traits> | |
1053 | void basic_regex_creator<charT, traits>::create_startmap(re_syntax_base* state, unsigned char* l_map, unsigned int* pnull, unsigned char mask) | |
1054 | { | |
1055 | int not_last_jump = 1; | |
1056 | re_syntax_base* recursion_start = 0; | |
1057 | int recursion_sub = 0; | |
1058 | re_syntax_base* recursion_restart = 0; | |
1059 | ||
1060 | // track case sensitivity: | |
1061 | bool l_icase = m_icase; | |
1062 | ||
1063 | while(state) | |
1064 | { | |
1065 | switch(state->type) | |
1066 | { | |
1067 | case syntax_element_toggle_case: | |
1068 | l_icase = static_cast<re_case*>(state)->icase; | |
1069 | state = state->next.p; | |
1070 | break; | |
1071 | case syntax_element_literal: | |
1072 | { | |
1073 | // don't set anything in *pnull, set each element in l_map | |
1074 | // that could match the first character in the literal: | |
1075 | if(l_map) | |
1076 | { | |
1077 | l_map[0] |= mask_init; | |
1078 | charT first_char = *static_cast<charT*>(static_cast<void*>(static_cast<re_literal*>(state) + 1)); | |
1079 | for(unsigned int i = 0; i < (1u << CHAR_BIT); ++i) | |
1080 | { | |
1081 | if(m_traits.translate(static_cast<charT>(i), l_icase) == first_char) | |
1082 | l_map[i] |= mask; | |
1083 | } | |
1084 | } | |
1085 | return; | |
1086 | } | |
1087 | case syntax_element_end_line: | |
1088 | { | |
1089 | // next character must be a line separator (if there is one): | |
1090 | if(l_map) | |
1091 | { | |
1092 | l_map[0] |= mask_init; | |
1093 | l_map[static_cast<unsigned>('\n')] |= mask; | |
1094 | l_map[static_cast<unsigned>('\r')] |= mask; | |
1095 | l_map[static_cast<unsigned>('\f')] |= mask; | |
1096 | l_map[0x85] |= mask; | |
1097 | } | |
1098 | // now figure out if we can match a NULL string at this point: | |
1099 | if(pnull) | |
1100 | create_startmap(state->next.p, 0, pnull, mask); | |
1101 | return; | |
1102 | } | |
1103 | case syntax_element_recurse: | |
1104 | { | |
1105 | if(state->type == syntax_element_startmark) | |
1106 | recursion_sub = static_cast<re_brace*>(state)->index; | |
1107 | else | |
1108 | recursion_sub = 0; | |
1109 | if(m_recursion_checks[recursion_sub]) | |
1110 | { | |
1111 | // Infinite recursion!! | |
1112 | if(0 == this->m_pdata->m_status) // update the error code if not already set | |
1113 | this->m_pdata->m_status = boost::regex_constants::error_bad_pattern; | |
1114 | // | |
1115 | // clear the expression, we should be empty: | |
1116 | // | |
1117 | this->m_pdata->m_expression = 0; | |
1118 | this->m_pdata->m_expression_len = 0; | |
1119 | // | |
1120 | // and throw if required: | |
1121 | // | |
1122 | if(0 == (this->flags() & regex_constants::no_except)) | |
1123 | { | |
1124 | std::string message = "Encountered an infinite recursion."; | |
1125 | boost::regex_error e(message, boost::regex_constants::error_bad_pattern, 0); | |
1126 | e.raise(); | |
1127 | } | |
1128 | } | |
1129 | else if(recursion_start == 0) | |
1130 | { | |
1131 | recursion_start = state; | |
1132 | recursion_restart = state->next.p; | |
1133 | state = static_cast<re_jump*>(state)->alt.p; | |
1134 | m_recursion_checks[recursion_sub] = true; | |
1135 | break; | |
1136 | } | |
1137 | m_recursion_checks[recursion_sub] = true; | |
1138 | // can't handle nested recursion here... | |
1139 | BOOST_FALLTHROUGH; | |
1140 | } | |
1141 | case syntax_element_backref: | |
1142 | // can be null, and any character can match: | |
1143 | if(pnull) | |
1144 | *pnull |= mask; | |
1145 | BOOST_FALLTHROUGH; | |
1146 | case syntax_element_wild: | |
1147 | { | |
1148 | // can't be null, any character can match: | |
1149 | set_all_masks(l_map, mask); | |
1150 | return; | |
1151 | } | |
1152 | case syntax_element_accept: | |
1153 | case syntax_element_match: | |
1154 | { | |
1155 | // must be null, any character can match: | |
1156 | set_all_masks(l_map, mask); | |
1157 | if(pnull) | |
1158 | *pnull |= mask; | |
1159 | return; | |
1160 | } | |
1161 | case syntax_element_word_start: | |
1162 | { | |
1163 | // recurse, then AND with all the word characters: | |
1164 | create_startmap(state->next.p, l_map, pnull, mask); | |
1165 | if(l_map) | |
1166 | { | |
1167 | l_map[0] |= mask_init; | |
1168 | for(unsigned int i = 0; i < (1u << CHAR_BIT); ++i) | |
1169 | { | |
1170 | if(!m_traits.isctype(static_cast<charT>(i), m_word_mask)) | |
1171 | l_map[i] &= static_cast<unsigned char>(~mask); | |
1172 | } | |
1173 | } | |
1174 | return; | |
1175 | } | |
1176 | case syntax_element_word_end: | |
1177 | { | |
1178 | // recurse, then AND with all the word characters: | |
1179 | create_startmap(state->next.p, l_map, pnull, mask); | |
1180 | if(l_map) | |
1181 | { | |
1182 | l_map[0] |= mask_init; | |
1183 | for(unsigned int i = 0; i < (1u << CHAR_BIT); ++i) | |
1184 | { | |
1185 | if(m_traits.isctype(static_cast<charT>(i), m_word_mask)) | |
1186 | l_map[i] &= static_cast<unsigned char>(~mask); | |
1187 | } | |
1188 | } | |
1189 | return; | |
1190 | } | |
1191 | case syntax_element_buffer_end: | |
1192 | { | |
1193 | // we *must be null* : | |
1194 | if(pnull) | |
1195 | *pnull |= mask; | |
1196 | return; | |
1197 | } | |
1198 | case syntax_element_long_set: | |
1199 | if(l_map) | |
1200 | { | |
1201 | typedef typename traits::char_class_type m_type; | |
1202 | if(static_cast<re_set_long<m_type>*>(state)->singleton) | |
1203 | { | |
1204 | l_map[0] |= mask_init; | |
1205 | for(unsigned int i = 0; i < (1u << CHAR_BIT); ++i) | |
1206 | { | |
1207 | charT c = static_cast<charT>(i); | |
1208 | if(&c != re_is_set_member(&c, &c + 1, static_cast<re_set_long<m_type>*>(state), *m_pdata, l_icase)) | |
1209 | l_map[i] |= mask; | |
1210 | } | |
1211 | } | |
1212 | else | |
1213 | set_all_masks(l_map, mask); | |
1214 | } | |
1215 | return; | |
1216 | case syntax_element_set: | |
1217 | if(l_map) | |
1218 | { | |
1219 | l_map[0] |= mask_init; | |
1220 | for(unsigned int i = 0; i < (1u << CHAR_BIT); ++i) | |
1221 | { | |
1222 | if(static_cast<re_set*>(state)->_map[ | |
1223 | static_cast<unsigned char>(m_traits.translate(static_cast<charT>(i), l_icase))]) | |
1224 | l_map[i] |= mask; | |
1225 | } | |
1226 | } | |
1227 | return; | |
1228 | case syntax_element_jump: | |
1229 | // take the jump: | |
1230 | state = static_cast<re_alt*>(state)->alt.p; | |
1231 | not_last_jump = -1; | |
1232 | break; | |
1233 | case syntax_element_alt: | |
1234 | case syntax_element_rep: | |
1235 | case syntax_element_dot_rep: | |
1236 | case syntax_element_char_rep: | |
1237 | case syntax_element_short_set_rep: | |
1238 | case syntax_element_long_set_rep: | |
1239 | { | |
1240 | re_alt* rep = static_cast<re_alt*>(state); | |
1241 | if(rep->_map[0] & mask_init) | |
1242 | { | |
1243 | if(l_map) | |
1244 | { | |
1245 | // copy previous results: | |
1246 | l_map[0] |= mask_init; | |
1247 | for(unsigned int i = 0; i <= UCHAR_MAX; ++i) | |
1248 | { | |
1249 | if(rep->_map[i] & mask_any) | |
1250 | l_map[i] |= mask; | |
1251 | } | |
1252 | } | |
1253 | if(pnull) | |
1254 | { | |
1255 | if(rep->can_be_null & mask_any) | |
1256 | *pnull |= mask; | |
1257 | } | |
1258 | } | |
1259 | else | |
1260 | { | |
1261 | // we haven't created a startmap for this alternative yet | |
1262 | // so take the union of the two options: | |
1263 | if(is_bad_repeat(state)) | |
1264 | { | |
1265 | set_all_masks(l_map, mask); | |
1266 | if(pnull) | |
1267 | *pnull |= mask; | |
1268 | return; | |
1269 | } | |
1270 | set_bad_repeat(state); | |
1271 | create_startmap(state->next.p, l_map, pnull, mask); | |
1272 | if((state->type == syntax_element_alt) | |
1273 | || (static_cast<re_repeat*>(state)->min == 0) | |
1274 | || (not_last_jump == 0)) | |
1275 | create_startmap(rep->alt.p, l_map, pnull, mask); | |
1276 | } | |
1277 | } | |
1278 | return; | |
1279 | case syntax_element_soft_buffer_end: | |
1280 | // match newline or null: | |
1281 | if(l_map) | |
1282 | { | |
1283 | l_map[0] |= mask_init; | |
1284 | l_map[static_cast<unsigned>('\n')] |= mask; | |
1285 | l_map[static_cast<unsigned>('\r')] |= mask; | |
1286 | } | |
1287 | if(pnull) | |
1288 | *pnull |= mask; | |
1289 | return; | |
1290 | case syntax_element_endmark: | |
1291 | // need to handle independent subs as a special case: | |
1292 | if(static_cast<re_brace*>(state)->index < 0) | |
1293 | { | |
1294 | // can be null, any character can match: | |
1295 | set_all_masks(l_map, mask); | |
1296 | if(pnull) | |
1297 | *pnull |= mask; | |
1298 | return; | |
1299 | } | |
1300 | else if(recursion_start && (recursion_sub != 0) && (recursion_sub == static_cast<re_brace*>(state)->index)) | |
1301 | { | |
1302 | // recursion termination: | |
1303 | recursion_start = 0; | |
1304 | state = recursion_restart; | |
1305 | break; | |
1306 | } | |
1307 | ||
1308 | // | |
1309 | // Normally we just go to the next state... but if this sub-expression is | |
1310 | // the target of a recursion, then we might be ending a recursion, in which | |
1311 | // case we should check whatever follows that recursion, as well as whatever | |
1312 | // follows this state: | |
1313 | // | |
1314 | if(m_pdata->m_has_recursions && static_cast<re_brace*>(state)->index) | |
1315 | { | |
1316 | bool ok = false; | |
1317 | re_syntax_base* p = m_pdata->m_first_state; | |
1318 | while(p) | |
1319 | { | |
1320 | if(p->type == syntax_element_recurse) | |
1321 | { | |
1322 | re_brace* p2 = static_cast<re_brace*>(static_cast<re_jump*>(p)->alt.p); | |
1323 | if((p2->type == syntax_element_startmark) && (p2->index == static_cast<re_brace*>(state)->index)) | |
1324 | { | |
1325 | ok = true; | |
1326 | break; | |
1327 | } | |
1328 | } | |
1329 | p = p->next.p; | |
1330 | } | |
1331 | if(ok) | |
1332 | { | |
1333 | create_startmap(p->next.p, l_map, pnull, mask); | |
1334 | } | |
1335 | } | |
1336 | state = state->next.p; | |
1337 | break; | |
1338 | ||
1339 | case syntax_element_commit: | |
1340 | set_all_masks(l_map, mask); | |
1341 | // Continue scanning so we can figure out whether we can be null: | |
1342 | state = state->next.p; | |
1343 | break; | |
1344 | case syntax_element_startmark: | |
1345 | // need to handle independent subs as a special case: | |
1346 | if(static_cast<re_brace*>(state)->index == -3) | |
1347 | { | |
1348 | state = state->next.p->next.p; | |
1349 | break; | |
1350 | } | |
1351 | BOOST_FALLTHROUGH; | |
1352 | default: | |
1353 | state = state->next.p; | |
1354 | } | |
1355 | ++not_last_jump; | |
1356 | } | |
1357 | } | |
1358 | ||
1359 | template <class charT, class traits> | |
1360 | unsigned basic_regex_creator<charT, traits>::get_restart_type(re_syntax_base* state) | |
1361 | { | |
1362 | // | |
1363 | // find out how the machine starts, so we can optimise the search: | |
1364 | // | |
1365 | while(state) | |
1366 | { | |
1367 | switch(state->type) | |
1368 | { | |
1369 | case syntax_element_startmark: | |
1370 | case syntax_element_endmark: | |
1371 | state = state->next.p; | |
1372 | continue; | |
1373 | case syntax_element_start_line: | |
1374 | return regbase::restart_line; | |
1375 | case syntax_element_word_start: | |
1376 | return regbase::restart_word; | |
1377 | case syntax_element_buffer_start: | |
1378 | return regbase::restart_buf; | |
1379 | case syntax_element_restart_continue: | |
1380 | return regbase::restart_continue; | |
1381 | default: | |
1382 | state = 0; | |
1383 | continue; | |
1384 | } | |
1385 | } | |
1386 | return regbase::restart_any; | |
1387 | } | |
1388 | ||
1389 | template <class charT, class traits> | |
1390 | void basic_regex_creator<charT, traits>::set_all_masks(unsigned char* bits, unsigned char mask) | |
1391 | { | |
1392 | // | |
1393 | // set mask in all of bits elements, | |
1394 | // if bits[0] has mask_init not set then we can | |
1395 | // optimise this to a call to memset: | |
1396 | // | |
1397 | if(bits) | |
1398 | { | |
1399 | if(bits[0] == 0) | |
1400 | (std::memset)(bits, mask, 1u << CHAR_BIT); | |
1401 | else | |
1402 | { | |
1403 | for(unsigned i = 0; i < (1u << CHAR_BIT); ++i) | |
1404 | bits[i] |= mask; | |
1405 | } | |
1406 | bits[0] |= mask_init; | |
1407 | } | |
1408 | } | |
1409 | ||
1410 | template <class charT, class traits> | |
1411 | bool basic_regex_creator<charT, traits>::is_bad_repeat(re_syntax_base* pt) | |
1412 | { | |
1413 | switch(pt->type) | |
1414 | { | |
1415 | case syntax_element_rep: | |
1416 | case syntax_element_dot_rep: | |
1417 | case syntax_element_char_rep: | |
1418 | case syntax_element_short_set_rep: | |
1419 | case syntax_element_long_set_rep: | |
1420 | { | |
1421 | unsigned state_id = static_cast<re_repeat*>(pt)->state_id; | |
1422 | if(state_id > sizeof(m_bad_repeats) * CHAR_BIT) | |
1423 | return true; // run out of bits, assume we can't traverse this one. | |
1424 | static const boost::uintmax_t one = 1uL; | |
1425 | return m_bad_repeats & (one << state_id); | |
1426 | } | |
1427 | default: | |
1428 | return false; | |
1429 | } | |
1430 | } | |
1431 | ||
1432 | template <class charT, class traits> | |
1433 | void basic_regex_creator<charT, traits>::set_bad_repeat(re_syntax_base* pt) | |
1434 | { | |
1435 | switch(pt->type) | |
1436 | { | |
1437 | case syntax_element_rep: | |
1438 | case syntax_element_dot_rep: | |
1439 | case syntax_element_char_rep: | |
1440 | case syntax_element_short_set_rep: | |
1441 | case syntax_element_long_set_rep: | |
1442 | { | |
1443 | unsigned state_id = static_cast<re_repeat*>(pt)->state_id; | |
1444 | static const boost::uintmax_t one = 1uL; | |
1445 | if(state_id <= sizeof(m_bad_repeats) * CHAR_BIT) | |
1446 | m_bad_repeats |= (one << state_id); | |
1447 | } | |
1448 | break; | |
1449 | default: | |
1450 | break; | |
1451 | } | |
1452 | } | |
1453 | ||
1454 | template <class charT, class traits> | |
1455 | syntax_element_type basic_regex_creator<charT, traits>::get_repeat_type(re_syntax_base* state) | |
1456 | { | |
1457 | typedef typename traits::char_class_type m_type; | |
1458 | if(state->type == syntax_element_rep) | |
1459 | { | |
1460 | // check to see if we are repeating a single state: | |
1461 | if(state->next.p->next.p->next.p == static_cast<re_alt*>(state)->alt.p) | |
1462 | { | |
1463 | switch(state->next.p->type) | |
1464 | { | |
1465 | case BOOST_REGEX_DETAIL_NS::syntax_element_wild: | |
1466 | return BOOST_REGEX_DETAIL_NS::syntax_element_dot_rep; | |
1467 | case BOOST_REGEX_DETAIL_NS::syntax_element_literal: | |
1468 | return BOOST_REGEX_DETAIL_NS::syntax_element_char_rep; | |
1469 | case BOOST_REGEX_DETAIL_NS::syntax_element_set: | |
1470 | return BOOST_REGEX_DETAIL_NS::syntax_element_short_set_rep; | |
1471 | case BOOST_REGEX_DETAIL_NS::syntax_element_long_set: | |
1472 | if(static_cast<BOOST_REGEX_DETAIL_NS::re_set_long<m_type>*>(state->next.p)->singleton) | |
1473 | return BOOST_REGEX_DETAIL_NS::syntax_element_long_set_rep; | |
1474 | break; | |
1475 | default: | |
1476 | break; | |
1477 | } | |
1478 | } | |
1479 | } | |
1480 | return state->type; | |
1481 | } | |
1482 | ||
1483 | template <class charT, class traits> | |
1484 | void basic_regex_creator<charT, traits>::probe_leading_repeat(re_syntax_base* state) | |
1485 | { | |
1486 | // enumerate our states, and see if we have a leading repeat | |
1487 | // for which failed search restarts can be optimised; | |
1488 | do | |
1489 | { | |
1490 | switch(state->type) | |
1491 | { | |
1492 | case syntax_element_startmark: | |
1493 | if(static_cast<re_brace*>(state)->index >= 0) | |
1494 | { | |
1495 | state = state->next.p; | |
1496 | continue; | |
1497 | } | |
1498 | if((static_cast<re_brace*>(state)->index == -1) | |
1499 | || (static_cast<re_brace*>(state)->index == -2)) | |
1500 | { | |
1501 | // skip past the zero width assertion: | |
1502 | state = static_cast<const re_jump*>(state->next.p)->alt.p->next.p; | |
1503 | continue; | |
1504 | } | |
1505 | if(static_cast<re_brace*>(state)->index == -3) | |
1506 | { | |
1507 | // Have to skip the leading jump state: | |
1508 | state = state->next.p->next.p; | |
1509 | continue; | |
1510 | } | |
1511 | return; | |
1512 | case syntax_element_endmark: | |
1513 | case syntax_element_start_line: | |
1514 | case syntax_element_end_line: | |
1515 | case syntax_element_word_boundary: | |
1516 | case syntax_element_within_word: | |
1517 | case syntax_element_word_start: | |
1518 | case syntax_element_word_end: | |
1519 | case syntax_element_buffer_start: | |
1520 | case syntax_element_buffer_end: | |
1521 | case syntax_element_restart_continue: | |
1522 | state = state->next.p; | |
1523 | break; | |
1524 | case syntax_element_dot_rep: | |
1525 | case syntax_element_char_rep: | |
1526 | case syntax_element_short_set_rep: | |
1527 | case syntax_element_long_set_rep: | |
1528 | if(this->m_has_backrefs == 0) | |
1529 | static_cast<re_repeat*>(state)->leading = true; | |
1530 | BOOST_FALLTHROUGH; | |
1531 | default: | |
1532 | return; | |
1533 | } | |
1534 | }while(state); | |
1535 | } | |
1536 | ||
1537 | ||
1538 | } // namespace BOOST_REGEX_DETAIL_NS | |
1539 | ||
1540 | } // namespace boost | |
1541 | ||
1542 | #ifdef BOOST_MSVC | |
1543 | # pragma warning(pop) | |
1544 | #endif | |
1545 | ||
1546 | #ifdef BOOST_MSVC | |
1547 | #pragma warning(push) | |
1548 | #pragma warning(disable: 4103) | |
1549 | #endif | |
1550 | #ifdef BOOST_HAS_ABI_HEADERS | |
1551 | # include BOOST_ABI_SUFFIX | |
1552 | #endif | |
1553 | #ifdef BOOST_MSVC | |
1554 | #pragma warning(pop) | |
1555 | #endif | |
1556 | ||
1557 | #endif | |
1558 |