1 // Copyright 2003-2009 The RE2 Authors. All Rights Reserved.
2 // Use of this source code is governed by a BSD-style
3 // license that can be found in the LICENSE file.
4
5 #ifndef RE2_RE2_H_
6 #define RE2_RE2_H_
7
8 // C++ interface to the re2 regular-expression library.
9 // RE2 supports Perl-style regular expressions (with extensions like
10 // \d, \w, \s, ...).
11 //
12 // -----------------------------------------------------------------------
13 // REGEXP SYNTAX:
14 //
15 // This module uses the re2 library and hence supports
16 // its syntax for regular expressions, which is similar to Perl's with
17 // some of the more complicated things thrown away. In particular,
18 // backreferences and generalized assertions are not available, nor is \Z.
19 //
20 // See https://github.com/google/re2/wiki/Syntax for the syntax
21 // supported by RE2, and a comparison with PCRE and PERL regexps.
22 //
23 // For those not familiar with Perl's regular expressions,
24 // here are some examples of the most commonly used extensions:
25 //
26 // "hello (\\w+) world" -- \w matches a "word" character
27 // "version (\\d+)" -- \d matches a digit
28 // "hello\\s+world" -- \s matches any whitespace character
29 // "\\b(\\w+)\\b" -- \b matches non-empty string at word boundary
30 // "(?i)hello" -- (?i) turns on case-insensitive matching
31 // "/\\*(.*?)\\*/" -- .*? matches . minimum no. of times possible
32 //
33 // The double backslashes are needed when writing C++ string literals.
34 // However, they should NOT be used when writing C++11 raw string literals:
35 //
36 // R"(hello (\w+) world)" -- \w matches a "word" character
37 // R"(version (\d+))" -- \d matches a digit
38 // R"(hello\s+world)" -- \s matches any whitespace character
39 // R"(\b(\w+)\b)" -- \b matches non-empty string at word boundary
40 // R"((?i)hello)" -- (?i) turns on case-insensitive matching
41 // R"(/\*(.*?)\*/)" -- .*? matches . minimum no. of times possible
42 //
43 // -----------------------------------------------------------------------
44 // MATCHING INTERFACE:
45 //
46 // The "FullMatch" operation checks that supplied text matches a
47 // supplied pattern exactly.
48 //
49 // Example: successful match
50 // CHECK(RE2::FullMatch("hello", "h.*o"));
51 //
52 // Example: unsuccessful match (requires full match):
53 // CHECK(!RE2::FullMatch("hello", "e"));
54 //
55 // -----------------------------------------------------------------------
56 // UTF-8 AND THE MATCHING INTERFACE:
57 //
58 // By default, the pattern and input text are interpreted as UTF-8.
59 // The RE2::Latin1 option causes them to be interpreted as Latin-1.
60 //
61 // Example:
62 // CHECK(RE2::FullMatch(utf8_string, RE2(utf8_pattern)));
63 // CHECK(RE2::FullMatch(latin1_string, RE2(latin1_pattern, RE2::Latin1)));
64 //
65 // -----------------------------------------------------------------------
66 // MATCHING WITH SUBSTRING EXTRACTION:
67 //
68 // You can supply extra pointer arguments to extract matched substrings.
69 // On match failure, none of the pointees will have been modified.
70 // On match success, the substrings will be converted (as necessary) and
71 // their values will be assigned to their pointees until all conversions
72 // have succeeded or one conversion has failed.
73 // On conversion failure, the pointees will be in an indeterminate state
74 // because the caller has no way of knowing which conversion failed.
75 // However, conversion cannot fail for types like string and StringPiece
76 // that do not inspect the substring contents. Hence, in the common case
77 // where all of the pointees are of such types, failure is always due to
78 // match failure and thus none of the pointees will have been modified.
79 //
80 // Example: extracts "ruby" into "s" and 1234 into "i"
81 // int i;
82 // std::string s;
83 // CHECK(RE2::FullMatch("ruby:1234", "(\\w+):(\\d+)", &s, &i));
84 //
85 // Example: fails because string cannot be stored in integer
86 // CHECK(!RE2::FullMatch("ruby", "(.*)", &i));
87 //
88 // Example: fails because there aren't enough sub-patterns
89 // CHECK(!RE2::FullMatch("ruby:1234", "\\w+:\\d+", &s));
90 //
91 // Example: does not try to extract any extra sub-patterns
92 // CHECK(RE2::FullMatch("ruby:1234", "(\\w+):(\\d+)", &s));
93 //
94 // Example: does not try to extract into NULL
95 // CHECK(RE2::FullMatch("ruby:1234", "(\\w+):(\\d+)", NULL, &i));
96 //
97 // Example: integer overflow causes failure
98 // CHECK(!RE2::FullMatch("ruby:1234567891234", "\\w+:(\\d+)", &i));
99 //
100 // NOTE(rsc): Asking for substrings slows successful matches quite a bit.
101 // This may get a little faster in the future, but right now is slower
102 // than PCRE. On the other hand, failed matches run *very* fast (faster
103 // than PCRE), as do matches without substring extraction.
104 //
105 // -----------------------------------------------------------------------
106 // PARTIAL MATCHES
107 //
108 // You can use the "PartialMatch" operation when you want the pattern
109 // to match any substring of the text.
110 //
111 // Example: simple search for a string:
112 // CHECK(RE2::PartialMatch("hello", "ell"));
113 //
114 // Example: find first number in a string
115 // int number;
116 // CHECK(RE2::PartialMatch("x*100 + 20", "(\\d+)", &number));
117 // CHECK_EQ(number, 100);
118 //
119 // -----------------------------------------------------------------------
120 // PRE-COMPILED REGULAR EXPRESSIONS
121 //
122 // RE2 makes it easy to use any string as a regular expression, without
123 // requiring a separate compilation step.
124 //
125 // If speed is of the essence, you can create a pre-compiled "RE2"
126 // object from the pattern and use it multiple times. If you do so,
127 // you can typically parse text faster than with sscanf.
128 //
129 // Example: precompile pattern for faster matching:
130 // RE2 pattern("h.*o");
131 // while (ReadLine(&str)) {
132 // if (RE2::FullMatch(str, pattern)) ...;
133 // }
134 //
135 // -----------------------------------------------------------------------
136 // SCANNING TEXT INCREMENTALLY
137 //
138 // The "Consume" operation may be useful if you want to repeatedly
139 // match regular expressions at the front of a string and skip over
140 // them as they match. This requires use of the "StringPiece" type,
141 // which represents a sub-range of a real string.
142 //
143 // Example: read lines of the form "var = value" from a string.
144 // std::string contents = ...; // Fill string somehow
145 // StringPiece input(contents); // Wrap a StringPiece around it
146 //
147 // std::string var;
148 // int value;
149 // while (RE2::Consume(&input, "(\\w+) = (\\d+)\n", &var, &value)) {
150 // ...;
151 // }
152 //
153 // Each successful call to "Consume" will set "var/value", and also
154 // advance "input" so it points past the matched text. Note that if the
155 // regular expression matches an empty string, input will advance
156 // by 0 bytes. If the regular expression being used might match
157 // an empty string, the loop body must check for this case and either
158 // advance the string or break out of the loop.
159 //
160 // The "FindAndConsume" operation is similar to "Consume" but does not
161 // anchor your match at the beginning of the string. For example, you
162 // could extract all words from a string by repeatedly calling
163 // RE2::FindAndConsume(&input, "(\\w+)", &word)
164 //
165 // -----------------------------------------------------------------------
166 // USING VARIABLE NUMBER OF ARGUMENTS
167 //
168 // The above operations require you to know the number of arguments
169 // when you write the code. This is not always possible or easy (for
170 // example, the regular expression may be calculated at run time).
171 // You can use the "N" version of the operations when the number of
172 // match arguments are determined at run time.
173 //
174 // Example:
175 // const RE2::Arg* args[10];
176 // int n;
177 // // ... populate args with pointers to RE2::Arg values ...
178 // // ... set n to the number of RE2::Arg objects ...
179 // bool match = RE2::FullMatchN(input, pattern, args, n);
180 //
181 // The last statement is equivalent to
182 //
183 // bool match = RE2::FullMatch(input, pattern,
184 // *args[0], *args[1], ..., *args[n - 1]);
185 //
186 // -----------------------------------------------------------------------
187 // PARSING HEX/OCTAL/C-RADIX NUMBERS
188 //
189 // By default, if you pass a pointer to a numeric value, the
190 // corresponding text is interpreted as a base-10 number. You can
191 // instead wrap the pointer with a call to one of the operators Hex(),
192 // Octal(), or CRadix() to interpret the text in another base. The
193 // CRadix operator interprets C-style "0" (base-8) and "0x" (base-16)
194 // prefixes, but defaults to base-10.
195 //
196 // Example:
197 // int a, b, c, d;
198 // CHECK(RE2::FullMatch("100 40 0100 0x40", "(.*) (.*) (.*) (.*)",
199 // RE2::Octal(&a), RE2::Hex(&b), RE2::CRadix(&c), RE2::CRadix(&d));
200 // will leave 64 in a, b, c, and d.
201
202 #include <stddef.h>
203 #include <stdint.h>
204 #include <algorithm>
205 #include <map>
206 #include <mutex>
207 #include <string>
208 #include <vector>
209
210 #if defined(__APPLE__)
211 #include <TargetConditionals.h>
212 #endif
213
214 #include "re2/stringpiece.h"
215
216 namespace re2 {
217 class Prog;
218 class Regexp;
219 } // namespace re2
220
221 namespace re2 {
222
223 // Interface for regular expression matching. Also corresponds to a
224 // pre-compiled regular expression. An "RE2" object is safe for
225 // concurrent use by multiple threads.
226 class RE2 {
227 public:
228 // We convert user-passed pointers into special Arg objects
229 class Arg;
230 class Options;
231
232 // Defined in set.h.
233 class Set;
234
235 enum ErrorCode {
236 NoError = 0,
237
238 // Unexpected error
239 ErrorInternal,
240
241 // Parse errors
242 ErrorBadEscape, // bad escape sequence
243 ErrorBadCharClass, // bad character class
244 ErrorBadCharRange, // bad character class range
245 ErrorMissingBracket, // missing closing ]
246 ErrorMissingParen, // missing closing )
247 ErrorTrailingBackslash, // trailing \ at end of regexp
248 ErrorRepeatArgument, // repeat argument missing, e.g. "*"
249 ErrorRepeatSize, // bad repetition argument
250 ErrorRepeatOp, // bad repetition operator
251 ErrorBadPerlOp, // bad perl operator
252 ErrorBadUTF8, // invalid UTF-8 in regexp
253 ErrorBadNamedCapture, // bad named capture group
254 ErrorPatternTooLarge // pattern too large (compile failed)
255 };
256
257 // Predefined common options.
258 // If you need more complicated things, instantiate
259 // an Option class, possibly passing one of these to
260 // the Option constructor, change the settings, and pass that
261 // Option class to the RE2 constructor.
262 enum CannedOptions {
263 DefaultOptions = 0,
264 Latin1, // treat input as Latin-1 (default UTF-8)
265 POSIX, // POSIX syntax, leftmost-longest match
266 Quiet // do not log about regexp parse errors
267 };
268
269 // Need to have the const char* and const std::string& forms for implicit
270 // conversions when passing string literals to FullMatch and PartialMatch.
271 // Otherwise the StringPiece form would be sufficient.
272 #ifndef SWIG
273 RE2(const char* pattern);
274 RE2(const std::string& pattern);
275 #endif
276 RE2(const StringPiece& pattern);
277 RE2(const StringPiece& pattern, const Options& options);
278 ~RE2();
279
280 // Returns whether RE2 was created properly.
ok()281 bool ok() const { return error_code() == NoError; }
282
283 // The string specification for this RE2. E.g.
284 // RE2 re("ab*c?d+");
285 // re.pattern(); // "ab*c?d+"
pattern()286 const std::string& pattern() const { return pattern_; }
287
288 // If RE2 could not be created properly, returns an error string.
289 // Else returns the empty string.
error()290 const std::string& error() const { return *error_; }
291
292 // If RE2 could not be created properly, returns an error code.
293 // Else returns RE2::NoError (== 0).
error_code()294 ErrorCode error_code() const { return error_code_; }
295
296 // If RE2 could not be created properly, returns the offending
297 // portion of the regexp.
error_arg()298 const std::string& error_arg() const { return error_arg_; }
299
300 // Returns the program size, a very approximate measure of a regexp's "cost".
301 // Larger numbers are more expensive than smaller numbers.
302 int ProgramSize() const;
303 int ReverseProgramSize() const;
304
305 // If histogram is not null, outputs the program fanout
306 // as a histogram bucketed by powers of 2.
307 // Returns the number of the largest non-empty bucket.
308 int ProgramFanout(std::vector<int>* histogram) const;
309 int ReverseProgramFanout(std::vector<int>* histogram) const;
310
311 // Returns the underlying Regexp; not for general use.
312 // Returns entire_regexp_ so that callers don't need
313 // to know about prefix_ and prefix_foldcase_.
Regexp()314 re2::Regexp* Regexp() const { return entire_regexp_; }
315
316 /***** The array-based matching interface ******/
317
318 // The functions here have names ending in 'N' and are used to implement
319 // the functions whose names are the prefix before the 'N'. It is sometimes
320 // useful to invoke them directly, but the syntax is awkward, so the 'N'-less
321 // versions should be preferred.
322 static bool FullMatchN(const StringPiece& text, const RE2& re,
323 const Arg* const args[], int n);
324 static bool PartialMatchN(const StringPiece& text, const RE2& re,
325 const Arg* const args[], int n);
326 static bool ConsumeN(StringPiece* input, const RE2& re,
327 const Arg* const args[], int n);
328 static bool FindAndConsumeN(StringPiece* input, const RE2& re,
329 const Arg* const args[], int n);
330
331 #ifndef SWIG
332 private:
333 template <typename F, typename SP>
Apply(F f,SP sp,const RE2 & re)334 static inline bool Apply(F f, SP sp, const RE2& re) {
335 return f(sp, re, NULL, 0);
336 }
337
338 template <typename F, typename SP, typename... A>
Apply(F f,SP sp,const RE2 & re,const A &...a)339 static inline bool Apply(F f, SP sp, const RE2& re, const A&... a) {
340 const Arg* const args[] = {&a...};
341 const int n = sizeof...(a);
342 return f(sp, re, args, n);
343 }
344
345 public:
346 // In order to allow FullMatch() et al. to be called with a varying number
347 // of arguments of varying types, we use two layers of variadic templates.
348 // The first layer constructs the temporary Arg objects. The second layer
349 // (above) constructs the array of pointers to the temporary Arg objects.
350
351 /***** The useful part: the matching interface *****/
352
353 // Matches "text" against "re". If pointer arguments are
354 // supplied, copies matched sub-patterns into them.
355 //
356 // You can pass in a "const char*" or a "std::string" for "text".
357 // You can pass in a "const char*" or a "std::string" or a "RE2" for "re".
358 //
359 // The provided pointer arguments can be pointers to any scalar numeric
360 // type, or one of:
361 // std::string (matched piece is copied to string)
362 // StringPiece (StringPiece is mutated to point to matched piece)
363 // T (where "bool T::ParseFrom(const char*, size_t)" exists)
364 // (void*)NULL (the corresponding matched sub-pattern is not copied)
365 //
366 // Returns true iff all of the following conditions are satisfied:
367 // a. "text" matches "re" exactly
368 // b. The number of matched sub-patterns is >= number of supplied pointers
369 // c. The "i"th argument has a suitable type for holding the
370 // string captured as the "i"th sub-pattern. If you pass in
371 // NULL for the "i"th argument, or pass fewer arguments than
372 // number of sub-patterns, "i"th captured sub-pattern is
373 // ignored.
374 //
375 // CAVEAT: An optional sub-pattern that does not exist in the
376 // matched string is assigned the empty string. Therefore, the
377 // following will return false (because the empty string is not a
378 // valid number):
379 // int number;
380 // RE2::FullMatch("abc", "[a-z]+(\\d+)?", &number);
381 template <typename... A>
FullMatch(const StringPiece & text,const RE2 & re,A &&...a)382 static bool FullMatch(const StringPiece& text, const RE2& re, A&&... a) {
383 return Apply(FullMatchN, text, re, Arg(std::forward<A>(a))...);
384 }
385
386 // Exactly like FullMatch(), except that "re" is allowed to match
387 // a substring of "text".
388 template <typename... A>
PartialMatch(const StringPiece & text,const RE2 & re,A &&...a)389 static bool PartialMatch(const StringPiece& text, const RE2& re, A&&... a) {
390 return Apply(PartialMatchN, text, re, Arg(std::forward<A>(a))...);
391 }
392
393 // Like FullMatch() and PartialMatch(), except that "re" has to match
394 // a prefix of the text, and "input" is advanced past the matched
395 // text. Note: "input" is modified iff this routine returns true
396 // and "re" matched a non-empty substring of "text".
397 template <typename... A>
Consume(StringPiece * input,const RE2 & re,A &&...a)398 static bool Consume(StringPiece* input, const RE2& re, A&&... a) {
399 return Apply(ConsumeN, input, re, Arg(std::forward<A>(a))...);
400 }
401
402 // Like Consume(), but does not anchor the match at the beginning of
403 // the text. That is, "re" need not start its match at the beginning
404 // of "input". For example, "FindAndConsume(s, "(\\w+)", &word)" finds
405 // the next word in "s" and stores it in "word".
406 template <typename... A>
FindAndConsume(StringPiece * input,const RE2 & re,A &&...a)407 static bool FindAndConsume(StringPiece* input, const RE2& re, A&&... a) {
408 return Apply(FindAndConsumeN, input, re, Arg(std::forward<A>(a))...);
409 }
410 #endif
411
412 // Replace the first match of "re" in "str" with "rewrite".
413 // Within "rewrite", backslash-escaped digits (\1 to \9) can be
414 // used to insert text matching corresponding parenthesized group
415 // from the pattern. \0 in "rewrite" refers to the entire matching
416 // text. E.g.,
417 //
418 // std::string s = "yabba dabba doo";
419 // CHECK(RE2::Replace(&s, "b+", "d"));
420 //
421 // will leave "s" containing "yada dabba doo"
422 //
423 // Returns true if the pattern matches and a replacement occurs,
424 // false otherwise.
425 static bool Replace(std::string* str,
426 const RE2& re,
427 const StringPiece& rewrite);
428
429 // Like Replace(), except replaces successive non-overlapping occurrences
430 // of the pattern in the string with the rewrite. E.g.
431 //
432 // std::string s = "yabba dabba doo";
433 // CHECK(RE2::GlobalReplace(&s, "b+", "d"));
434 //
435 // will leave "s" containing "yada dada doo"
436 // Replacements are not subject to re-matching.
437 //
438 // Because GlobalReplace only replaces non-overlapping matches,
439 // replacing "ana" within "banana" makes only one replacement, not two.
440 //
441 // Returns the number of replacements made.
442 static int GlobalReplace(std::string* str,
443 const RE2& re,
444 const StringPiece& rewrite);
445
446 // Like Replace, except that if the pattern matches, "rewrite"
447 // is copied into "out" with substitutions. The non-matching
448 // portions of "text" are ignored.
449 //
450 // Returns true iff a match occurred and the extraction happened
451 // successfully; if no match occurs, the string is left unaffected.
452 //
453 // REQUIRES: "text" must not alias any part of "*out".
454 static bool Extract(const StringPiece& text,
455 const RE2& re,
456 const StringPiece& rewrite,
457 std::string* out);
458
459 // Escapes all potentially meaningful regexp characters in
460 // 'unquoted'. The returned string, used as a regular expression,
461 // will match exactly the original string. For example,
462 // 1.5-2.0?
463 // may become:
464 // 1\.5\-2\.0\?
465 static std::string QuoteMeta(const StringPiece& unquoted);
466
467 // Computes range for any strings matching regexp. The min and max can in
468 // some cases be arbitrarily precise, so the caller gets to specify the
469 // maximum desired length of string returned.
470 //
471 // Assuming PossibleMatchRange(&min, &max, N) returns successfully, any
472 // string s that is an anchored match for this regexp satisfies
473 // min <= s && s <= max.
474 //
475 // Note that PossibleMatchRange() will only consider the first copy of an
476 // infinitely repeated element (i.e., any regexp element followed by a '*' or
477 // '+' operator). Regexps with "{N}" constructions are not affected, as those
478 // do not compile down to infinite repetitions.
479 //
480 // Returns true on success, false on error.
481 bool PossibleMatchRange(std::string* min, std::string* max,
482 int maxlen) const;
483
484 // Generic matching interface
485
486 // Type of match.
487 enum Anchor {
488 UNANCHORED, // No anchoring
489 ANCHOR_START, // Anchor at start only
490 ANCHOR_BOTH // Anchor at start and end
491 };
492
493 // Return the number of capturing subpatterns, or -1 if the
494 // regexp wasn't valid on construction. The overall match ($0)
495 // does not count: if the regexp is "(a)(b)", returns 2.
NumberOfCapturingGroups()496 int NumberOfCapturingGroups() const { return num_captures_; }
497
498 // Return a map from names to capturing indices.
499 // The map records the index of the leftmost group
500 // with the given name.
501 // Only valid until the re is deleted.
502 const std::map<std::string, int>& NamedCapturingGroups() const;
503
504 // Return a map from capturing indices to names.
505 // The map has no entries for unnamed groups.
506 // Only valid until the re is deleted.
507 const std::map<int, std::string>& CapturingGroupNames() const;
508
509 // General matching routine.
510 // Match against text starting at offset startpos
511 // and stopping the search at offset endpos.
512 // Returns true if match found, false if not.
513 // On a successful match, fills in submatch[] (up to nsubmatch entries)
514 // with information about submatches.
515 // I.e. matching RE2("(foo)|(bar)baz") on "barbazbla" will return true, with
516 // submatch[0] = "barbaz", submatch[1].data() = NULL, submatch[2] = "bar",
517 // submatch[3].data() = NULL, ..., up to submatch[nsubmatch-1].data() = NULL.
518 // Caveat: submatch[] may be clobbered even on match failure.
519 //
520 // Don't ask for more match information than you will use:
521 // runs much faster with nsubmatch == 1 than nsubmatch > 1, and
522 // runs even faster if nsubmatch == 0.
523 // Doesn't make sense to use nsubmatch > 1 + NumberOfCapturingGroups(),
524 // but will be handled correctly.
525 //
526 // Passing text == StringPiece(NULL, 0) will be handled like any other
527 // empty string, but note that on return, it will not be possible to tell
528 // whether submatch i matched the empty string or did not match:
529 // either way, submatch[i].data() == NULL.
530 bool Match(const StringPiece& text,
531 size_t startpos,
532 size_t endpos,
533 Anchor re_anchor,
534 StringPiece* submatch,
535 int nsubmatch) const;
536
537 // Check that the given rewrite string is suitable for use with this
538 // regular expression. It checks that:
539 // * The regular expression has enough parenthesized subexpressions
540 // to satisfy all of the \N tokens in rewrite
541 // * The rewrite string doesn't have any syntax errors. E.g.,
542 // '\' followed by anything other than a digit or '\'.
543 // A true return value guarantees that Replace() and Extract() won't
544 // fail because of a bad rewrite string.
545 bool CheckRewriteString(const StringPiece& rewrite,
546 std::string* error) const;
547
548 // Returns the maximum submatch needed for the rewrite to be done by
549 // Replace(). E.g. if rewrite == "foo \\2,\\1", returns 2.
550 static int MaxSubmatch(const StringPiece& rewrite);
551
552 // Append the "rewrite" string, with backslash subsitutions from "vec",
553 // to string "out".
554 // Returns true on success. This method can fail because of a malformed
555 // rewrite string. CheckRewriteString guarantees that the rewrite will
556 // be sucessful.
557 bool Rewrite(std::string* out,
558 const StringPiece& rewrite,
559 const StringPiece* vec,
560 int veclen) const;
561
562 // Constructor options
563 class Options {
564 public:
565 // The options are (defaults in parentheses):
566 //
567 // utf8 (true) text and pattern are UTF-8; otherwise Latin-1
568 // posix_syntax (false) restrict regexps to POSIX egrep syntax
569 // longest_match (false) search for longest match, not first match
570 // log_errors (true) log syntax and execution errors to ERROR
571 // max_mem (see below) approx. max memory footprint of RE2
572 // literal (false) interpret string as literal, not regexp
573 // never_nl (false) never match \n, even if it is in regexp
574 // dot_nl (false) dot matches everything including new line
575 // never_capture (false) parse all parens as non-capturing
576 // case_sensitive (true) match is case-sensitive (regexp can override
577 // with (?i) unless in posix_syntax mode)
578 //
579 // The following options are only consulted when posix_syntax == true.
580 // When posix_syntax == false, these features are always enabled and
581 // cannot be turned off; to perform multi-line matching in that case,
582 // begin the regexp with (?m).
583 // perl_classes (false) allow Perl's \d \s \w \D \S \W
584 // word_boundary (false) allow Perl's \b \B (word boundary and not)
585 // one_line (false) ^ and $ only match beginning and end of text
586 //
587 // The max_mem option controls how much memory can be used
588 // to hold the compiled form of the regexp (the Prog) and
589 // its cached DFA graphs. Code Search placed limits on the number
590 // of Prog instructions and DFA states: 10,000 for both.
591 // In RE2, those limits would translate to about 240 KB per Prog
592 // and perhaps 2.5 MB per DFA (DFA state sizes vary by regexp; RE2 does a
593 // better job of keeping them small than Code Search did).
594 // Each RE2 has two Progs (one forward, one reverse), and each Prog
595 // can have two DFAs (one first match, one longest match).
596 // That makes 4 DFAs:
597 //
598 // forward, first-match - used for UNANCHORED or ANCHOR_START searches
599 // if opt.longest_match() == false
600 // forward, longest-match - used for all ANCHOR_BOTH searches,
601 // and the other two kinds if
602 // opt.longest_match() == true
603 // reverse, first-match - never used
604 // reverse, longest-match - used as second phase for unanchored searches
605 //
606 // The RE2 memory budget is statically divided between the two
607 // Progs and then the DFAs: two thirds to the forward Prog
608 // and one third to the reverse Prog. The forward Prog gives half
609 // of what it has left over to each of its DFAs. The reverse Prog
610 // gives it all to its longest-match DFA.
611 //
612 // Once a DFA fills its budget, it flushes its cache and starts over.
613 // If this happens too often, RE2 falls back on the NFA implementation.
614
615 // For now, make the default budget something close to Code Search.
616 static const int kDefaultMaxMem = 8<<20;
617
618 enum Encoding {
619 EncodingUTF8 = 1,
620 EncodingLatin1
621 };
622
Options()623 Options() :
624 encoding_(EncodingUTF8),
625 posix_syntax_(false),
626 longest_match_(false),
627 log_errors_(true),
628 max_mem_(kDefaultMaxMem),
629 literal_(false),
630 never_nl_(false),
631 dot_nl_(false),
632 never_capture_(false),
633 case_sensitive_(true),
634 perl_classes_(false),
635 word_boundary_(false),
636 one_line_(false) {
637 }
638
639 /*implicit*/ Options(CannedOptions);
640
encoding()641 Encoding encoding() const { return encoding_; }
set_encoding(Encoding encoding)642 void set_encoding(Encoding encoding) { encoding_ = encoding; }
643
posix_syntax()644 bool posix_syntax() const { return posix_syntax_; }
set_posix_syntax(bool b)645 void set_posix_syntax(bool b) { posix_syntax_ = b; }
646
longest_match()647 bool longest_match() const { return longest_match_; }
set_longest_match(bool b)648 void set_longest_match(bool b) { longest_match_ = b; }
649
log_errors()650 bool log_errors() const { return log_errors_; }
set_log_errors(bool b)651 void set_log_errors(bool b) { log_errors_ = b; }
652
max_mem()653 int64_t max_mem() const { return max_mem_; }
set_max_mem(int64_t m)654 void set_max_mem(int64_t m) { max_mem_ = m; }
655
literal()656 bool literal() const { return literal_; }
set_literal(bool b)657 void set_literal(bool b) { literal_ = b; }
658
never_nl()659 bool never_nl() const { return never_nl_; }
set_never_nl(bool b)660 void set_never_nl(bool b) { never_nl_ = b; }
661
dot_nl()662 bool dot_nl() const { return dot_nl_; }
set_dot_nl(bool b)663 void set_dot_nl(bool b) { dot_nl_ = b; }
664
never_capture()665 bool never_capture() const { return never_capture_; }
set_never_capture(bool b)666 void set_never_capture(bool b) { never_capture_ = b; }
667
case_sensitive()668 bool case_sensitive() const { return case_sensitive_; }
set_case_sensitive(bool b)669 void set_case_sensitive(bool b) { case_sensitive_ = b; }
670
perl_classes()671 bool perl_classes() const { return perl_classes_; }
set_perl_classes(bool b)672 void set_perl_classes(bool b) { perl_classes_ = b; }
673
word_boundary()674 bool word_boundary() const { return word_boundary_; }
set_word_boundary(bool b)675 void set_word_boundary(bool b) { word_boundary_ = b; }
676
one_line()677 bool one_line() const { return one_line_; }
set_one_line(bool b)678 void set_one_line(bool b) { one_line_ = b; }
679
Copy(const Options & src)680 void Copy(const Options& src) {
681 *this = src;
682 }
683
684 int ParseFlags() const;
685
686 private:
687 Encoding encoding_;
688 bool posix_syntax_;
689 bool longest_match_;
690 bool log_errors_;
691 int64_t max_mem_;
692 bool literal_;
693 bool never_nl_;
694 bool dot_nl_;
695 bool never_capture_;
696 bool case_sensitive_;
697 bool perl_classes_;
698 bool word_boundary_;
699 bool one_line_;
700 };
701
702 // Returns the options set in the constructor.
options()703 const Options& options() const { return options_; }
704
705 // Argument converters; see below.
706 static inline Arg CRadix(short* x);
707 static inline Arg CRadix(unsigned short* x);
708 static inline Arg CRadix(int* x);
709 static inline Arg CRadix(unsigned int* x);
710 static inline Arg CRadix(long* x);
711 static inline Arg CRadix(unsigned long* x);
712 static inline Arg CRadix(long long* x);
713 static inline Arg CRadix(unsigned long long* x);
714
715 static inline Arg Hex(short* x);
716 static inline Arg Hex(unsigned short* x);
717 static inline Arg Hex(int* x);
718 static inline Arg Hex(unsigned int* x);
719 static inline Arg Hex(long* x);
720 static inline Arg Hex(unsigned long* x);
721 static inline Arg Hex(long long* x);
722 static inline Arg Hex(unsigned long long* x);
723
724 static inline Arg Octal(short* x);
725 static inline Arg Octal(unsigned short* x);
726 static inline Arg Octal(int* x);
727 static inline Arg Octal(unsigned int* x);
728 static inline Arg Octal(long* x);
729 static inline Arg Octal(unsigned long* x);
730 static inline Arg Octal(long long* x);
731 static inline Arg Octal(unsigned long long* x);
732
733 private:
734 void Init(const StringPiece& pattern, const Options& options);
735
736 bool DoMatch(const StringPiece& text,
737 Anchor re_anchor,
738 size_t* consumed,
739 const Arg* const args[],
740 int n) const;
741
742 re2::Prog* ReverseProg() const;
743
744 std::string pattern_; // string regular expression
745 Options options_; // option flags
746 re2::Regexp* entire_regexp_; // parsed regular expression
747 const std::string* error_; // error indicator (or points to empty string)
748 ErrorCode error_code_; // error code
749 std::string error_arg_; // fragment of regexp showing error
750 std::string prefix_; // required prefix (before suffix_regexp_)
751 bool prefix_foldcase_; // prefix_ is ASCII case-insensitive
752 re2::Regexp* suffix_regexp_; // parsed regular expression, prefix_ removed
753 re2::Prog* prog_; // compiled program for regexp
754 int num_captures_; // number of capturing groups
755 bool is_one_pass_; // can use prog_->SearchOnePass?
756
757 // Reverse Prog for DFA execution only
758 mutable re2::Prog* rprog_;
759 // Map from capture names to indices
760 mutable const std::map<std::string, int>* named_groups_;
761 // Map from capture indices to names
762 mutable const std::map<int, std::string>* group_names_;
763
764 mutable std::once_flag rprog_once_;
765 mutable std::once_flag named_groups_once_;
766 mutable std::once_flag group_names_once_;
767
768 RE2(const RE2&) = delete;
769 RE2& operator=(const RE2&) = delete;
770 };
771
772 /***** Implementation details *****/
773
774 // Hex/Octal/Binary?
775
776 // Special class for parsing into objects that define a ParseFrom() method
777 template <class T>
778 class _RE2_MatchObject {
779 public:
Parse(const char * str,size_t n,void * dest)780 static inline bool Parse(const char* str, size_t n, void* dest) {
781 if (dest == NULL) return true;
782 T* object = reinterpret_cast<T*>(dest);
783 return object->ParseFrom(str, n);
784 }
785 };
786
787 class RE2::Arg {
788 public:
789 // Empty constructor so we can declare arrays of RE2::Arg
790 Arg();
791
792 // Constructor specially designed for NULL arguments
793 Arg(void*);
794 Arg(std::nullptr_t);
795
796 typedef bool (*Parser)(const char* str, size_t n, void* dest);
797
798 // Type-specific parsers
799 #define MAKE_PARSER(type, name) \
800 Arg(type* p) : arg_(p), parser_(name) {} \
801 Arg(type* p, Parser parser) : arg_(p), parser_(parser) {}
802
MAKE_PARSER(char,parse_char)803 MAKE_PARSER(char, parse_char)
804 MAKE_PARSER(signed char, parse_schar)
805 MAKE_PARSER(unsigned char, parse_uchar)
806 MAKE_PARSER(float, parse_float)
807 MAKE_PARSER(double, parse_double)
808 MAKE_PARSER(std::string, parse_string)
809 MAKE_PARSER(StringPiece, parse_stringpiece)
810
811 MAKE_PARSER(short, parse_short)
812 MAKE_PARSER(unsigned short, parse_ushort)
813 MAKE_PARSER(int, parse_int)
814 MAKE_PARSER(unsigned int, parse_uint)
815 MAKE_PARSER(long, parse_long)
816 MAKE_PARSER(unsigned long, parse_ulong)
817 MAKE_PARSER(long long, parse_longlong)
818 MAKE_PARSER(unsigned long long, parse_ulonglong)
819
820 #undef MAKE_PARSER
821
822 // Generic constructor templates
823 template <class T> Arg(T* p)
824 : arg_(p), parser_(_RE2_MatchObject<T>::Parse) { }
Arg(T * p,Parser parser)825 template <class T> Arg(T* p, Parser parser)
826 : arg_(p), parser_(parser) { }
827
828 // Parse the data
829 bool Parse(const char* str, size_t n) const;
830
831 private:
832 void* arg_;
833 Parser parser_;
834
835 static bool parse_null (const char* str, size_t n, void* dest);
836 static bool parse_char (const char* str, size_t n, void* dest);
837 static bool parse_schar (const char* str, size_t n, void* dest);
838 static bool parse_uchar (const char* str, size_t n, void* dest);
839 static bool parse_float (const char* str, size_t n, void* dest);
840 static bool parse_double (const char* str, size_t n, void* dest);
841 static bool parse_string (const char* str, size_t n, void* dest);
842 static bool parse_stringpiece (const char* str, size_t n, void* dest);
843
844 #define DECLARE_INTEGER_PARSER(name) \
845 private: \
846 static bool parse_##name(const char* str, size_t n, void* dest); \
847 static bool parse_##name##_radix(const char* str, size_t n, void* dest, \
848 int radix); \
849 \
850 public: \
851 static bool parse_##name##_hex(const char* str, size_t n, void* dest); \
852 static bool parse_##name##_octal(const char* str, size_t n, void* dest); \
853 static bool parse_##name##_cradix(const char* str, size_t n, void* dest);
854
855 DECLARE_INTEGER_PARSER(short)
856 DECLARE_INTEGER_PARSER(ushort)
857 DECLARE_INTEGER_PARSER(int)
858 DECLARE_INTEGER_PARSER(uint)
859 DECLARE_INTEGER_PARSER(long)
860 DECLARE_INTEGER_PARSER(ulong)
861 DECLARE_INTEGER_PARSER(longlong)
862 DECLARE_INTEGER_PARSER(ulonglong)
863
864 #undef DECLARE_INTEGER_PARSER
865
866 };
867
Arg()868 inline RE2::Arg::Arg() : arg_(NULL), parser_(parse_null) { }
Arg(void * p)869 inline RE2::Arg::Arg(void* p) : arg_(p), parser_(parse_null) { }
Arg(std::nullptr_t p)870 inline RE2::Arg::Arg(std::nullptr_t p) : arg_(p), parser_(parse_null) { }
871
Parse(const char * str,size_t n)872 inline bool RE2::Arg::Parse(const char* str, size_t n) const {
873 return (*parser_)(str, n, arg_);
874 }
875
876 // This part of the parser, appropriate only for ints, deals with bases
877 #define MAKE_INTEGER_PARSER(type, name) \
878 inline RE2::Arg RE2::Hex(type* ptr) { \
879 return RE2::Arg(ptr, RE2::Arg::parse_##name##_hex); \
880 } \
881 inline RE2::Arg RE2::Octal(type* ptr) { \
882 return RE2::Arg(ptr, RE2::Arg::parse_##name##_octal); \
883 } \
884 inline RE2::Arg RE2::CRadix(type* ptr) { \
885 return RE2::Arg(ptr, RE2::Arg::parse_##name##_cradix); \
886 }
887
MAKE_INTEGER_PARSER(short,short)888 MAKE_INTEGER_PARSER(short, short)
889 MAKE_INTEGER_PARSER(unsigned short, ushort)
890 MAKE_INTEGER_PARSER(int, int)
891 MAKE_INTEGER_PARSER(unsigned int, uint)
892 MAKE_INTEGER_PARSER(long, long)
893 MAKE_INTEGER_PARSER(unsigned long, ulong)
894 MAKE_INTEGER_PARSER(long long, longlong)
895 MAKE_INTEGER_PARSER(unsigned long long, ulonglong)
896
897 #undef MAKE_INTEGER_PARSER
898
899 #ifndef SWIG
900 // Silence warnings about missing initializers for members of LazyRE2.
901 #if !defined(__clang__) && defined(__GNUC__) && __GNUC__ >= 6
902 #pragma GCC diagnostic ignored "-Wmissing-field-initializers"
903 #endif
904
905 // Helper for writing global or static RE2s safely.
906 // Write
907 // static LazyRE2 re = {".*"};
908 // and then use *re instead of writing
909 // static RE2 re(".*");
910 // The former is more careful about multithreaded
911 // situations than the latter.
912 //
913 // N.B. This class never deletes the RE2 object that
914 // it constructs: that's a feature, so that it can be used
915 // for global and function static variables.
916 class LazyRE2 {
917 private:
918 struct NoArg {};
919
920 public:
921 typedef RE2 element_type; // support std::pointer_traits
922
923 // Constructor omitted to preserve braced initialization in C++98.
924
925 // Pretend to be a pointer to Type (never NULL due to on-demand creation):
926 RE2& operator*() const { return *get(); }
927 RE2* operator->() const { return get(); }
928
929 // Named accessor/initializer:
930 RE2* get() const {
931 std::call_once(once_, &LazyRE2::Init, this);
932 return ptr_;
933 }
934
935 // All data fields must be public to support {"foo"} initialization.
936 const char* pattern_;
937 RE2::CannedOptions options_;
938 NoArg barrier_against_excess_initializers_;
939
940 mutable RE2* ptr_;
941 mutable std::once_flag once_;
942
943 private:
944 static void Init(const LazyRE2* lazy_re2) {
945 lazy_re2->ptr_ = new RE2(lazy_re2->pattern_, lazy_re2->options_);
946 }
947
948 void operator=(const LazyRE2&); // disallowed
949 };
950 #endif
951
952 namespace hooks {
953
954 // Most platforms support thread_local. Older versions of iOS don't support
955 // thread_local, but for the sake of brevity, we lump together all versions
956 // of Apple platforms that aren't macOS. If an iOS application really needs
957 // the context pointee someday, we can get more specific then...
958 #define RE2_HAVE_THREAD_LOCAL
959 #if defined(__APPLE__) && !TARGET_OS_OSX
960 #undef RE2_HAVE_THREAD_LOCAL
961 #endif
962
963 // A hook must not make any assumptions regarding the lifetime of the context
964 // pointee beyond the current invocation of the hook. Pointers and references
965 // obtained via the context pointee should be considered invalidated when the
966 // hook returns. Hence, any data about the context pointee (e.g. its pattern)
967 // would have to be copied in order for it to be kept for an indefinite time.
968 //
969 // A hook must not use RE2 for matching. Control flow reentering RE2::Match()
970 // could result in infinite mutual recursion. To discourage that possibility,
971 // RE2 will not maintain the context pointer correctly when used in that way.
972 #ifdef RE2_HAVE_THREAD_LOCAL
973 extern thread_local const RE2* context;
974 #endif
975
976 struct DFAStateCacheReset {
977 int64_t state_budget;
978 size_t state_cache_size;
979 };
980
981 struct DFASearchFailure {
982 // Nothing yet...
983 };
984
985 #define DECLARE_HOOK(type) \
986 using type##Callback = void(const type&); \
987 void Set##type##Hook(type##Callback* cb); \
988 type##Callback* Get##type##Hook();
989
990 DECLARE_HOOK(DFAStateCacheReset)
991 DECLARE_HOOK(DFASearchFailure)
992
993 #undef DECLARE_HOOK
994
995 } // namespace hooks
996
997 } // namespace re2
998
999 using re2::RE2;
1000 using re2::LazyRE2;
1001
1002 #endif // RE2_RE2_H_
1003