OLD | NEW |
(Empty) | |
| 1 // Copyright (c) 2012 The Chromium Authors. All rights reserved. |
| 2 // Use of this source code is governed by a BSD-style license that can be |
| 3 // found in the LICENSE file. |
| 4 |
| 5 #ifndef CHROME_COMMON_EXTENSIONS_MATCHER_REGEX_SET_MATCHER_H_ |
| 6 #define CHROME_COMMON_EXTENSIONS_MATCHER_REGEX_SET_MATCHER_H_ |
| 7 |
| 8 #include <map> |
| 9 #include <set> |
| 10 #include <string> |
| 11 #include <vector> |
| 12 |
| 13 #include "base/memory/scoped_ptr.h" |
| 14 #include "chrome/common/extensions/matcher/string_pattern.h" |
| 15 #include "chrome/common/extensions/matcher/substring_set_matcher.h" |
| 16 |
| 17 namespace re2 { |
| 18 class FilteredRE2; |
| 19 } |
| 20 |
| 21 namespace extensions { |
| 22 |
| 23 // Efficiently matches URLs against a collection of regular expressions, |
| 24 // using FilteredRE2 to reduce the number of regexes that must be matched |
| 25 // by pre-filtering with substring matching. See: |
| 26 // http://swtch.com/~rsc/regexp/regexp3.html#analysis |
| 27 class RegexSetMatcher { |
| 28 public: |
| 29 RegexSetMatcher(); |
| 30 virtual ~RegexSetMatcher(); |
| 31 |
| 32 // Adds the regex patterns in |regex_list| to the matcher. Also rebuilds |
| 33 // the FilteredRE2 matcher; thus, for efficiency, prefer adding multiple |
| 34 // patterns at once. |
| 35 // Ownership of the patterns remains with the caller. |
| 36 void AddPatterns(const std::vector<const StringPattern*>& regex_list); |
| 37 |
| 38 // Removes all regex patterns. |
| 39 void ClearPatterns(); |
| 40 |
| 41 // Appends the IDs of regular expressions in our set that match the |text| |
| 42 // to |matches|. |
| 43 bool Match(const std::string& text, |
| 44 std::set<StringPattern::ID>* matches) const; |
| 45 |
| 46 private: |
| 47 typedef int RE2ID; |
| 48 typedef std::map<StringPattern::ID, const StringPattern*> RegexMap; |
| 49 typedef std::vector<StringPattern::ID> RE2IDMap; |
| 50 |
| 51 // Use Aho-Corasick SubstringSetMatcher to find which literal patterns |
| 52 // match the |text|. |
| 53 std::vector<RE2ID> FindSubstringMatches(const std::string& text) const; |
| 54 |
| 55 // Rebuild FilteredRE2 from scratch. Needs to be called whenever |
| 56 // our set of regexes changes. |
| 57 // TODO(yoz): investigate if it could be done incrementally; |
| 58 // apparently not supported by FilteredRE2. |
| 59 void RebuildMatcher(); |
| 60 |
| 61 // Clean up StringPatterns in |substring_patterns_|. |
| 62 void DeleteSubstringPatterns(); |
| 63 |
| 64 // Mapping of regex StringPattern::IDs to regexes. |
| 65 RegexMap regexes_; |
| 66 // Mapping of RE2IDs from FilteredRE2 (which are assigned in order) |
| 67 // to regex StringPattern::IDs. |
| 68 RE2IDMap re2_id_map_; |
| 69 |
| 70 scoped_ptr<re2::FilteredRE2> filtered_re2_; |
| 71 scoped_ptr<SubstringSetMatcher> substring_matcher_; |
| 72 |
| 73 // The substring patterns from FilteredRE2, which are used in |
| 74 // |substring_matcher_| but whose lifetime is managed here. |
| 75 std::vector<const StringPattern*> substring_patterns_; |
| 76 }; |
| 77 |
| 78 } // namespace extensions |
| 79 |
| 80 #endif // CHROME_COMMON_EXTENSIONS_MATCHER_REGEX_SET_MATCHER_H_ |
OLD | NEW |