Andrew Walbran | 3d2c197 | 2020-04-07 12:24:26 +0100 | [diff] [blame^] | 1 | //===-- RegularExpression.h -------------------------------------*- C++ -*-===// |
| 2 | // |
| 3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 | // See https://llvm.org/LICENSE.txt for license information. |
| 5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 | // |
| 7 | //===----------------------------------------------------------------------===// |
| 8 | |
| 9 | #ifndef liblldb_RegularExpression_h_ |
| 10 | #define liblldb_RegularExpression_h_ |
| 11 | |
| 12 | #ifdef _WIN32 |
| 13 | #include "../lib/Support/regex_impl.h" |
| 14 | |
| 15 | typedef llvm_regmatch_t regmatch_t; |
| 16 | typedef llvm_regex_t regex_t; |
| 17 | |
| 18 | inline int regcomp(llvm_regex_t *a, const char *b, int c) { |
| 19 | return llvm_regcomp(a, b, c); |
| 20 | } |
| 21 | |
| 22 | inline size_t regerror(int a, const llvm_regex_t *b, char *c, size_t d) { |
| 23 | return llvm_regerror(a, b, c, d); |
| 24 | } |
| 25 | |
| 26 | inline int regexec(const llvm_regex_t *a, const char *b, size_t c, |
| 27 | llvm_regmatch_t d[], int e) { |
| 28 | return llvm_regexec(a, b, c, d, e); |
| 29 | } |
| 30 | |
| 31 | inline void regfree(llvm_regex_t *a) { llvm_regfree(a); } |
| 32 | #else |
| 33 | #ifdef __ANDROID__ |
| 34 | #include <regex> |
| 35 | #endif |
| 36 | #include <regex.h> |
| 37 | #endif |
| 38 | |
| 39 | #include <string> |
| 40 | #include <vector> |
| 41 | |
| 42 | #include <stddef.h> |
| 43 | #include <stdint.h> |
| 44 | |
| 45 | namespace llvm { |
| 46 | class StringRef; |
| 47 | } // namespace llvm |
| 48 | |
| 49 | namespace lldb_private { |
| 50 | |
| 51 | /// \class RegularExpression RegularExpression.h |
| 52 | /// "lldb/Utility/RegularExpression.h" |
| 53 | /// A C++ wrapper class for regex. |
| 54 | /// |
| 55 | /// This regular expression class wraps the posix regex functions \c |
| 56 | /// regcomp(), \c regerror(), \c regexec(), and \c regfree() from the header |
| 57 | /// file in \c /usr/include/regex\.h. |
| 58 | class RegularExpression { |
| 59 | public: |
| 60 | class Match { |
| 61 | public: |
| 62 | Match(uint32_t max_matches) : m_matches() { |
| 63 | if (max_matches > 0) |
| 64 | m_matches.resize(max_matches + 1); |
| 65 | } |
| 66 | |
| 67 | void Clear() { |
| 68 | const size_t num_matches = m_matches.size(); |
| 69 | regmatch_t invalid_match = {-1, -1}; |
| 70 | for (size_t i = 0; i < num_matches; ++i) |
| 71 | m_matches[i] = invalid_match; |
| 72 | } |
| 73 | |
| 74 | size_t GetSize() const { return m_matches.size(); } |
| 75 | |
| 76 | regmatch_t *GetData() { |
| 77 | return (m_matches.empty() ? nullptr : m_matches.data()); |
| 78 | } |
| 79 | |
| 80 | bool GetMatchAtIndex(llvm::StringRef s, uint32_t idx, |
| 81 | std::string &match_str) const; |
| 82 | |
| 83 | bool GetMatchAtIndex(llvm::StringRef s, uint32_t idx, |
| 84 | llvm::StringRef &match_str) const; |
| 85 | |
| 86 | bool GetMatchSpanningIndices(llvm::StringRef s, uint32_t idx1, |
| 87 | uint32_t idx2, |
| 88 | llvm::StringRef &match_str) const; |
| 89 | |
| 90 | protected: |
| 91 | std::vector<regmatch_t> |
| 92 | m_matches; ///< Where parenthesized subexpressions results are stored |
| 93 | }; |
| 94 | |
| 95 | /// Default constructor. |
| 96 | /// |
| 97 | /// The default constructor that initializes the object state such that it |
| 98 | /// contains no compiled regular expression. |
| 99 | RegularExpression(); |
| 100 | |
| 101 | explicit RegularExpression(llvm::StringRef string); |
| 102 | |
| 103 | /// Destructor. |
| 104 | /// |
| 105 | /// Any previously compiled regular expression contained in this object will |
| 106 | /// be freed. |
| 107 | ~RegularExpression(); |
| 108 | |
| 109 | RegularExpression(const RegularExpression &rhs); |
| 110 | |
| 111 | const RegularExpression &operator=(const RegularExpression &rhs); |
| 112 | |
| 113 | /// Compile a regular expression. |
| 114 | /// |
| 115 | /// Compile a regular expression using the supplied regular expression text. |
| 116 | /// The compiled regular expression lives in this object so that it can be |
| 117 | /// readily used for regular expression matches. Execute() can be called |
| 118 | /// after the regular expression is compiled. Any previously compiled |
| 119 | /// regular expression contained in this object will be freed. |
| 120 | /// |
| 121 | /// \param[in] re |
| 122 | /// A NULL terminated C string that represents the regular |
| 123 | /// expression to compile. |
| 124 | /// |
| 125 | /// \return |
| 126 | /// \b true if the regular expression compiles successfully, |
| 127 | /// \b false otherwise. |
| 128 | bool Compile(llvm::StringRef string); |
| 129 | bool Compile(const char *) = delete; |
| 130 | |
| 131 | /// Executes a regular expression. |
| 132 | /// |
| 133 | /// Execute a regular expression match using the compiled regular expression |
| 134 | /// that is already in this object against the match string \a s. If any |
| 135 | /// parens are used for regular expression matches \a match_count should |
| 136 | /// indicate the number of regmatch_t values that are present in \a |
| 137 | /// match_ptr. |
| 138 | /// |
| 139 | /// \param[in] string |
| 140 | /// The string to match against the compile regular expression. |
| 141 | /// |
| 142 | /// \param[in] match |
| 143 | /// A pointer to a RegularExpression::Match structure that was |
| 144 | /// properly initialized with the desired number of maximum |
| 145 | /// matches, or nullptr if no parenthesized matching is needed. |
| 146 | /// |
| 147 | /// \return |
| 148 | /// \b true if \a string matches the compiled regular |
| 149 | /// expression, \b false otherwise. |
| 150 | bool Execute(llvm::StringRef string, Match *match = nullptr) const; |
| 151 | bool Execute(const char *, Match * = nullptr) = delete; |
| 152 | |
| 153 | size_t GetErrorAsCString(char *err_str, size_t err_str_max_len) const; |
| 154 | |
| 155 | /// Free the compiled regular expression. |
| 156 | /// |
| 157 | /// If this object contains a valid compiled regular expression, this |
| 158 | /// function will free any resources it was consuming. |
| 159 | void Free(); |
| 160 | |
| 161 | /// Access the regular expression text. |
| 162 | /// |
| 163 | /// Returns the text that was used to compile the current regular |
| 164 | /// expression. |
| 165 | /// |
| 166 | /// \return |
| 167 | /// The NULL terminated C string that was used to compile the |
| 168 | /// current regular expression |
| 169 | llvm::StringRef GetText() const; |
| 170 | |
| 171 | /// Test if valid. |
| 172 | /// |
| 173 | /// Test if this object contains a valid regular expression. |
| 174 | /// |
| 175 | /// \return |
| 176 | /// \b true if the regular expression compiled and is ready |
| 177 | /// for execution, \b false otherwise. |
| 178 | bool IsValid() const; |
| 179 | |
| 180 | void Clear() { |
| 181 | Free(); |
| 182 | m_re.clear(); |
| 183 | m_comp_err = 1; |
| 184 | } |
| 185 | |
| 186 | int GetErrorCode() const { return m_comp_err; } |
| 187 | |
| 188 | bool operator<(const RegularExpression &rhs) const; |
| 189 | |
| 190 | private: |
| 191 | // Member variables |
| 192 | std::string m_re; ///< A copy of the original regular expression text |
| 193 | int m_comp_err; ///< Status code for the regular expression compilation |
| 194 | regex_t m_preg; ///< The compiled regular expression |
| 195 | }; |
| 196 | |
| 197 | } // namespace lldb_private |
| 198 | |
| 199 | #endif // liblldb_RegularExpression_h_ |