Torok Edwin | ce0c81e | 2009-08-30 08:24:09 +0000 | [diff] [blame] | 1 | //===-- Regex.h - Regular Expression matcher implementation -*- C++ -*-----===// |
| 2 | // |
| 3 | // The LLVM Compiler Infrastructure |
| 4 | // |
| 5 | // This file is distributed under the University of Illinois Open Source |
| 6 | // License. See LICENSE.TXT for details. |
| 7 | // |
| 8 | //===----------------------------------------------------------------------===// |
| 9 | // |
| 10 | // This file implements a POSIX regular expression matcher. |
| 11 | // |
| 12 | //===----------------------------------------------------------------------===// |
| 13 | |
| 14 | #include "llvm/ADT/SmallVector.h" |
| 15 | #include "llvm/ADT/StringRef.h" |
| 16 | |
| 17 | struct llvm_regex; |
| 18 | namespace llvm { |
| 19 | class Regex { |
| 20 | public: |
| 21 | enum { |
| 22 | /// Compile with support for subgroup matches, this is just to make |
| 23 | /// constructs like Regex("...", 0) more readable as Regex("...", Sub). |
| 24 | Sub=0, |
| 25 | /// Compile for matching that ignores upper/lower case distinctions. |
| 26 | IgnoreCase=1, |
| 27 | /// Compile for matching that need only report success or failure, |
| 28 | /// not what was matched. |
| 29 | NoSub=2, |
| 30 | /// Compile for newline-sensitive matching. With this flag '[^' bracket |
| 31 | /// expressions and '.' never match newline. A ^ anchor matches the |
| 32 | /// null string after any newline in the string in addition to its normal |
| 33 | /// function, and the $ anchor matches the null string before any |
| 34 | /// newline in the string in addition to its normal function. |
| 35 | Newline=4 |
| 36 | }; |
| 37 | |
| 38 | /// Compiles the given POSIX Extended Regular Expression \arg Regex. |
| 39 | /// This implementation supports regexes and matching strings with embedded |
| 40 | /// NUL characters. |
| 41 | Regex(const StringRef &Regex, unsigned Flags=NoSub); |
| 42 | ~Regex(); |
| 43 | |
| 44 | /// isValid - returns the error encountered during regex compilation, or |
| 45 | /// matching, if any. |
| 46 | bool isValid(std::string &Error); |
| 47 | |
| 48 | /// matches - Match the regex against a given \arg String. |
| 49 | /// |
| 50 | /// \param Matches - If given, on a succesful match this will be filled in |
| 51 | /// with references to the matched group expressions (inside \arg String), |
| 52 | /// the first group is always the entire pattern. |
| 53 | /// By default the regex is compiled with NoSub, which disables support for |
| 54 | /// Matches. |
| 55 | /// For this feature to be enabled you must construct the regex using |
| 56 | /// Regex("...", Regex::Sub) constructor. |
Torok Edwin | ce0c81e | 2009-08-30 08:24:09 +0000 | [diff] [blame] | 57 | bool match(const StringRef &String, SmallVectorImpl<StringRef> *Matches=0); |
| 58 | private: |
| 59 | struct llvm_regex *preg; |
| 60 | int error; |
| 61 | bool sub; |
| 62 | }; |
| 63 | } |