1 -----------------------------------------------------------------------------
3 -- Module : Text.Regex.Posix
4 -- Copyright : (c) The University of Glasgow 2001
5 -- License : BSD-style (see the file libraries/core/LICENSE)
7 -- Maintainer : libraries@haskell.org
8 -- Stability : experimental
9 -- Portability : non-portable (only on platforms that provide POSIX regexps)
11 -- $Id: Posix.hsc,v 1.1 2001/08/02 11:20:50 simonmar Exp $
13 -- Interface to the POSIX regular expression library.
14 -- ToDo: detect regex library with configure.
15 -- ToDo: should have an interface using PackedStrings.
17 -----------------------------------------------------------------------------
19 module Text.Regex.Posix (
22 regcomp, -- :: String -> Int -> IO Regex
24 regexec, -- :: Regex -- pattern
25 -- -> String -- string to match
26 -- -> IO (Maybe (String, -- everything before match
27 -- String, -- matched portion
28 -- String, -- everything after match
29 -- [String])) -- subexpression matches
31 regExtended, -- (flag to regcomp) use extended regex syntax
32 regIgnoreCase, -- (flag to regcomp) ignore case when matching
33 regNewline -- (flag to regcomp) '.' doesn't match newline
43 newtype Regex = Regex (ForeignPtr CRegex)
45 -- -----------------------------------------------------------------------------
48 regcomp :: String -> Int -> IO Regex
49 regcomp pattern flags = do
50 regex_ptr <- mallocBytes (#const sizeof(regex_t))
51 regex_fptr <- newForeignPtr regex_ptr (regfree regex_ptr)
52 withCString pattern $ \cstr -> do
53 r <- c_regcomp regex_fptr cstr (fromIntegral flags)
55 then return (Regex regex_fptr)
56 else error "Text.Regex.Posix.regcomp: error in pattern" -- ToDo
58 regfree :: Ptr CRegex -> IO ()
63 -- -----------------------------------------------------------------------------
66 regexec :: Regex -- pattern
67 -> String -- string to match
68 -> IO (Maybe (String, -- everything before match
69 String, -- matched portion
70 String, -- everything after match
71 [String])) -- subexpression matches
73 regexec (Regex regex_fptr) str = do
74 withUnsafeCString str $ \cstr -> do
75 nsub <- withForeignPtr regex_fptr $ \p -> (#peek regex_t, re_nsub) p
76 let nsub_int = fromIntegral (nsub :: CSize)
77 allocaBytes ((1 + nsub_int) * (#const sizeof(regmatch_t))) $ \p_match -> do
78 -- add one because index zero covers the whole match
79 r <- c_regexec regex_fptr cstr (1 + nsub) p_match 0{-no flags for now-}
81 if (r /= 0) then return Nothing else do
83 (before,match,after) <- matched_parts str p_match
86 mapM (unpack str) $ take nsub_int $ tail $
87 iterate (`plusPtr` (#const sizeof(regmatch_t))) p_match
89 return (Just (before, match, after, sub_strs))
91 matched_parts :: String -> Ptr CRegMatch -> IO (String, String, String)
92 matched_parts string p_match = do
93 start <- (#peek regmatch_t, rm_so) p_match :: IO CInt
94 end <- (#peek regmatch_t, rm_eo) p_match :: IO CInt
95 let s = fromIntegral start; e = fromIntegral end
96 return ( take (s-1) string,
97 take (e-s) (drop s string),
100 unpack :: String -> Ptr CRegMatch -> IO (String)
101 unpack string p_match = do
102 start <- (#peek regmatch_t, rm_so) p_match :: IO CInt
103 end <- (#peek regmatch_t, rm_eo) p_match :: IO CInt
104 -- the subexpression may not have matched at all, perhaps because it
105 -- was optional. In this case, the offsets are set to -1.
106 if (start == -1) then return "" else do
107 return (take (fromIntegral (end-start)) (drop (fromIntegral start) string))
109 -- -----------------------------------------------------------------------------
110 -- The POSIX regex C interface
117 -- Return values from regexec
125 regIgnoreCase = REG_ICASE, \
129 -- Error codes from regcomp
147 foreign import "regcomp" unsafe
148 c_regcomp :: ForeignPtr CRegex -> CString -> CInt -> IO CInt
150 foreign import "regfree" unsafe
151 c_regfree :: Ptr CRegex -> IO ()
153 foreign import "regexec" unsafe
154 c_regexec :: ForeignPtr CRegex -> UnsafeCString -> CSize
155 -> Ptr CRegMatch -> CInt -> IO CInt