1 -----------------------------------------------------------------------------
3 -- Module : Text.Regex.Posix
4 -- Copyright : (c) The University of Glasgow 2002
5 -- License : BSD-style (see the file libraries/base/LICENSE)
7 -- Maintainer : libraries@haskell.org
8 -- Stability : experimental
9 -- Portability : non-portable (needs POSIX regexps)
11 -- Interface to the POSIX regular expression library.
13 -----------------------------------------------------------------------------
15 -- ToDo: should have an interface using PackedStrings.
17 module Text.Regex.Posix (
21 -- * Compiling a regular expression
22 regcomp, -- :: String -> Int -> IO Regex
24 -- ** Flags for regcomp
25 regExtended, -- (flag to regcomp) use extended regex syntax
26 regIgnoreCase, -- (flag to regcomp) ignore case when matching
27 regNewline, -- (flag to regcomp) '.' doesn't match newline
29 -- * Matching a regular expression
30 regexec, -- :: Regex -- pattern
31 -- -> String -- string to match
32 -- -> IO (Maybe (String, -- everything before match
33 -- String, -- matched portion
34 -- String, -- everything after match
35 -- [String])) -- subexpression matches
39 #include <sys/types.h>
47 -- | A compiled regular expression
48 newtype Regex = Regex (ForeignPtr CRegex)
50 -- -----------------------------------------------------------------------------
53 -- | Compiles a regular expression
55 :: String -- ^ The regular expression to compile
56 -> Int -- ^ Flags (summed together)
57 -> IO Regex -- ^ Returns: the compiled regular expression
58 regcomp pattern flags = do
59 regex_fptr <- mallocForeignPtrBytes (#const sizeof(regex_t))
60 r <- withCString pattern $ \cstr ->
61 withForeignPtr regex_fptr $ \p ->
62 c_regcomp p cstr (fromIntegral flags)
64 then do addForeignPtrFinalizer regex_fptr ptr_regfree
65 return (Regex regex_fptr)
66 else error "Text.Regex.Posix.regcomp: error in pattern" -- ToDo
68 -- -----------------------------------------------------------------------------
71 -- | Matches a regular expression against a string
72 regexec :: Regex -- ^ Compiled regular expression
73 -> String -- ^ String to match against
74 -> IO (Maybe (String, String, String, [String]))
75 -- ^ Returns: 'Nothing' if the regex did not match the
79 -- 'Just' (everything before match,
81 -- everything after match,
82 -- subexpression matches)
85 regexec (Regex regex_fptr) str = do
86 withCString str $ \cstr -> do
87 withForeignPtr regex_fptr $ \regex_ptr -> do
88 nsub <- (#peek regex_t, re_nsub) regex_ptr
89 let nsub_int = fromIntegral (nsub :: CSize)
90 allocaBytes ((1 + nsub_int) * (#const sizeof(regmatch_t))) $ \p_match -> do
91 -- add one because index zero covers the whole match
92 r <- c_regexec regex_ptr cstr (1 + nsub) p_match 0{-no flags for now-}
94 if (r /= 0) then return Nothing else do
96 (before,match,after) <- matched_parts str p_match
99 mapM (unpack str) $ take nsub_int $ tail $
100 iterate (`plusPtr` (#const sizeof(regmatch_t))) p_match
102 return (Just (before, match, after, sub_strs))
104 matched_parts :: String -> Ptr CRegMatch -> IO (String, String, String)
105 matched_parts string p_match = do
106 start <- (#peek regmatch_t, rm_so) p_match :: IO (#type regoff_t)
107 end <- (#peek regmatch_t, rm_eo) p_match :: IO (#type regoff_t)
108 let s = fromIntegral start; e = fromIntegral end
109 return ( take s string,
110 take (e-s) (drop s string),
113 unpack :: String -> Ptr CRegMatch -> IO (String)
114 unpack string p_match = do
115 start <- (#peek regmatch_t, rm_so) p_match :: IO (#type regoff_t)
116 end <- (#peek regmatch_t, rm_eo) p_match :: IO (#type regoff_t)
117 -- the subexpression may not have matched at all, perhaps because it
118 -- was optional. In this case, the offsets are set to -1.
119 if (start == -1) then return "" else do
120 return (take (fromIntegral (end-start)) (drop (fromIntegral start) string))
122 -- -----------------------------------------------------------------------------
123 -- The POSIX regex C interface
130 -- Return values from regexec
138 regIgnoreCase = REG_ICASE, \
142 -- Error codes from regcomp
160 foreign import ccall unsafe "regcomp"
161 c_regcomp :: Ptr CRegex -> CString -> CInt -> IO CInt
163 foreign import ccall unsafe "®free"
164 ptr_regfree :: FunPtr (Ptr CRegex -> IO ())
166 foreign import ccall unsafe "regexec"
167 c_regexec :: Ptr CRegex -> CString -> CSize
168 -> Ptr CRegMatch -> CInt -> IO CInt