diff --git a/Codec/Binary/UTF8/Generic.hs b/Codec/Binary/UTF8/Generic.hs
index c807c4b6a5c0743c2e027c8018c1e8da4293400d..f03cec68ce8d0aaaf6a91bb810d421e4c891202e 100644
--- a/Codec/Binary/UTF8/Generic.hs
+++ b/Codec/Binary/UTF8/Generic.hs
@@ -32,6 +32,10 @@ import Prelude hiding (take,drop,splitAt,span,break,foldr,foldl,length,lines,nul
import Codec.Binary.UTF8.String(encode)
+#ifdef BYTESTRING_IN_BASE
+import Data.ByteString.Base (unsafeHead, unsafeTail)
+#endif
+
class (Num s, Ord s) => UTF8Bytes b s | b -> s where
bsplit :: s -> b -> (b,b)
bdrop :: s -> b -> b
@@ -45,7 +49,7 @@ class (Num s, Ord s) => UTF8Bytes b s | b -> s where
instance UTF8Bytes B.ByteString Int where
bsplit = B.splitAt
bdrop = B.drop
- buncons = B.uncons
+ buncons = unconsB
elemIndex = B.elemIndex
empty = B.empty
null = B.null
@@ -55,7 +59,7 @@ instance UTF8Bytes B.ByteString Int where
instance UTF8Bytes L.ByteString Int64 where
bsplit = L.splitAt
bdrop = L.drop
- buncons = L.uncons
+ buncons = unconsL
elemIndex = L.elemIndex
empty = L.empty
null = L.null
@@ -267,3 +271,20 @@ lines' bs = case elemIndex 10 bs of
in xs : lines' ys
Nothing -> [bs]
+-----------
+-- Compatibility functions for base-2
+
+unconsB :: B.ByteString -> Maybe (Word8,B.ByteString)
+unconsL :: L.ByteString -> Maybe (Word8,L.ByteString)
+
+#ifdef BYTESTRING_IN_BASE
+unconsB bs | B.null bs = Nothing
+ | otherwise = Just (unsafeHead bs, unsafeTail bs)
+
+unconsL bs = case L.toChunks bs of
+ (x:xs) | not (B.null x) -> Just (unsafeHead x, L.fromChunks (unsafeTail x:xs))
+ _ -> Nothing
+#else
+unconsB = B.uncons
+unconsL = L.uncons
+#endif
diff --git a/Data/ByteString/Lazy/UTF8.hs b/Data/ByteString/Lazy/UTF8.hs
index 267f4b0821e064390642d1c885d9523fd6262ec3..b1ec7ad7419a3a3cbfdf6456f61f9371f8a0b376 100644
--- a/Data/ByteString/Lazy/UTF8.hs
+++ b/Data/ByteString/Lazy/UTF8.hs
@@ -28,6 +28,7 @@ import qualified Data.ByteString.Lazy as B
import Prelude hiding (take,drop,splitAt,span,break,foldr,foldl,length,lines)
import Codec.Binary.UTF8.String(encode)
+import Codec.Binary.UTF8.Generic (buncons)
-- | Converts a Haskell string into a UTF8 encoded bytestring.
fromString :: String -> B.ByteString
@@ -51,7 +52,7 @@ replacement_char = '\xfffd'
-- XXX: Should we combine sequences of errors into a single replacement
-- character?
decode :: B.ByteString -> Maybe (Char,Int64)
-decode bs = do (c,cs) <- B.uncons bs
+decode bs = do (c,cs) <- buncons bs
return (choose (fromEnum c) cs)
where
choose :: Int -> B.ByteString -> (Char, Int64)
@@ -75,7 +76,7 @@ decode bs = do (c,cs) <- B.uncons bs
{-# INLINE get_follower #-}
get_follower :: Int -> B.ByteString -> Maybe (Int, B.ByteString)
- get_follower acc cs = do (x,xs) <- B.uncons cs
+ get_follower acc cs = do (x,xs) <- buncons cs
acc1 <- follower acc x
return (acc1,xs)
diff --git a/Data/ByteString/UTF8.hs b/Data/ByteString/UTF8.hs
index e6741153ceeb4ef340a442c018d34930265fbebf..76acaed31e493571010e93785387cd74075826f2 100644
--- a/Data/ByteString/UTF8.hs
+++ b/Data/ByteString/UTF8.hs
@@ -27,6 +27,7 @@ import qualified Data.ByteString as B
import Prelude hiding (take,drop,splitAt,span,break,foldr,foldl,length,lines)
import Codec.Binary.UTF8.String(encode)
+import Codec.Binary.UTF8.Generic (buncons)
-- | Converts a Haskell string into a UTF8 encoded bytestring.
fromString :: String -> B.ByteString
@@ -50,7 +51,7 @@ replacement_char = '\xfffd'
-- XXX: Should we combine sequences of errors into a single replacement
-- character?
decode :: B.ByteString -> Maybe (Char,Int)
-decode bs = do (c,cs) <- B.uncons bs
+decode bs = do (c,cs) <- buncons bs
return (choose (fromEnum c) cs)
where
choose :: Int -> B.ByteString -> (Char, Int)
@@ -74,7 +75,7 @@ decode bs = do (c,cs) <- B.uncons bs
{-# INLINE get_follower #-}
get_follower :: Int -> B.ByteString -> Maybe (Int, B.ByteString)
- get_follower acc cs = do (x,xs) <- B.uncons cs
+ get_follower acc cs = do (x,xs) <- buncons cs
acc1 <- follower acc x
return (acc1,xs)
diff --git a/System/IO/UTF8.hs b/System/IO/UTF8.hs
index be4796c72bede0d41f2b83840edc60e1f209f308..e5f88fcf3b451407bdcd1380f238b3d105646932 100644
--- a/System/IO/UTF8.hs
+++ b/System/IO/UTF8.hs
@@ -36,6 +36,7 @@ import Prelude (String, (=<<), (.), map, Enum(toEnum, fromEnum), Read,
Show(..))
import System.IO (Handle, IO, FilePath, IOMode(AppendMode, ReadMode, WriteMode))
import qualified System.IO as IO
+import Control.Exception (bracket)
import Codec.Binary.UTF8.String (encode, decode)
@@ -83,7 +84,7 @@ openBinaryFile :: FilePath -> IOMode -> IO Handle
openBinaryFile n m = IO.openBinaryFile (encodeString n) m
withBinaryFile :: FilePath -> IOMode -> (Handle -> IO a) -> IO a
-withBinaryFile n m f = IO.withBinaryFile (encodeString n) m f
+withBinaryFile n m f = bracket (openBinaryFile n m) IO.hClose f
-- | The 'readFile' function reads a file and
-- returns the contents of the file as a UTF8 string.
diff --git a/utf8-string.cabal b/utf8-string.cabal
index 437d86529062e40b976c34bd5dc63e389966c3e1..0585cf35bb0503c18f91d0c3f6c7b275cb0d01ad 100644
--- a/utf8-string.cabal
+++ b/utf8-string.cabal
@@ -1,5 +1,5 @@
Name: utf8-string
-Version: 0.3.3
+Version: 0.3.4
Author: Eric Mertens
Maintainer: emertens@galois.com
License: BSD3
@@ -21,9 +21,11 @@ library
if flag(bytestring-in-base)
build-depends: base >= 2.0 && < 2.2
+ cpp-options: -DBYTESTRING_IN_BASE
else
build-depends: base < 2.0 || >= 3, bytestring >= 0.9
+ Extensions: CPP
Exposed-modules: Codec.Binary.UTF8.String
Codec.Binary.UTF8.Generic
System.IO.UTF8