Check upper code point limit when decoding 4-byte UTF-8 characters

glguy · Nov 16, 2009 · 97025cd · 97025cd
1 parent f8d8d70
commit 97025cd
Show file tree

Hide file tree

Showing 3 changed files with 6 additions and 6 deletions.
diff --git a/Codec/Binary/UTF8/Generic.hs b/Codec/Binary/UTF8/Generic.hs
@@ -158,8 +158,8 @@ decode bs = do (c,cs) <- buncons bs
         case get_follower d1 cs1 of
           Just (d2, cs2) ->
             case get_follower d2 cs2 of
-              Just (d,_) | d >= 0x10000 -> (toEnum d, 4)
-                         | otherwise    -> (replacement_char, 4)
+              Just (d,_) | d >= 0x10000 && d < 0x110000 -> (toEnum d, 4)
+                         | otherwise                    -> (replacement_char, 4)
               _ -> (replacement_char, 3)
           _ -> (replacement_char, 2)
       _ -> (replacement_char, 1)

diff --git a/Data/ByteString/Lazy/UTF8.hs b/Data/ByteString/Lazy/UTF8.hs
@@ -114,8 +114,8 @@ decode bs = do (c,cs) <- buncons bs
         case get_follower d1 cs1 of
           Just (d2, cs2) ->
             case get_follower d2 cs2 of
-              Just (d,_) | d >= 0x10000 -> (toEnum d, 4)
-                         | otherwise    -> (replacement_char, 4)
+              Just (d,_) | d >= 0x10000 && d < 0x110000 -> (toEnum d, 4)
+                         | otherwise                    -> (replacement_char, 4)
               _ -> (replacement_char, 3)
           _ -> (replacement_char, 2)
       _ -> (replacement_char, 1)

diff --git a/Data/ByteString/UTF8.hs b/Data/ByteString/UTF8.hs
@@ -113,8 +113,8 @@ decode bs = do (c,cs) <- buncons bs
         case get_follower d1 cs1 of
           Just (d2, cs2) ->
             case get_follower d2 cs2 of
-              Just (d,_) | d >= 0x10000 -> (toEnum d, 4)
-                         | otherwise    -> (replacement_char, 4)
+              Just (d,_) | d >= 0x10000 && d < 0x110000 -> (toEnum d, 4)
+                         | otherwise                    -> (replacement_char, 4)
               _ -> (replacement_char, 3)
           _ -> (replacement_char, 2)
       _ -> (replacement_char, 1)