[r2005-05-24 13:05:26 by afrisch] Empty log message

Original author: afrisch
Date: 2005-05-24 13:05:26+00:00

misc/encodings.ml

@@ -22,43 +22,37 @@ struct
     (i < j) &&
     match s.[i] with
       | '\000'..'\127' -> check s (i+1) j
-      | '\128'..'\223' as c->
+      | '\192'..'\223' as c ->
 	  (i+1 < j) &&
 	  let n1 = Char.code c 
 	  and n2 = Char.code s.[i+1] in
-	  (n2 >= 128) && (n2 <= 191) &&
-	  (((n1 land 0b11111) lsl 6) lor (n2 land 0b111111) >= 128) &&
-	  check s (i+2) j
+	  (n2 lsr 6 == 0b10) 
+	  && let p = ((n1 land 0x1f) lsl 6) lor (n2 land 0x3f) in
+	  (p >= 0x80) && (p < 0x800) 
+	  && check s (i+2) j
       | '\224'..'\239' as c -> 
 	  (i+2 < j) &&
 	  let n1 = Char.code c 
 	  and n2 = Char.code s.[i+1] 
 	  and n3 = Char.code s.[i+2] in
-	  (n2 >= 128) && (n2 <= 191) &&
-	  (n3 >= 128) && (n3 <= 191) &&
-	  let p = ((n1 land 0b1111) lsl 12) lor
-                  ((n2 land 0b111111) lsl 6) lor
-                  (n3 land 0b111111) in
-	  (p >= 0x800) &&
-	  ((p < 0xd800) || (p >= 0xe000)) &&
-	  ((p < 0xfffe) || (p >  0xffff)) &&
-	  check s (i+3) j
+	  (n2 lsr 6 == 0b10) && (n3 lsr 6 == 0b10) 
+	  && let p = ((n1 land 0x0f) lsl 12)  lor ((n2 land 0x3f) lsl 6) 
+	    lor (n3 land 0x3f) in
+	  (p >= 0x800) 
+	  && ((p < 0xd800) || (p >= 0xe000)) 
+	  && ((p < 0xfffe) || (p >  0xffff)) 
+	  &&  check s (i+3) j
       | '\240'..'\247' as c -> 
 	  (i+3 < j) &&
 	  let n1 = Char.code c 
 	  and n2 = Char.code s.[i+1] 
 	  and n3 = Char.code s.[i+2] 
 	  and n4 = Char.code s.[i+3] in
-	  (n2 >= 128) && (n2 <= 191) &&
-	  (n3 >= 128) && (n3 <= 191) &&
-	  (n4 >= 128) && (n4 <= 191) &&
-          let p = ((n1 land 0b111) lsl 18) lor
-                  ((n2 land 0b111111) lsl 12) lor
-                  ((n3 land 0b111111) lsl 6) lor
-                  (n4 land 0b111111)
-          in
-	  (p >= 0x10000) && (p < 0x110000) &&
-	  check s (i+4) j
+	  (n2 lsr 6 == 0b10) && (n3 lsr 6 == 0b10) && (n4 lsr 6 == 0b10) 
+	  && let p = ((n1 land 0x07) lsl 18) lor ((n2 land 0x3f) lsl 12) lor
+            ((n3 land 0x3f) lsl 6) lor (n4 land 0x3f) in
+	  (p >= 0x10000) && (p < 0x110000) 
+	  && check s (i+4) j
       | _ -> false
 
 
@@ -78,7 +72,7 @@ struct
 	  ((Char.code c land 0b11111) lsl 12) lor
 	  ((Char.code s.[i+1] land 0b111111) lsl 6) lor
 	  ((Char.code s.[i+2] land 0b111111))
-      | '\240'..'\248' as c ->
+      | '\240'..'\247' as c ->
 	  ((Char.code c land 0b11111) lsl 18) lor
 	  ((Char.code s.[i+1] land 0b111111) lsl 12) lor
 	  ((Char.code s.[i+2] land 0b111111) lsl 6) lor
@@ -96,7 +90,7 @@ struct
 	  ((Char.code c land 0b1111) lsl 12) lor
 	  ((Char.code s.[i+1] land 0b111111) lsl 6) lor
 	  ((Char.code s.[i+2] land 0b111111)), i + 3
-      | '\240'..'\248' as c ->
+      | '\240'..'\247' as c ->
 	  ((Char.code c land 0b111) lsl 18) lor
 	  ((Char.code s.[i+1] land 0b111111) lsl 12) lor
 	  ((Char.code s.[i+2] land 0b111111) lsl 6) lor
@@ -127,7 +121,7 @@ struct
       | '\000'..'\127' as c -> i + 1
       | '\192'..'\223' as c -> i + 2
       | '\224'..'\239' as c -> i + 3 
-      | '\240'..'\248' as c -> i + 4
+      | '\240'..'\247' as c -> i + 4
       | _ -> failwith "Malformed UTF-8 bufffer"
 (*
   let width = Array.create 256 1