Version française
Home     About     Download     Resources     Contact us    
Browse thread
[Caml-list] A grave bug in logical shift right op?
[ Home ] [ Index: by date | by threads ]
[ Search: ]

[ Message by date: previous | next ] [ Message in thread: previous | next ] [ Thread: previous | next ]
Date: -- (:)
From: Jacques Garrigue <garrigue@k...>
Subject: Re: [Caml-list] A grave bug in logical shift right op?
From: Eray Ozkural <exa@kablonet.com.tr>

> I'm using the debian package in testing....
> 
> orion:fealn-attack$ ledit ocaml
>         Objective Caml version 3.07+2
> 
> # Printf.printf "%X" 0xffffffff;;
> 7FFFFFFF- : unit = ()
> # Printf.printf "%lx" (Int32.of_int 0xffffffff);;
> ffffffff- : unit = ()
> # -1 lsr 24;;
> - : int = 127
> # Int32.shift_right_logical (Int32.of_int (-1)) 24;;
> - : int32 = 255l
> 
> What's going on here? Something's eating the most significant bit in
> both hexadecimal output routine and the logical shift right
> operator. Somebody please help me! My cipher isn't working!!!

Sure, the int type has only 31 bits. In good math,
(2 ^ 31 - 1) quo 2 ^ 24 is 127, as far as I know.

Note that the problem with your first example has already been
discussed. A litteral bigger than the biggest possible integer should
probably trigger an error/exception.
A bit more curions is your second example: since
0xffffffff = 0x7fffffff = -1, converting it to int32 gives -1l (the
32-bit value for -1), which happens to be 0xffffffffl.

If you want to do 32-bit computations, you must use only 32-bit
integers. This shouldn't be too inefficient as long as you use loops
rather than function calls (the same recommandations as for floats
apply, I believe).

Jacques Garrigue

-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners