That's an interesting idea.  SO the key string and the value string will really point to the same
location in memory.  

So if I did (string,string ref) Hashtbl.t,
then the string ref would be a pointer to the key string.
Would that allow me to compare two string ref variables together?  Would the comparason of the
locations of the strings mean I wouldn't have to do a full string compare?

so in my module
type index = string ref
type dict = (string,index) Hashtbl.t

So in the case of
let a = Hashtbl.find dict "a"
let b = Hashtbl.find dict "a"

would a==b be the same as !a == !b?  
Also, would a==b be a faster comparason in this case?

thanx
bob




--- Yamagata Yoriyuki <yoriyuki@mbg.ocn.ne.jp> wrote:
> From: Bob Bailey <bobbaileyjr@yahoo.com>
> Subject: [Caml-list] Question about string ref and Hashtbl.
> Date: Fri, 19 Mar 2004 14:03:23 -0800 (PST)
> 
> > type t = 
> >   {
> >    mutable last:int;
> >    names: (string,int) Hashtbl.t;
> >    indexes:(int,string ref) Hashtbl.t;
> >   }
> 
> If I were you, I would define the dictonary as
> 
> type t = (string, string) Hashtbl.t
> 
> and add function as
> 
> let add dict str =
>     try
> 	Hashtbl.find dict str
>     with  Not_found ->
> 	  Hashtbl.add dict str str
> 
> In addition, if some of your strings become unused during exectuion,
> I would recommend useing WeakHash.
> 
> module Dict = Weak.Make (struct 
>        type t = string 
>        let equal = (=) 
>        let hash = Hashtbl.hash
> end)
> 
> and use Dict as Hasthtbl.  In this way, if a string becomes unused,
> it is GC-ed and removed from the dict.
> 
> --
> Yamagata Yoriyuki


__________________________________
Do you Yahoo!?
Yahoo! Mail - More reliable, more storage, less spam
http://mail.yahoo.com

-------------------
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