Re: [Caml-list] Better option to read a file

[Jean-Christophe Filliatre <Jean-Christophe.Filliatre@lri.fr>]

Agustín Valverde writes:

 > Second:
 > 
 > let rec unir c ac = unir ac^(Char.escaped c);;
 > 
 > let leer2 fl =
 >       let form = ref "" in
 >       let c = ref '-' in
 >       let arch = open_in fl in
 >       (try
 >         (while true do (c := input_char arch); (if !c != '\n' then (form 
 > := unir !c !form) else ()) done)
 >       with End_of_file -> close_in arch);
 >       !form;;

Note that this function is very inefficient: you are indeed building a
lot  of intermediate  strings with  "unir", resulting  in  a quadratic
space (even if the final string occupies linear space).

Using a buffer as suggested by Christoph is clearly better (the module
Buffer  from ocaml standard  library is  doubling its  internal string
buffer as needed, without you to worry about it).

 > I also have a parser to convert the string, could I to improve these 
 > functions merging them with the parser in some way?

The use  of ocamllex  in combination  with a buffer  is both  easy and
efficient. I sketch it:

======================================================================
{ 
  open Lexing 
  let buf = Buffer.create 1024
}

rule read = parse
  | "\\n"  { Buffer.add_char buf '\n'; read lexbuf }
  | "\\t"  { Buffer.add_char buf '\t'; read lexbuf }
  | "\\\\" { Buffer.add_char buf '\\'; read lexbuf }
  | _      { Buffer.add_string buf (lexeme lexbuf); read lexbuf }
  | eof    { Buffer.contents buf }

{
  let read_file f =
    let cin = open_in f in
    let lb = from_channel cin in
    let s = read lb in
    close_in cin;
    s
}
======================================================================

-- 
Jean-Christophe Filliâtre (http://www.lri.fr/~filliatr)

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