printf "%a" vs sprintf "%a"

printf "%a" vs sprintf "%a"

[Tiphaine Turpin]

Hi,

I'am confused by the interpretation of "%a":

# Printf.printf "%a";;
- : (out_channel -> '_a -> unit) -> '_a -> unit = <fun>
# Printf.sprintf "%a";;
- : (unit -> '_a -> string) -> '_a -> string = <fun>

Usually, the typing of formatting functions is such that

printf something

has type unit if and only if

sprintf something

has type string. But %a breaks this rule. Wouldn't it be simpler to have
two separate directives which accept respectively string printers and
channel printers, regardless of the outer printing function ?

Tiphaine

Re: [Caml-list] printf "%a" vs sprintf "%a"

[David Rajchenbach-Teller]

Yeah, the behaviour of *printf is weird in this respect (and that of
Extlib's printf is even a tad worse). For this reason, in Batteries
Included, we have 
 Printf.sprintf  (which behaves as the base Printf.sprintf)
and
 Printf.sprintf2 (which behaves as you expected)

All of this in addition to our [Print] module, which offers a syntax
comparable with [Printf], but allows extending the format rather than
using %a and %t.

Cheers,
 David


On Thu, 2009-03-26 at 17:53 +0100, Tiphaine Turpin wrote:
> Wouldn't it be simpler to have
> two separate directives which accept respectively string printers and
> channel printers, regardless of the outer printing function ?
> 
> Tiphaine
> 
> _______________________________________________
> Caml-list mailing list. Subscription management:
> http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
> Archives: http://caml.inria.fr
> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
> Bug reports: http://caml.inria.fr/bin/caml-bugs

-- 
David Teller-Rajchenbach
 Security of Distributed Systems
  http://www.univ-orleans.fr/lifo/Members/David.Teller
   « Ce matin Un crétin A tué un chercheur. » (air connu)
   Latest News of French Research: System being liquidated. Researchers angry.

Re: [Caml-list] printf "%a" vs sprintf "%a"

[Jérémie Dimino]

Tiphaine Turpin wrote:
> has type string. But %a breaks this rule. Wouldn't it be simpler to have
> two separate directives which accept respectively string printers and
> channel printers, regardless of the outer printing function ?

Note that you can do it with batteries and the new printf.

The default "%a" directive always uses a channel printer:

# Print.printf p"%a";;
- : (unit Batteries.IO.output -> '_a -> unit) -> '_a -> unit = <fun>
# Print.sprintf p"%a";;
- : (unit Batteries.IO.output -> '_a -> unit) -> '_a -> string = <fun>

And you can define a "%foo" directive which always uses a string printer:

# let printer_foo k f x = k (fun out -> String.print out (f x));;
val printer_as : (('a Batteries.IO.output -> unit) -> 'b) -> ('c -> string) -> 'c -> 'b = <fun>
# Print.printf p"%foo";;
- : ('_a -> string) -> '_a -> unit = <fun>
# Print.sprintf p"%foo";;
- : ('_a -> string) -> '_a -> string = <fun>

Cheers,
Jérémie

Re: [Caml-list] printf "%a" vs sprintf "%a"

[Tiphaine Turpin]

Thanks for the replies. sprintf2 is just fine for me.

Tiphaine


Jérémie Dimino a écrit :
> Tiphaine Turpin wrote:
>   
>> has type string. But %a breaks this rule. Wouldn't it be simpler to have
>> two separate directives which accept respectively string printers and
>> channel printers, regardless of the outer printing function ?
>>     
>
> Note that you can do it with batteries and the new printf.
>
> The default "%a" directive always uses a channel printer:
>
> # Print.printf p"%a";;
> - : (unit Batteries.IO.output -> '_a -> unit) -> '_a -> unit = <fun>
> # Print.sprintf p"%a";;
> - : (unit Batteries.IO.output -> '_a -> unit) -> '_a -> string = <fun>
>
> And you can define a "%foo" directive which always uses a string printer:
>
> # let printer_foo k f x = k (fun out -> String.print out (f x));;
> val printer_as : (('a Batteries.IO.output -> unit) -> 'b) -> ('c -> string) -> 'c -> 'b = <fun>
> # Print.printf p"%foo";;
> - : ('_a -> string) -> '_a -> unit = <fun>
> # Print.sprintf p"%foo";;
> - : ('_a -> string) -> '_a -> string = <fun>
>
> Cheers,
> Jérémie
>
> _______________________________________________
> Caml-list mailing list. Subscription management:
> http://yquem.inria.fr/cgi-bin/mailman/listinfo/caml-list
> Archives: http://caml.inria.fr
> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
> Bug reports: http://caml.inria.fr/bin/caml-bugs
>
>