If you take a look at, for instance, the Num library, or the types of
the infixes in Pervasives, you'll notice that the infix is parenthesized.
That is also how you declare new ones, or redefine the existing ones.
For example,
# let (+@) x y = x + y - 1;;
val ( +@ ) : int -> int -> int = <fun>
# 1 +@ 0;;
- : int = 0
and the new (+@) gets its precedence from the first character of the
lexeme, the "+".
Prefixes are defined similarly,
# let (~@) = fun x -> String.uppercase x ;;
val ( ~@ ) : string -> string = <fun>
# ~@ "hi there";;
- : string = "HI THERE"
You can also redefine the existing infixes and prefixes, including the
ones like mod, not, and or, but that would be very nasty.
I wish that there was some mechanism like in Haskell to make infixes out
of identifiers, but of course we would be back to having directives again
to resolve ambiguities in precedence and associativity if this were the
case. Camlp4 helps here.
-- Brian
On Tue, 30 May 2000, Steve Stevenson wrote:
> In caml there are directives to do this. The obvious doesn't work in
> Ocaml.
>
> best regards
>
> steve
>
>
This archive was generated by hypermail 2b29 : Sat Jun 03 2000 - 09:36:48 MET DST