On Mon, Jul 6, 2015 at 10:08 AM, Joe Conway <m...@joeconway.com> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 07/06/2015 12:39 AM, Michael Paquier wrote:
> >> Yeah, particularly the use of first_optarg makes things harder
> >> to follow in the code with this patch. A C wrapper has the
> >> disadvantage to decentralize the argument checks to many places
> >> making the flow harder to follow hence using
> >> get_fn_expr_argtype() with PG_NARGS would be the way to go, at
> >> least to me. This way, you can easily find how many arguments
> >> there are, and which value is assigned to which variable before
> >> moving on to the real processing.
> >
> > Just to be clear I mean that: if (PG_NARGS() == 5) { if
> > (get_fn_expr_argtype(fcinfo->flinfo, 1) == TYPEOID) var =
> > PG_GETARG_BOOL(1) [...]
>
> Actually, I had in mind something like:
> 8<---------------------
> int i;
> int numargs;
> int *argtypes;
>
> numargs = PG_NARGS();
> argtypes = palloc(numargs * sizeof(int));
> for (i = 0; i < numargs; i++)
> argtypes[i] = get_fn_expr_argtype(fcinfo->flinfo, i);
>
> if ((numargs == 4 || numargs == 5) &&
> argtypes[0] == TEXTOID &&
> argtypes[1] == TEXTOID &&
> argtypes[2] == INT4OID &&
> argtypes[3] == BOOLOID)
> {
> [...]
> }
> else if ((numargs == 3 || numargs == 4) &&
> argtypes[0] == TEXTOID &&
> argtypes[1] == INT4OID &&
> argtypes[2] == BOOLOID)
> {
> [...]
> 8<---------------------
> etc.
>
> If the highest number of arguments is always checked first, the
> pattern is not ambiguous even with the extra anyelement.
>
>
Utterly reasonable and do-able. I'll get to work on that soonish, pending
resolution of other's concerns.
