Joe Conway wrote:
> Tom Lane wrote:
>> With Joe's recent additions for polymorphic functions, it's even
>> possible for the function to discover what it's supposed to return
>> at runtime. (Hey Joe, did we make that work for functions called
>> from the FROM clause? If not, seems like something to fix up.)
>
> Yeah -- the best example is the new hash based crosstab function in
> contrib/tablefunc:
>
Dooh! I just reread what you wrote. Your referring to the FuncExpr
addition to the FmgrInfo struct. I'll have to check that out and let you
know. I'm getting ready to dive back into that stuff in the next day or so.
Joe
