"" <wcting163@163.com> writes:
> 1). although i kown that, "the system will convert **now** to a timestamp as
> soon as the constant is parsed",
> i think this is a bug.
Sorry, it's not a bug, and we're not going to change it.
> but, what about "prepare p1 as select 'today'::timestamp"??
> today() is not a builtin function, we can't change it to "select
> today()::timestamp";
You can get that by using date_trunc on the result of now(); or there's
current_date, which is actually SQL-standard unlike these other things.
regards, tom lane