Re: WIP: Allow SQL-language functions to reference parameters by parameter name

On Tue, Apr 5, 2011 at 3:47 PM, Peter Eisentraut <peter_e(at)gmx(dot)net> wrote:
> On tis, 2011-04-05 at 14:45 -0500, Merlin Moncure wrote:
>> Talking about the standards compliance of functions is a bit silly:
>> our implementation of functions isn't even close to approximating what
>> looks to be the standard
>
> That doesn't mean it couldn't be better in the future.  We shouldn't
> take it further away, in any case.
>
> As long as we use LANGUAGE SQL, we are both technically and morally in
> standards-space.

sql standard functions are psm routines aiui. Are you making the case
that 'language sql' in postgresql could or should in fact be psm at
some point in the future? I say that's not the case -- our 'language
sql' is not psm.

That said, if you well and truly stated that it was project objective
to allow psm constructions in 'language sql', and you could figure out
a way to do that without breaking current sql code, I would have to
say i'm coming around to your point of view. Either way, our $N
notation is already non-standard and highly in use - what's the big
deal about making it more useful?

merlin