David Wheeler wrote:
> On Saturday, December 14, 2002, at 03:24  PM, Bruce Momjian wrote:
> 
> > I can't think of an ultimate solution.  If we added a length to the
> > passed query string, so you could embed nulls, we would still have the
> > problem of handling the null once we split the query up into data
> > elements.  The only solution there would be to add a length to all
> > passed data values, so we could handle nulls in them, but I doubt it is
> > worth the effort and added code complexity.
> 
> No, but it might be worth it if PostgreSQL's PREPAREd statements 
> created real functions. Then each value could be passed to the back-end 
> individually, rather than in one big string. so instead of
> 
>    "execute stmt('val1', 'val2')"
> 
> we could do:
> 
>    stmt(val1, val2);
> 
> where "val1" and "val2" are actually variables that hold the values 
> being passed.

Yes, thanks David.  You are correct, and I meant to talk about that, but
by the time I got to the end of the email, I had forgotten.

We are looking to overhaul the wire protocol for 7.4, and one of the
TODO items is:

        o Special passing of binary values in platform-neutral format (bytea?)

What this will allow us to do is to pass binary to the server as a
self-contained value, rather than inside the query string.  This will
make things much easier for binary data.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  [EMAIL PROTECTED]               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073

Reply via email to