Tom Lane wrote:
> BTW, I also had a bunch of concerns having to do with odd-seeming
> choices about what information would be wired into libpq and what
> would be retrieved at runtime from the backend. I don't recall the
> details at the moment, but I want to raise a flag that that is still
> an issue for me. I'd like to see some explicit design decisions
> about what information will be treated in which way.
I noticed that too, and looked into it. I didn't see any hard-wired
oids (at least that I remember), but I did see cases where the
scale/precision results had to be accessed based on the specific type
involved, e.g. NUMERIC. I don't see a way around this, and in fact most
people are doing this type-speicific stuff in their apps. The only
other solution I can see is adding a backend function that does these
type-specific manipulations and returns them to the client. This seems
quite attractive, especially considering changes in internal type
representations between releases.
-- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610)
853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill,
Pennsylvania19026