Re: Oid registry - Mailing list pgsql-hackers

From Peter Eisentraut
Subject Re: Oid registry
Date
Msg-id 506227C0.5080605@gmx.net
Whole thread Raw
In response to Re: Oid registry  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On 9/24/12 11:39 PM, Tom Lane wrote:
> My recollection of the PGCon discussion is that people wanted to allow
> client-side code to hard-wire type OIDs for add-on types, in more or
> less the same way that things like JDBC know that "25" is "text".

If I write a custom uint type and want to explain to JDBC that it is
number-like, I don't think it's very attractive to have to register a
hard-wired OID for it.  There would not only be the burden of
maintaining this OID database, JDBC and every driver would on the other
hand have to maintain its own database of all custom types in the world.

I would rather imagine a system where the type communicates its
properties to the client using some sort of label system, like "number",
"string", "pair of number" (point), "pair of string" (hstore).  Or
something with more or less detail.  Maybe something like that already
exists (maybe somewhere in thrift, protocol buffers, etc.?).



pgsql-hackers by date:

Previous
From: Peter Eisentraut
Date:
Subject: Re: Oid registry
Next
From: Peter Eisentraut
Date:
Subject: Re: Oid registry