On 1/10/16 3:44 PM, Andreas Joseph Krogh wrote:
>
> It might be about time to come up with an extension that's a replacement
> for large objects.
>
> What would it take to fund such an extension?
Time and/or money.
> It would "have to" support:
> - Migrate existing LOs away from pg_largeobject
> - Proper driver-support (JDBC)
Would probably be possible to extend JDBC (looks like that's what the
current support does).
> - Possible to exclude from pg_dump
That'd be up to the extension dump control semantics.
> - Support pg_upgrade
>
> And -hackers should agree on the goal to ultimately being merged into
> core and replace pg_largeobject.
Well, there's a reason I suggested an extension. I think it's very
unlikely -hackers would want to add another LO format to the database.
Now-a-days, it's generally preferred to do most things as extensions,
and only incorporate things in the backend that really can't be done
with an extension.
If this theoretical new replacement for LOs took the world by storm and
everyone was using it, maybe it'd be a different thing. The xml and JSON
types are examples of that; they started life as add-ons and were
eventually pulled in because they became extremely popular.
--
Jim Nasby, Data Architect, Blue Treble Consulting, Austin TX
Experts in Analytics, Data Architecture and PostgreSQL
Data in Trouble? Get it in Treble! http://BlueTreble.com