Re: pg_dump / copy bugs with "big lines" ?

Alvaro Herrera wrote:
> 
> I have now pushed this to 9.5, 9.6 and master.  It could be backpatched
> to 9.4 with ease (just a small change in heap_form_tuple); anything
> further back would require much more effort.

I had to revert this on 9.5 and 9.6 -- it is obvious (in hindsight) that
changing StringInfoData is an ABI break, so we can't do it in back
branches; see
https://www.postgresql.org/message-id/27737.1480993846@sss.pgh.pa.us
The patch still remains in master, with the bugs you pointed out.  I
suppose if somebody is desperate about getting data out from a table
with large tuples, they'd need to use pg10's pg_dump for it.

We could use the highest-order bit in StringInfoData->maxlen to
represent the boolean flag instead, if we really cared.  But I'm not
going to sweat over it ...

-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services