Re: and it's not a bunny rabbit, either

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>
To: Heikki Linnakangas <heikki.linnakangas@enterprisedb.com>
Cc: Robert Haas <robertmhaas@gmail.com>, Itagaki Takahiro <itagaki.takahiro@gmail.com>, pgsql-hackers@postgresql.org
Date: 2010-12-29T17:49:19Z
Lists: pgsql-hackers
Heikki Linnakangas <heikki.linnakangas@enterprisedb.com> writes:
> Hmm, I believe the idea of heap_open is to check that the relation is 
> backed by a heap that you can read with heap_beginscan+heap_next. At the 
> moment that includes normal tables, sequences and toast tables. Foreign 
> tables would not fall into that category.

I don't believe that that definition is documented anyplace; if we
decide that's what we want it to mean, some code comments would be in
order.

> Yeah, you're right that most of the callers of heap_open actually want 
> to a tighter check than that.

I think probably most of the physical calls of heap_open are actually
associated with system catalog accesses, and the fact that the code says
heap_open not relation_open has got more to do with copy&paste than any
real thought about what we're specifying.

			regards, tom lane