Horton

Re: BUG #17994: Invalidating relcache corrupts tupDesc inside ExecEvalFieldStoreDeForm()

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

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Andrew Dunstan <andrew@dunslane.net>
Cc: Alvaro Herrera <alvherre@alvh.no-ip.org>, Andres Freund <andres@anarazel.de>, exclusion@gmail.com, pgsql-bugs@lists.postgresql.org
Date: 2023-08-21T16:30:22Z
Lists: pgsql-bugs 
Andrew Dunstan <andrew@dunslane.net> writes:
> On 2023-08-21 Mo 08:33, Alvaro Herrera wrote:
>> Hmm ... when exactly do these values get freed if no longer needed?  Is
>> the theory that leaking them is not relevant?

> Not sure I understand "relevant" here. They don't get freed. There will 
> be at most one entry per row in pg_attribute where atthasmissing is 
> true.

Yeah.  My feeling is that as long as we don't make duplicate hashtable
entries, there will never be so many entries that it'd be worth the
cost and intellectual complexity of trying to clean them up.  When
and if that theory is disproven, we can think harder; but for now
I think this approach is good enough.

			regards, tom lane

Commits

  1. Cache by-reference missing values in a long lived context

  2. Fix order of operations in ExecEvalFieldStoreDeForm().