Re: BUG #17477: A crash bug in transformValuesClause() - Mailing list pgsql-bugs

From Jonathan S. Katz
Subject Re: BUG #17477: A crash bug in transformValuesClause()
Date
Msg-id 6f7f0d34-0f58-41a9-98b5-bf86e1960165@postgresql.org
Whole thread Raw
In response to Re: BUG #17477: A crash bug in transformValuesClause()  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: BUG #17477: A crash bug in transformValuesClause()
Re: BUG #17477: A crash bug in transformValuesClause()
List pgsql-bugs
On 5/9/22 11:25 AM, Tom Lane wrote:
> Masahiko Sawada <sawada.mshk@gmail.com> writes:
>> It seems like transformValuesClause() cannot handle properly the value
>> clause having a relation that has an empty column. Should we raise an
>> error in this case?
> 
> Given that we try to support zero-column relations, I'm not sure why
> we'd insist on disallowing zero-column VALUES.  I think the problem
> is that the code in transformValuesClause needs to be tweaked to
> make that work.  The attached quick hack seems to do the trick.

Agree with the reasoning.

Confirmed reproducing the crash and that this fixes it. I did a short 
double-take on the error message:

    ERROR:  subquery must return only one column

but it is accurate, given this is what the subquery must do, and zero != 
one.

I don't see anything glaring in the code (though I'm not that familiar 
with this part of the codebase), but given this seems like an extreme 
edge case and protects against a crash, I'm satisfied with this.

Jonathan

Attachment

pgsql-bugs by date:

Previous
From: Tom Lane
Date:
Subject: Re: BUG #17478: Missing documents in the index after CREATE INDEX CONCURRENTLY (but existing in the table)
Next
From: Tom Lane
Date:
Subject: Re: BUG #17477: A crash bug in transformValuesClause()