more elogs hit by sqlsmith (Re: amvalidate(): cache lookup failed for operator class 123)

From: Justin Pryzby <pryzby(at)telsasoft(dot)com>
To: Robert Haas <robertmhaas(at)gmail(dot)com>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: more elogs hit by sqlsmith (Re: amvalidate(): cache lookup failed for operator class 123)
Date: 2023-04-12 21:40:20
Message-ID: ZDclRM/jaTe66nce@telsasoft.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Mon, Feb 13, 2023 at 07:50:53AM -0600, Justin Pryzby wrote:
> On Thu, May 13, 2021 at 12:01:22PM -0500, Justin Pryzby wrote:
> > postgres=# select amvalidate(123);
> > ERROR: cache lookup failed for operator class 123
> > The usual expectation is that sql callable functions should return null rather
> > than hitting elog().
>
> On Thu, May 13, 2021 at 2:22 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> > Meh. I'm not convinced that that position ought to apply to amvalidate.
>
> On Thu, May 13, 2021 at 04:12:10PM -0400, Robert Haas wrote:
> > I am still of the opinion that we ought to apply it across the board,
> > for consistency. It makes it easier for humans to know which problems
> > are known to be reachable and which are thought to be can't-happen and
> > thus bugs. If we fix cases like this to return a real error code, then
> > anything that comes up as XX000 is likely to be a real bug, whereas if
> > we don't, the things that we're not concerned about have to be
> > filtered out by some other method, probably involving a human being.
> > If the filter that human being has to apply further involves reading
> > Tom Lane's mind and knowing what he will think about a particular
> > report, or alternatively asking him, it just makes complicated
> > something that we could have made simple.
>
> FWIW, here are some other cases from sqlsmith which hit elog()/XX000:
>
> postgres=# select unknownin('');
> ERROR: failed to find conversion function from unknown to text
> postgres=# \errverbose
> ERROR: XX000: failed to find conversion function from unknown to text
> LOCATION: coerce_type, parse_coerce.c:542
>
> postgres=# SELECT pg_catalog.interval( '12 seconds'::interval ,3);
> ERROR: unrecognized interval typmod: 3
>
> postgres=# SELECT pg_describe_object(1,0,1);
> ERROR: invalid non-zero objectSubId for object class 1
>
> postgres=# SELECT pg_read_file( repeat('a',333));
> ERROR: could not open file "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa" for reading: File name too long

postgres=# SELECT acldefault('a',0);
ERROR: unrecognized object type abbreviation: a

postgres=# SELECT setweight('a','1');
ERROR: unrecognized weight: 49

regression=# select float8_regr_intercept(ARRAY[1]) ;
ERROR: float8_regr_intercept: expected 6-element float8 array

None of these is new in v16.

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Daniel Gustafsson 2023-04-12 21:46:01 Re: [PATCH] Add `verify-system` sslmode to use system CA pool for server cert
Previous Message Tom Lane 2023-04-12 21:40:18 Re: [PATCH] Add `verify-system` sslmode to use system CA pool for server cert