Consider the following small testcase:
BEGIN;
CREATE OR REPLACE FUNCTION upper(IN varchar, OUT varchar)
LANGUAGE SQL STRICT IMMUTABLE
AS
$$
SELECT pg_catalog.upper($1)::varchar;
$$;
CREATE TABLE foo(value varchar);
INSERT INTO foo SELECT 'helmle' FROM generate_series(1, 1000);
CREATE INDEX foo_upper_index ON foo(public.UPPER(value));
ANALYZE foo;
-- assertion failure
EXPLAIN SELECT 1 FROM foo WHERE UPPER(value) = 'xyz';
COMMIT;
While the sense of the UPPER() function is surely debatable (but it is used
in this way in the database where the assertion failure was triggered), the
EXPLAIN in the script crashes with
TRAP: FailedAssertion("!(((array)->elemtype) == elmtype)", File:
"arrayfuncs.c", Line: 2970)
LOG: server process (PID 15451) was terminated by signal 6: Abort trap
It took me a while to strip that down to this script, because it only
happens when the statistics slots of the index are examined for the
constant value in the where clause. The stavalues array holds type oid
1043, whereas type oid 25 is expected.
I tried it back from current -HEAD to 8.3.11 and managed to reproduce it
everywhere. Non-cassert builds are working correctly, so i'm not sure
wether this is an over-aggressive assert() or masks a problem from
somewhere else.
--
Thanks
Bernd
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
