Horton

Re: Index-only scan returns incorrect results when using a composite GIST index with a gist_trgm_ops column.

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

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp>
Cc: sk@zsrv.org, michael.paquier@gmail.com, x4mmm@yandex-team.ru, david@nlpgo.com, pgsql-bugs@lists.postgresql.org, pgsql-hackers@postgresql.org
Date: 2018-03-01T20:39:18Z
Lists: pgsql-bugs, pgsql-hackers 
Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp> writes:
> At Wed, 24 Jan 2018 00:13:51 +0300, Sergei Kornilov <sk@zsrv.org> wrote in <348951516742031@web54j.yandex.ru>
>> Should we also make backport to older versions? I test on REL_10_STABLE - patch builds and works ok, but "make check" fails on new testcase with error:
> CREATE INDEX ON t USING gist (a test_inet_ops, a inet_ops);
> + ERROR:  missing support function 4 for attribute 1 of index "t_a_a1_idx"
>> and with different explain results.

> Thank you for checking that. d3a4f89 allowed that and
> inet_gist_decompress is removed at the same time. Unfortunately I
> didn't find a type on master that has both decompress and fetch
> methods so I prefer to split the regression patch among target
> versions.

I pushed this fix with minor adjustments.  I did not like the proposed
regression test at all: it was overcomplicated and the need for different
versions for different back branches wasn't fun either.  After some poking
around I found that the bug could be exhibited using just btree_gist's
gist_inet_ops, since the core inet_ops class indexes the same datatype and
it does have a fetch function.  So I added a test case in btree_gist.

			regards, tom lane

Commits

  1. Improve stability of recently-added regression test case.

  2. Fix IOS planning when only some index columns can return an attribute.

  3. Ability to advance replication slots

  4. Allow no-op GiST support functions to be omitted.