Horton

Re: BUG #17950: Incorrect memory access in gtsvector_picksplit()

Alexander Law <exclusion@gmail.com>

View thread View on postgr.es
From: Alexander Lakhin <exclusion@gmail.com>
To: pgsql-bugs@lists.postgresql.org
Date: 2023-08-13T13:00:01Z
Lists: pgsql-bugs

Attachments

17.06.2023 17:00, Alexander Lakhin wrote:
> 29.05.2023 23:00, PG Bug reporting form wrote:
>> The following bug has been logged on the website:
>>
>> Bug reference:      17950
>
> I managed to reduce the reproducer to the following:
> CREATE TABLE tst(t tsvector);
> INSERT INTO tst SELECT array_to_string(array(SELECT 'a' || x::text FROM generate_series(1, 125) x), ' ')::tsvector 
> FROM generate_series(1, 3000);
> INSERT INTO tst SELECT '' FROM generate_series(1, 100);
> CREATE INDEX gistidx ON tst USING gist (t tsvector_ops(siglen=1));
>
>
> A trivial fix for the issue is attached.
>

I can also propose a regression test addition that demonstrates the valgrind
complaint and also the output of gtsvectorout() for the case ISALLTRUE and
the opposite.

This addition increases the duration of `make check -C contrib/pageinspect`
under valgrind by 7-8 seconds for me:
ok 5         - gist                                     2496 ms
->
ok 5         - gist                                     9890 ms

In absence of any objections or other propositions, I'm inclined to register
this bugfix on the commitfest.

Best regards,
Alexander

Commits

  1. Improve description of keys in tsvector

  2. Fix out-of-bound read in gtsvector_picksplit()