From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Michael Paquier <michael(at)paquier(dot)xyz> |
Cc: | exclusion(at)gmail(dot)com, pgsql-bugs(at)lists(dot)postgresql(dot)org |
Subject: | Re: BUG #18135: Incorrect memory access occurs when attaching a partition with an index |
Date: | 2023-10-02 03:04:11 |
Message-ID: | 1914031.1696215851@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-bugs |
Michael Paquier <michael(at)paquier(dot)xyz> writes:
> FWIW, I was thinking about a case like that with 2 index attributes:
> info1->attr[0] = 0, info1->attr[1] = 1
> info2->attr[0] = 1, info2->attr[1] = 0
> With a mapping from info2 to info1 like that:
> attmap[0] = 1, attmap[1] = 0.
> The code before 9f71e10d6 would have accessed an incorrect pointer.
> The new code would return false, which would not be correct if the
> expressions stored in ii_Expressions for info1/attr0 and info2/attr1
> match?
I think "false" is correct. Otherwise you're claiming that an
index on (expr, col1) is equivalent to an index on (col1, expr).
Maybe it is equivalent for some purposes, but I don't think this
function has any business assuming that it is so for all purposes.
Also, the pre-existing comment clearly intends that false is
the proper result here:
* might differ!) Note that this implies that index columns that are
* expressions appear in the same positions. We will next compare the
* expressions themselves.
The code just failed to enforce that fully.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | PG Bug reporting form | 2023-10-02 08:01:30 | BUG #18142: strange behaviour of "UPDATE" with id_encode() |
Previous Message | Michael Paquier | 2023-10-02 02:56:35 | Re: BUG #18135: Incorrect memory access occurs when attaching a partition with an index |