| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Andrew Dunstan <andrew(at)dunslane(dot)net> |
| Cc: | Valeriy Meleshkin <valeriy(at)meleshk(dot)in>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Reproducible coliisions in jsonb_hash |
| Date: | 2022-05-12 13:57:24 |
| Message-ID: | 309877.1652363844@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
> On 2022-05-12 Th 07:02, Valeriy Meleshkin wrote:
>> AFAICT it happens because when iterating over a jsonb the hash function makes no distinction between raw scalars and arrays (it doesn't inspect v.val.array.rawScalar)
> It does look rather like a bug, but I'm unclear about the implications
> of fixing it.
Changing this hash algorithm would break existing hash indexes on jsonb
columns. Maybe there aren't any, but even if so I can't get very excited
about changing this. Hash algorithms always have collisions, and we have
never made any promise that ours are cryptographically strong.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Bruce Momjian | 2022-05-12 14:01:31 | Re: First draft of the PG 15 release notes |
| Previous Message | Bruce Momjian | 2022-05-12 13:53:45 | Re: First draft of the PG 15 release notes |