| From: | Tatsuo Ishii <ishii(at)postgresql(dot)org> |
|---|---|
| To: | li(dot)evan(dot)chao(at)gmail(dot)com |
| Cc: | kdbase(dot)hack(at)gmail(dot)com, pgsql-hackers(at)lists(dot)postgresql(dot)org |
| Subject: | Re: Fix duplicate detection for null-treatment window functions |
| Date: | 2026-07-18 02:19:20 |
| Message-ID: | 20260718.111920.2170688754062169922.ishii@postgresql.org |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
>>>> Happy to send a patch along these lines if you agree.
>>>>
>>>
>>> I think your analysis is correct, please feel free to post your version.
>>
>> Thanks, Chao. Patch attached.
>>
>> It drops the redundant "&& wfunc->ignore_nulls == perfunc[i].ignore_nulls"
>> term together with WindowStatePerFuncData.ignore_nulls, and rewords the
>> now-stale comment, leaving equal() to distinguish calls that differ only
>> in null treatment -- which it already does, since WindowFunc.ignore_nulls
>> is a plain scalar field that _equalWindowFunc() compares. This fixes the
>> same bug your v2 fixed, but removes the hand-maintained shadow copy that
>> was the root cause, so it can't silently drift out of sync again.
Yeah, in nodeWindowAgg.c, actually ignore_nulls checking is performed
using WindowObject.ignore_nulls. I agree with your analysis.
>> Testing:
>> - Since the patch changes the layout of WindowStatePerFuncData, I also
>> built the whole tree with -fsanitize=alignment,undefined
>> -fno-sanitize-recover=alignment and re-ran make check under it: 245/245,
>> no runtime errors. The removed byte sat right before an 8-byte-aligned
>> pointer (winobj), so it was absorbed by padding and no following field
>> offset changes.
Thanks for the checking.
> V2 overall looks good. One comment: I don’t think we need to add a comment explaining the equal() behavior for ignore_nulls.
>
> The duplicate detection logic is legacy. Commit 25a30bbd4 intended to handle ignore_nulls specially, so it added “which needs the same ignore_nulls value” to the comment. Since equal() already handles ignore_nulls natively, the code change that added special handling for ignore_nulls has been reverted, so I think the comment should just be reverted as well.
Fair enough.
> PFA v3, reverted the comment.
Looks good to me except the comment line needs indentation fixing. v4
patch attached. If there's no objection, I will commit it.
Regards,
--
Tatsuo Ishii
SRA OSS K.K.
English: http://www.sraoss.co.jp/index_en/
Japanese:http://www.sraoss.co.jp
| Attachment | Content-Type | Size |
|---|---|---|
| v4-0001-Remove-redundant-null-treatment-check-in-window-f.patch | application/octet-stream | 2.9 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Chao Li | 2026-07-18 02:28:47 | Re: Fix duplicate detection for null-treatment window functions |
| Previous Message | Lukas Fittl | 2026-07-18 01:04:22 | Re: pg_plan_advice: fix parsing underscore in numbers |