Re: pgsql: Allow for plugin control over path generation strategies.

From: Aleksander Alekseev <aleksander(at)tigerdata(dot)com>
To: pgsql-committers(at)lists(dot)postgresql(dot)org
Cc: Robert Haas <robertmhaas(at)gmail(dot)com>, Michael Paquier <michael(at)paquier(dot)xyz>
Subject: Re: pgsql: Allow for plugin control over path generation strategies.
Date: 2026-01-29 13:53:29
Message-ID: CAJ7c6TPtYQc6wFwhSDU5u5Fs5nE676VUX6MGrT3Db-oZ+0gUiQ@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-committers

Hi Robert,

> Thanks for the report. 4020b370f214315b8c10430301898ac21658143f
> includes an attempt at a fix.

Not sure if keeping unitized structure is a good solution in the long
run. However if you are aiming for performance instead of:

```
+ /* Clear fields that cost_material() will consult */
+ matpath.parallel_workers = 0;
matpath.parent = NULL;
```

... you could just rewrite the `if` condition in cost_material():

```
if (path->parent != NULL && /* <--- is false */
path->parallel_workers == 0 &&
```

--
Best regards,
Aleksander Alekseev

In response to

Responses

Browse pgsql-committers by date

  From Date Subject
Next Message Robert Haas 2026-01-29 14:11:55 Re: pgsql: Allow for plugin control over path generation strategies.
Previous Message Aleksander Alekseev 2026-01-29 13:48:47 Re: pgsql: Allow for plugin control over path generation strategies.