Re: Avoid excessive inlining? - Mailing list pgsql-general

From Philip Semanchuk
Subject Re: Avoid excessive inlining?
Date
Msg-id B419713D-368F-433A-986F-D6027921D743@americanefficient.com
Whole thread Raw
In response to Re: Avoid excessive inlining?  (Laurenz Albe <laurenz.albe@cybertec.at>)
List pgsql-general

> On Dec 22, 2020, at 8:40 AM, Laurenz Albe <laurenz.albe@cybertec.at> wrote:
>
> On Mon, 2020-12-21 at 11:45 -0500, Philip Semanchuk wrote:
>>> On Dec 19, 2020, at 12:59 AM, Joel Jacobson <joel@compiler.org> wrote:
>>> Is there a way to avoid excessive inlining when writing pure SQL functions, without having to use PL/pgSQL?
>>
>> The rules for inlining are here:
>> https://wiki.postgresql.org/wiki/Inlining_of_SQL_functions
>>
>> According to those rules, if you declared your SQL function as VOLATILE, then Postgres wouldn’t
>> inline it. From your question, I’m not sure if you want to have the same function inlined
>> sometimes and not others. I can’t think of a way to do that offhand.
>
> Where do you see that?  As far as I know, VOLATILE is the best choice if you
> want the function to be inlined.

Ugh, you’re absolutely right, and I’m sorry for spreading misinformation. That’s what I get from quoting from memory
ratherthan reading the link that I posted.  


>
> I would say that the simplest way to prevent a function from being inlined
> is to set a parameter on it:
>
>  ALTER FUNCTION f() SET enable_seqscan = on;

I appreciate the correction and education.

Cheers
Philip


pgsql-general by date:

Previous
From: Laurenz Albe
Date:
Subject: Re: Avoid excessive inlining?
Next
From: "Joel Jacobson"
Date:
Subject: Re: Avoid excessive inlining?