Re: recovery modules - Mailing list pgsql-hackers

From Andres Freund
Subject Re: recovery modules
Date
Msg-id 20230208173344.ku22sogm6jgr27cl@awork3.anarazel.de
Whole thread Raw
In response to Re: recovery modules  (Nathan Bossart <nathandbossart@gmail.com>)
Responses Re: recovery modules
List pgsql-hackers
Hi,

On 2023-02-08 09:27:05 -0800, Nathan Bossart wrote:
> On Wed, Feb 08, 2023 at 08:27:13AM -0800, Andres Freund wrote:
> > One minor thing: I don't think we really need the AssertVariableIsOfType() for
> > anything but the Init() one?
> 
> This is another part that was borrowed from logical decoding output
> plugins.

I know :(. It was needed in an earlier version of the output plugin interface,
where all the different callbacks were looked up via dlsym(), but should have
been removed after that.


> I'm not sure this adds much since f2b73c8 ("Add central
> declarations for dlsym()ed symbols").  Perhaps we should remove all of
> these assertions for functions that now have central declarations.

Most of them weren't needed even before that.

And yes, I'd be for a patch to remove all of those assertions.

Greetings,

Andres Freund



pgsql-hackers by date:

Previous
From: Nathan Bossart
Date:
Subject: Re: recovery modules
Next
From: Peter Geoghegan
Date:
Subject: Re: [PATCH] Make ON CONFLICT DO NOTHING and ON CONFLICT DO UPDATE consistent