Re: Index AM API cleanup - Mailing list pgsql-hackers

From Peter Eisentraut
Subject Re: Index AM API cleanup
Date
Msg-id f0feec67-31ea-4124-9139-d16abc98ee50@eisentraut.org
Whole thread Raw
In response to Re: Index AM API cleanup  (Peter Eisentraut <peter@eisentraut.org>)
Responses Re: Index AM API cleanup
List pgsql-hackers
On 20.03.25 12:59, Peter Eisentraut wrote:
> v22-0006-Convert-from-StrategyNumber-to-CompareType.patch
> 
> This is all that remains now.  I think with a bit more polishing around 
> the edges, some comment updates, etc., this is close to ready.

Here is an updated version of this patch.  I have left out all the extra 
tests and WIP patches etc. from the series for now so that the focus is 
clear.

This patch is mostly unchanged from the above, except some small amount 
of updating comments, as well as the following.

I've done a fair bit of performance testing to make sure there are no 
noticeable regressions from this patch.  I've found that the function 
get_mergejoin_opfamilies() is quite critical to the planning time of 
even simple queries (such as pgbench --select-only), so I played around 
with various caching schemes.  In the end, I just settled on hardcoding 
information about the built-in index AM types.  Which is of course ugly, 
but at least it's essentially the same as before.  If we find other 
affected hotspots, we could apply similar workarounds, but so far I 
haven't found any.

Attachment

pgsql-hackers by date:

Previous
From: Nazir Bilal Yavuz
Date:
Subject: Re: Using read stream in autoprewarm
Next
From: Christoph Berg
Date:
Subject: Re: pgsql: Add support for OAUTHBEARER SASL mechanism