Content-Type: multipart/alternative;
 boundary="------------KZvnoqUz2Cz67CIyLDPbiPWB"
Message-ID: <e0258f8a-b8f5-411b-963f-5499afe78bf2@gmail.com>
Date: Sat, 19 Sep 2026 21:00:00 +0300
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: pgsql: Invalidate RI fast-path metadata on operator family
 changes
To: Amit Langote <amitlangote09@gmail.com>,
 Amit Langote <amitlan@postgresql.org>
Cc: pgsql-committers@lists.postgresql.org
References: <E1x7p34-00000000Mv7-2O5S@gemulon.postgresql.org>
 <CA+HiwqGitHCmO7nvV4nXXYWDkNYA+zVcJj=JNmL3C1O9B0Gz0A@mail.gmail.com>
 <CA+HiwqHO2xu3eR6bZ99qSQgCTnoS9VC4e0KW25bxE5kmkRMGyQ@mail.gmail.com>
Content-Language: en-US
From: Alexander Lakhin <exclusion@gmail.com>
In-Reply-To: 
 <CA+HiwqHO2xu3eR6bZ99qSQgCTnoS9VC4e0KW25bxE5kmkRMGyQ@mail.gmail.com>
Archived-At: 
 <https://www.postgresql.org/message-id/e0258f8a-b8f5-411b-963f-5499afe78bf2%40gmail.com>
Precedence: bulk

This is a multi-part message in MIME format.
--------------KZvnoqUz2Cz67CIyLDPbiPWB
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit

Hello Amit,

19.09.2026 12:14, Amit Langote wrote:
>> I noticed a few BF failures, after pushing the above changes:
>>
>> ...
>>
>> I don't see any relation between these failures and my commit. window
>> runs concurrently with foreign_key, but this commit's additions to the
>> latter are self-contained (a private operator family in its own
>> schema), so I don't see why window's plans would be affected.
>>
>> Does anyone see it differently?
> Forgot to add that the code change is isolated too in that it only
> changes what ri_triggers.c does on a pg_amop invalidation during an RI
> fast-path check, not anything the planner consults. It doesn't touch
> pg_constraint or how FKs feed planning, so I don't see a route from
> this commit's changes to these window plans.  Any effect would be
> timing at most.  I would not claim that I have figured this out.

I've reproduced such failures locally with:
test: test_setup
test: create_index
test: foreign_key window
x100

I think the window failures are caused by these additions:
+create operator family fam using btree;
+create operator class int_ops for type integer using btree family fam as
+  operator 1 <(integer,integer), operator 2 <=(integer,integer),
+  operator 3 =(integer,integer), operator 4 >=(integer,integer),
+  operator 5 >(integer,integer), function 1 btint4cmp(integer,integer);

Best regards,
Alexander
--------------KZvnoqUz2Cz67CIyLDPbiPWB
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: 8bit

<!DOCTYPE html>
<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <div class="moz-cite-prefix">Hello Amit,<br>
      <br>
      19.09.2026 12:14, Amit Langote wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CA+HiwqHO2xu3eR6bZ99qSQgCTnoS9VC4e0KW25bxE5kmkRMGyQ@mail.gmail.com">
      <blockquote type="cite">
        <pre wrap="" class="moz-quote-pre">
I noticed a few BF failures, after pushing the above changes:

...

I don't see any relation between these failures and my commit. window
runs concurrently with foreign_key, but this commit's additions to the
latter are self-contained (a private operator family in its own
schema), so I don't see why window's plans would be affected.

Does anyone see it differently?
</pre>
      </blockquote>
      <pre wrap="" class="moz-quote-pre">
Forgot to add that the code change is isolated too in that it only
changes what ri_triggers.c does on a pg_amop invalidation during an RI
fast-path check, not anything the planner consults. It doesn't touch
pg_constraint or how FKs feed planning, so I don't see a route from
this commit's changes to these window plans.  Any effect would be
timing at most.  I would not claim that I have figured this out.
</pre>
    </blockquote>
    <br>
    I've reproduced such failures locally with:<br>
    test: test_setup<br>
    test: create_index<br>
    test: foreign_key window<br>
    x100<br>
    <br>
    I think the window failures are caused by these additions:<br>
    +create operator family fam using btree;<br>
    +create operator class int_ops for type integer using btree family
    fam as<br>
    +  operator 1 &lt;(integer,integer), operator 2
    &lt;=(integer,integer),<br>
    +  operator 3 =(integer,integer), operator 4 &gt;=(integer,integer),<br>
    +  operator 5 &gt;(integer,integer), function 1
    btint4cmp(integer,integer);<br>
    <br>
    Best regards,<br>
    Alexander<br>
  </body>
</html>

--------------KZvnoqUz2Cz67CIyLDPbiPWB--