Re: tupdesc: simplify assert in equalTupleDescs()

From: Nazir Bilal Yavuz <byavuz81(at)gmail(dot)com>
To: Chao Li <li(dot)evan(dot)chao(at)gmail(dot)com>
Cc: PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: tupdesc: simplify assert in equalTupleDescs()
Date: 2026-03-17 08:12:45
Message-ID: CAN55FZ2rRmjAdV-yv6xO6HYpwT1HwXp0vUt1Bse-kmwew_qi-Q@mail.gmail.com
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi,

On Tue, 17 Mar 2026 at 10:22, Chao Li <li(dot)evan(dot)chao(at)gmail(dot)com> wrote:
>
> While reading the code, I saw these assertions in equalTupleDescs():
> ```
> CompactAttribute *cattr1 = TupleDescCompactAttr(tupdesc1, i);
> CompactAttribute *cattr2 = TupleDescCompactAttr(tupdesc2, i);
>
> Assert(cattr1->attnullability != ATTNULLABLE_UNKNOWN);
> Assert((cattr1->attnullability == ATTNULLABLE_UNKNOWN) ==
> (cattr2->attnullability == ATTNULLABLE_UNKNOWN));
>
> ```
>
> The first assertion already guarantees that cattr1->attnullability is not ATTNULLABLE_UNKNOWN, so in the second one the expression cattr1->attnullability == ATTNULLABLE_UNKNOWN will always be false, That means the second assertion is effectively just checking that cattr2->attnullability is also not ATTNULLABLE_UNKNOWN.
>
> So the current code is correct, but it feels a bit harder to read than necessary. This patch just simplifies the second assertion in a direct way.

Thank you for the report! You are right and the patch looks good to me.

Nitpick: It is still a bit hard to understand why
'cattr2->attnullability' should not be equal to 'ATTNULLABLE_UNKNOWN'.
It would be good to add a comment explaining that 'attr2->attnotnull'
should be true too because 'if (attr1->attnotnull !=
attr2->attnotnull)' is returning false.

--
Regards,
Nazir Bilal Yavuz
Microsoft

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Marco Nenciarini 2026-03-17 08:12:50 Re: BUG: Cascading standby fails to reconnect after falling back to archive recovery
Previous Message jian he 2026-03-17 08:05:58 Re: [PATCH] no table rewrite when set column type to constrained domain