public inbox for [email protected]  

help / color / mirror / Atom feed

From: Andrey Borodin <[email protected]>
To: surya poondla <[email protected]>
Cc: songjinzhou <[email protected]>
Cc: dllggyx <[email protected]>
Cc: pgsql-bugs <[email protected]>
Subject: Re: BUG #19382: Server crash at __nss_database_lookup
Date: Thu, 16 Apr 2026 17:00:14 +0500
Message-ID: <[email protected]> (raw)
In-Reply-To: <CAOVWO5o9YOpCTgg6FfNepCoH_6pFSa7TJ3SEWfJAoBvNOb0OdQ@mail.gmail.com>
References: <[email protected]>
	<CAOVWO5rVBKsjG4YwO_PJQu2OBGp8qUdF1jineYY6Lm3zc6-KWQ@mail.gmail.com>
	<CAOVWO5rbwKgHWLYJMvKuvGxW9eFSk7LADk=ZxDEvwA1uTefvAg@mail.gmail.com>
	<CAOVWO5qN8UXdwMnDa+a7aVq=irGSfm2aeYvEc-uV6j4QHZiyrA@mail.gmail.com>
	<CAOVWO5py4zLPYsnGK1EEzHnC4feTSmbsEZn-BiDgY_=J1P6Wmw@mail.gmail.com>
	<CAOVWO5pbgCVx0zgTr1mxZug2hoGwxZOk+-Owvwg0jaQv9JE3Fw@mail.gmail.com>
	<CAOVWO5r19cctAFKbW24jfMKsD-pkyV21w+z7L3pCPxM1CArtjQ@mail.gmail.com>
	<CAOVWO5oSeBouPv0ueVByh+_6EgRCjWh0spSmnF6Cv-TF1twqKg@mail.gmail.com>
	<[email protected]>
	<CAOVWO5oH37CETZuxxXw3dhCMOHPMA0xFoJBWTpfJ06OV7sGzTQ@mail.gmail.com>
	<[email protected]>
	<CAOVWO5oRGPd7mA3d85jNYmjLNfeBAca5oDcHTfRFxbAwPLxs5g@mail.gmail.com>
	<[email protected]>
	<CAOVWO5o9YOpCTgg6FfNepCoH_6pFSa7TJ3SEWfJAoBvNOb0OdQ@mail.gmail.com>



> On 9 Apr 2026, at 09:24, surya poondla <[email protected]> wrote:
> 
> <0005-Fix-bug-19382-server-crash-when-ALTER-TYPE-is-used-m.patch>

I’ve took a look into the patch and here are some thoughts:

1. collect_composite_type_versions() does this
/* Resolve domain types to their base type */
if (typtype == TYPTYPE_DOMAIN)
{
			attrtypid = getBaseType(attrtypid);
			typtype = get_typtype(attrtypid);
}

Only for nested types. Do we need this for root?

2. When we do this

/* Clear composite type snapshot */
rec->nCompTypes = 0;
rec->compTypeOids = NULL;
rec->compTypeVersions = NULL;

We also might need two pfree()s.

3. Here are few other test cases that crash with the patch.

-- Case 5: Dot assignment
create type bug19382_foo3 as (a int, b int);
create function bug19382_test_field_assign() returns record as $$
declare r bug19382_foo3;
begin
r.a := 123;
r.b := power(2, 30)::int4;
alter type bug19382_foo3 alter attribute b type text;
return r;
end;
$$ language plpgsql;
select bug19382_test_field_assign();
drop function bug19382_test_field_assign();
drop type bug19382_foo3 cascade;

-- Case 6: SELECT INTO field also bypasses snapshot.
create type bug19382_foo4 as (a int, b int);
create table bug19382_tbl (a int, b int);
insert into bug19382_tbl values (123, power(2, 30)::int4);
create function bug19382_test_select_into_field() returns record as $$
declare r bug19382_foo4;
begin
select a, b into r.a, r.b from bug19382_tbl;
alter type bug19382_foo4 alter attribute b type text;
return r;
end;
$$ language plpgsql;
select bug19382_test_select_into_field();
drop function bug19382_test_select_into_field();
drop table bug19382_tbl;
drop type bug19382_foo4 cascade;

Thanks!

Best regards, Andrey Borodin.

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients using the --to and --cc options:
  reply via email

  To: [email protected]
  Cc: [email protected], [email protected], [email protected], [email protected], [email protected]
  Subject: Re: BUG #19382: Server crash at __nss_database_lookup
  In-Reply-To: <[email protected]>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox