Andres mentioned to me that I might want to look at
HeapTupleHeaderGetOidOld() as there's a comment and some code there
that would lead you to believe we can still have tuples with oids. If
that were true, the code I recently added for getting 'tp' in
slotselectivelydeformheaptuple() is wrong.

#define HEAPHASOIDOLD 0x0008 / has an object-id field /

On 11.22, I tried:

create extension pageinspect;
create table t1 (a int) with oids;
insert into t1 select generate_Series(1,1000);
select count(*) from heappageitems(getrawpage('public.t1',0))

where t_infomask & 8 <> 0;

count
-------
185

alter table t1 set without oids;

select count(*) from heappageitems(getrawpage('public.t1',0))

where t_infomask & 8 <> 0;

count
-------
0

/*
 * If we dropped the OID column, must adjust pg_class.relhasoids and tell
 * Phase 3 to physically get rid of the column.  We formerly left the
 * column in place physically, but this caused subtle problems.  See
 * http://archives.postgresql.org/pgsql-hackers/2009-02/msg00363.php
 */
if (attnum == ObjectIdAttributeNumber)

And also, if I try to pg_upgrade before SET WITHOUT OIDS, I get:

$ pg_upgrade -d pgdata11 -D pgdata -b ~/pg11/bin -B ~/pg/bin
Performing Consistency Checks
-----------------------------
Checking cluster versions ok
Checking database user is the install user ok
Checking database connection settings ok
Checking for prepared transactions ok
Checking for system-defined composite types in user tables ok
Checking for reg* data types in user tables ok
Checking for contrib/isn with bigint-passing mismatch ok
Checking for user-defined encoding conversions ok
Checking for user-defined postfix operators ok
Checking for incompatible polymorphic functions ok
Checking for tables WITH OIDS fatal

So, I'm not following how we could get a tuple with OIDs in versions after 11.

Should we get rid of HEAPHASOIDOLD and the code that relates to it?