Submitted by: @dyemanov

Is related to CORE5600
Relate to CORE6090

Test case:

recreate table t (col1 int, col2 blob);

insert into t values (1, '1');
insert into t values (2, '2');
insert into t values (3, '3');
commit;

select * from t;

    COL1              COL2 

============ =================
1 84:0
2 84:1
3 84:2

update t set col1 = -col1;
select * from t;

    COL1              COL2 

============ =================
-1 84:0
-2 84:1
-3 84:2

-- So far so good: COL1 is negated, COL2 has the same BLOB IDs

rollback;

alter table t add col3 date;

select * from t;

    COL1              COL2        COL3 

============ ================= ===========
1 84:0 <null>
2 84:1 <null>
3 84:2 <null>

update t set col1 = -col1;
select * from t;

    COL1              COL2        COL3 

============ ================= ===========
-1 84:3 <null>
-2 84:4 <null>
-3 84:5 <null>

-- BUG: COL2 has different BLOB IDs, i.e. the whole blobs were copied rather than just their IDs

rollback;

This is a regression introduced while fixing CORE5600. v3.0.3 and later versions are affected.

Commits: 78b68a3 2e1f3f8