Its possible it's not really a concurrency bug and just a
bug that manifested itself because unchanged null context_ids where falling through (I really did not investigate it once I realized it should not have been
running at all in this case). I don't think thats the case
though since running a single thread should have the same
codepath and I didn't get the error there.
Anyway, here is the postgresql error log.....
2003-11-15 15:53:10 LOG: query:
select acs__add_user(
'3516',
'user',
now(),
null,
'127.0.0.1',
'9',
'test-user10551',
'test-10551@example.com',
NULL,
'user10551',
'test',
'F8DFBF53079A6306A7EBE11A36534549E09A70',
'A315004DDE1F3CA754561EFD287B5E5F06B9E4',
NULL,
't',
'approved'
);
2003-11-15 15:53:10 ERROR: Cannot insert a duplicate key into unique index acs_object_context_index_pk
2003-11-15 15:53:10 WARNING: Error occurred while executing PL/pgSQL function acs_objects_context_id_up_tr
2003-11-15 15:53:10 WARNING: line 28 at SQL statement
