Possible bug in 'set constraints all deferred';

David Jack Olrik wrote:

Hi,

I think I have found a bug in the handeling of 'deferred
contraints'. I have attached a small sql script that reproduces the
error.

The first transaction succedes, but the second one failes with
'psql:bug.sql:58: ERROR: <unnamed> referential integrity violation -
key in p still referenced from c'

Isn't it supposed succed just like the first one ??

[Attachment, skipping...]

WRT the primary key constraint, this should be considered a
"triggered data change violation", and the system should
already complain at the INSERT attempt's for p.

We don't track it that precise. So the error message isn't
the right one and it happens too late. But the overall
transactional behaviour is correct.

Jan

--

#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#================================================== JanWieck@Yahoo.com #