Kris Jurka wrote:
>
>
> On Mon, 1 Aug 2005, Oliver Jowett wrote:
>
>> I've committed my Timestamp/Date/Time changes to CVS HEAD.
>
>
> The regression tests fail on 7.2 and 7.3 with the below stacktrace. I
> understand that we may not be able to achieve 100% correctness with
> 7.2/7.3, but aren't there at least certain cases that work?
There certainly are cases that work, but that test isn't one of them.
Per my commit comments, this test is the case of setTimestamp() on a
'time without timezone' column. 7.4 was the first version to accept (and
discard) a timezone in this case. We don't know ahead of time that the
target parameter is a 'time without timezone' (and can't for pre-7.4
even if we wanted to, since there's no V3 protocol support), and if we
omitted the timezone information then using setTimestamp() against
timestamp fields would produce incorrect results.
If you're suggesting we skip that particular case so we can run the rest
of the testcase on 7.2/7.3, it's a good idea but more work than I had
time for -- feel free to update the test as necessary.
-O