Stewart Smith wrote:
On Fri, Jan 09, 2009 at 03:02:00PM -0500, Jay Pipes wrote:
For those interested, I have modified Drizzle to behave
accordingly. Code is in lp:~jaypipes/drizzle/temporal for those
interested in looking at the code. r754 contains the code.
I also added a new test case t/date_format.test to verify the error
and NULL behaviour.
May also want to look at my 'tests' branch that contains a bunch of
patches "improving" dates by setting NO_ZERO_DATES by default. catches a
few more things and weirdness.... but really it's just pain.
Yep. I have removed the NO_ZERO_DATES and fuzzy_date stuff from the
temporal calculations, with errors now being reported for *any* temporal
construct that isn't valid according to its column type.
There is a place in the optimizer where NO_ZERO_DATES is used in some
interesting ways, but I think the way I am storing temporal values will
alleviate the need for this.
perhaps resync when you think you're done and we can poke holes in each
others understanding of the basic concept of time, drink heavily and
then wonder where some period of time went... implementing a special
case in teh server as to not acknowledge anything that happened in the
time we don't have.
Any excuse for a drink is cool by me ;) I should be finished with the
temporal stuff this week and definitely will merge in your test patches
where I can. We can then discuss anything that didn't overlap.
Cheers!
Jay
_______________________________________________
Mailing list: https://launchpad.net/~drizzle-discuss
Post to : [email protected]
Unsubscribe : https://launchpad.net/~drizzle-discuss
More help : https://help.launchpad.net/ListHelp
