On 9 Jul, 2008, at 17:18, Jeffrey Harris wrote:
Hi Folks,
Re: the first occurrence not being included in Chandler, Katie said:
I assume this is known and mostly harmless, but thought I'd bring
it up.
Yeah, it kind of sucks that dateutil decided to implement the RFC's
definition of rrule in almost every way except this.
When I originally wrote vobject, this bothered me so much I wrote
code to always create an RDATE to match DTSTART, so the rules would
match. But that turns out not to really solve the problem, and it
causes a bunch of hassles in Chandler if you later decide to change
what day the first occurrence is on.
I think fixing this would basically require some major revisions to
dateutil. I periodically think of doing that, because I'm convinced
we could get a fair amount of performance improvement from updating
it (and we could make it easier to read/hack/update, to boot), but
it's never gotten near the top of my priority list.
Another approach would be to fix Chandler's rrule expansion (i.e. to
include the master's start if dateutil doesn't, and it's not an
exdate). That wouldn't have any benefit for dateutil, of course :o.
--Grant
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Open Source Applications Foundation "chandler-dev" mailing list
http://lists.osafoundation.org/mailman/listinfo/chandler-dev
