Re: daily recurrence

[torkel hpc2n umu se (Björn Torkelsson)]

Eskil Heyn Olsen <deity@dbc.dk> writes:

> That code was added by me. The problem was, people reported that
> end-on-date recurrences ended a day too early.
> 
> As I read the rfc for vcal/ical, end-on-date recurrence that eg. list
> eod=19991224 means the event must occur also on this day (correct me if
> I'm way off). All the examples on daily/monthly/etc recurrences I found
> had this described in the explanation.

That's the logical assumption, and I really think you are correct.

> The problem afaic determine was the eod was stored eg. to 19991224T000000,
> and if the event was timed to T140000, then since 140000 !< 000000, the
> event at 2pm was not scheduled on the eod at 24/12. Therefore by adding
> one day internally, the event was set to end on 19991225T000000, thereby
> effectively making the event recur until midnight (yes, arguably this
> should maybe by 19991224T235959)

Ah! Now I understand why the code is there.

> > For events that are repeated daily to a specific enddate, I get the one
> > day later then expected.
> 
> Is that with the current gnome-pim in cvs ? Because by removing the
> +86400's, I get the opposite behaviour, that the event ends a day too
> early.

Well, the problem is when syncing with a palm. I get the "famous" off by
on day error. When syncing, the end date is one day later in the pilot
than in gnomecal. Besides, in user-cal.vcf the enddate (of RRULE) is on
day later than I specified in gnomecal. I.e. if I specify an event
starting on 19991224T080000 and ending 16.30 same day, with a recurrece
every day, with an enddate of 19991226, the RRULE in user-cal.vcf says D1
19991227T000000 and my pilot thinks that I have an event on the 27th.

And yes, this is with the current gnome-pim in cvs.

Hm, try adding an event, recurring daily and with and enddate. After
OK:ing that event, edit the event. The enddate for the recurrence is now
one day later than the event i type in. Strange! Have I been fideling to
much with the code so I broke something in my local copy?

/torkel - looking further into the code... someday...