[kronolith] Invitation e-mail wrong?

[Jan Schneider]

Zitat von Jens-Uwe Mozdzen <jmozdzen at nde.ag>:

>> What I failed to check was the content of the invitation email - it  
>> was inconsistent as well... while HTML correctly showed 18 Jan  
>> 00:00 to Jan 19 23:59, the ICS parts were again a day longer: Jan  
>> 18 to 20. After accepting my own invitation, again the stored event  
>> was modified to the new date range, making it one day longer than  
>> originally planned.
>>
>> My conclusion: The invitation emails are in fact inconsistent and,  
>> concerning the ICS part, wrong.
>>
>
> Average life time of conclusions: Today shorter than ever.
>
> I just received an invitation from Google for an all-day, single-day  
> event. While I won't take them for the master of the Universe, I  
> expect them to send standard-conforming invitation emails - and that  
> specific one again had the end date as "start date + 1".
>
> Both invitations (Google/Horde) are of type text/calendar.
>
> For text/calendar, the following is defined in RFC 2445, 4.6.1:
>
> The "DTEND" property for a "VEVENT" calendar component specifies the  
> non-inclusive end of the event.
>
> I read that so that for a single-day all-day event, DTEND is to be  
> DTSTART+1... making the Horde invitation correct.

Indeed.

> But doesn't that mean that Horde's handling of the invitation is  
> incorrect? Shouldn't it be so in the case of all-day events, that  
> the internally stored end date is DTEND-1? Because kronolith seems  
> to interpret that value as inclusive.

Yes.
-- 
Jan Schneider
The Horde Project
http://www.horde.org/