[Vobject] no TZ info in parsed dates?

[Bill Janssen]

Jeffrey Harris <jeffrey at osafoundation.org> wrote:

> Hi Florian,
> 
> > I'm having a similar problem to the original one (timezone is not
> > parsed) with event coming from evolution reading via evolution-python
> > http://www.conduit-project.org/wiki/evolution-python , namely that
> > evolution does not add the VTimezone field. What do you suggest, is it
> > "ok" to just add it and set it to the value of DTSTART;TZID in case
> > VTIMEZONE is missing?
> > 
> > Just as a side note, the time zone is not ignored with evetns having  RRULE
> 
> If you need to import that sort of VEVENT, you can of course just make
> the ics valid by adding a relevant VTIMEZONE, or you can take the
> shortcut of registering the relevant evolution tzids (see Bill Janssen's
> helpful code snippet for a guide to how you might do that,
> http://skyhouseconsulting.markmail.org/message/ektmcktzgagdvb7e).
> 
> I think Outlook and Evolution are 80% of the culprits when it comes to
> this particular issue (no VTIMEZONE).  It would be plausible to work
> around both, for Outlook by using Bill Janssen's code, for Evolution by
> just noticing TZIDs that look like /softwarestudio.org/Tzfile/.  I've
> been hesitant to try this in the past, because I don't want to boil the
> ocean, but I'll give it a shot.
> 
> As an aside, I'm thinking of adding a compatibility-mode flag to
> vobject, which will default to on and silently tolerate invalid
> iCalendar like this (if vobject can do a workaround), but when off will
> fail.

That sounds like a good way to do it.  That way sticklers can turn the
flag off, and normal folk can still get timezones for the their Outlook
calendar mail.

>  I'd like to create a Google App Engine iCalendar validator, to go
> along with the Java and .Net versions that are already out there.  Jon
> Udell's been blogging about iCalendar compatibility and I'd like to make
> it easier for applications to play nice with vobject.

Bill