Re: [SyncEvolution] Store evolution items as "vCalendar" instead of "iCalendar"
Patrick wrote:
On Wed, 2009-12-09 at 00:46 +0000, Chen, Congwu wrote:
> On Tue, 08 Dec 2009, Ohly, Patrick wrote:
> > On Tue, 2009-12-08 at 14:57 +0000, Chen, Congwu wrote:
> > > On Tue, 08 Dec 2009, Ohly, Patrick wrote:
> > > > On Tue, 2009-12-08 at 09:57 +0000, Chen Congwu wrote:
> > > >
> > > > The Evolution backend really can't import vCalendar 1.0. libical
will
> > > > fail to parse it, because encoding rules are different. There might
be
> > > > specific examples that happen to work, but this is not something
that
we
> > > > should count on.
> > > Er, the testing sample items for vcalendar 1.0 all have an 'iCalendar
2.0'
type
> > > (config.insertItem, etc..). They are feed directly to the evolution
backend.
Does
> > > that mean all these sample items are guranteed to be correctly parsed
by libical?
> >
> > Aren't all of these sample items really iCalendar 2.0, taken from
> > test/testcases/ical20.ics?
> That's true only for testItems
It's also true for all other ical20 tests. Or asked the other way
around, were do we currently have vCalendar 1.0 test data? Are you
perhaps talking about code and tests that you have added, but not
published yet?
That's it. We had built-in tests for both vcalendar and icalendar 20 test
data, just without a registered test case for that. This leads me to think
adding such a test for this.
Now I understand the problem: the evolution backend has problems correctly
parse some vcalendar 1.0 items, better to rely on the icalendar test items.
Best Regards,
Congwu