Re: [Evolution] CalDAV error msg

[Milan Crha <mcrha redhat com>]

On Sun, 2010-02-21 at 08:08 -0500, Philippe LeCavalier wrote:
> On Sat, 2010-02-20 at 20:07 -0500, Adam Tauno Williams wrote:
> > < Content-Type: calendar/ics <-- you have to have a Content-Type
> > header.
> Mine is reported as: Content-Type: application/octet-stream

        Hi,
yes, yours is application/octet-stream, which is shown in your log in
the previous mail, two lines below suspicious ETag:-0. Adam probably
overlooked it.

> So obviously any software that adheres to strict standards won't like
> that. I thought Mozilla was in that boat too but I guess they're not
> as much as Evo.

Evolution uses different approach to fetch items from Sunbird. Where
Sunbird asks for data by the REPORT request, Evolution does that too,
but sometimes asks for the event itself, by direct URL, which is a legal
thing, and server is supposed to return exactly the same data as were
written with PUT, with a text/calendar Content-Type, when asking for a
calendar entry.

So they should fix two things:
a) ETag generation (I'm wondering what they are returning on initial
fetch with a REPORT, if there is a correct ETag, then they have only the
direct item fetch broken)
b) Content-type returned with a direct item fetch (and maybe returned
data itself, it depends).

You can check stored etags in a calendar cache, in
~/.evolution/cache/calendar/<caldav server url>/
are downloaded items, and when you'll search for
X-EVOLUTION-CALDAV-ETAG, then you'll see whether they all are "-0" or
they differ.

        Bye,
        Milan