On 8/30/06, Ted Leung <[EMAIL PROTECTED]> wrote:

What happens to clients like iCal, etc if we password protect
collections or if we require accounts in order to access calendar
collections?

the way the server works now, a collection can be accessed by:

* the owner of the collection, providing his username and password
* everybody else, providing a ticket

let's assume that when you share your calendar with chandler, the
"magic" read-write url that is produced is
<http://osaf.us/home/twl/calendar/?ticket=deadbeef>. let's further
assume that when julie clicks that link, she gets a web page showing
the calendar and/or providing urls for subscribing to the calendar
with dav, webcal and atom.

the question is - how do collection passwords fit into this workflow?

for the "magic url", it's easy. instead of getting the calendar page,
julie gets a page prompting her to enter the collection password. when
she submits that, she gets the calendar page.

but what about dav, webcal and atom access? do those protocols require
the collection password too? because the collection password is not a
standard authentication scheme (remember, we're authenticating with
ticket+collection password, not basic or digest authen), there's no
way for cosmo to tell ical to ask the user for the collection password
like it can do for basic authen with a 401 response.

so either the subscription urls that are provided on the calendar page
have to include both the ticket and the collection password, or we
don't use the collection password for protocol access.

seems to me that including the collection password in the
dav/webcal/atom url would defeat the purpose of the feature, but i
dunno.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Open Source Applications Foundation "Design" mailing list
http://lists.osafoundation.org/mailman/listinfo/design

Reply via email to