On 02/04/13 10:32, Dave Cridland wrote:
On 31 Mar 2013 10:33, "Kevin Smith" <[email protected] <mailto:[email protected]>> wrote: > > On Sat, Mar 30, 2013 at 1:08 PM, Yann Leboulanger <[email protected] <mailto:[email protected]>> wrote: > > While starting to implement XEP-0191, I realized that there is a regression > > in the feature Gajim offers if I don't use privacy lists: The ability to > > block a group by its name. > > Yann - I've been thinking about this, and I don't entirely understand > what the use case is here - in what situation would you want to add a > load of people to your roster, put them all in the same group, but > then block them? A work group, perhaps? So you could appear offline to your colleagues? I think the (possibly) correct method is to have groups be representable as jids.
I don't really see what this buys over PSA's suggestion of Gajim implementing "block this whole group" with "foreach user in group: block user", except more complexity which XEP-0191 (as I understand it) was explicitly designed to reduce.
If you get into blocking roster groups, you have to consider what happens if you unblock a JID which is inside a blocked group; how you represent to the user that moving a user out of a group they've blocked might (?) unblock them; etc.
-- Will
