"Hide empty groups"-checkbutton in conference-server-browser

[bb]

I was thinking this would be useful for people browsing for a conference room on servers like conference.jabber.org. Or, making the headers of the table clickable to sort the table. That way, one can sort by participants and have the same result, for example.

[asterix]

sortable columns sounds a good solution.

[junglecow]

I would go so far as to invert the option as "Show empty groups" (unchecked by default). All those empty groups bloat the list and are confusing. Sorting by usercount is not that intuitive for our average grandmother.

[anonymous]

Replying to junglecow:

I would go so far as to invert the option as "Show empty groups" (unchecked by default). All those empty groups bloat the list and are confusing. Sorting by usercount is not that intuitive for our average grandmother.

/agree although it would still be nice to be able to sort by participants, especially if the list is big (future!! ;P).

[junglecow]

Replying to anonymous:

although it would still be nice to be able to sort by participants, especially if the list is big (future!! ;P).

Yes, agreed ;-)

However, remember that in many cases the # of participants has to be requested separately for each room, and in a big list that can take a couple of years. (No idea why it's so slow, but it is.)

[asterix]

junglecow is right. we currently show number of participants of room we see in the window cause it's very long to get them from server. For exemple in jabber.org it must take several minutes. So I see no way to do this ticket.

[junglecow]

Well, actually, both jabber.org and jabber.ru add (numberofusers) to the room name when we get the list, so we could parse that.

There are two ways to detect:

if room_count > 10 and every_room_name_matches("\([0-9]+\)$"):
        parse_usernumbers_from_names()

Or alternatively, when we get the first detail record:

if new.name == old.name + " (%d)" % new.numberofusers:
        parse_usernumbers_from_names()

amessage.* doesn't do it though, so we need to handle that case in a good way as well.

[asterix]

I'm not sure it's a good thing thing to implement some server specific features. There should be a note in XEP instead. I'll try to talk about that to stpeter or in JIG ML

[steve-e]

See [8310] and [8312]