Persistent chat rooms

[Sadrul Habib Chowdhury]

* Sean Egan had this to say on [18 Sep 2007, 11:39:21 -0700]:
> On 9/17/07, Sadrul Habib Chowdhury <imadil at gmail.com> wrote:
> > > FYI: the two branches have been merged, and conversation hiding is now
> > > in im.pidgin.pidgin.
> 
> Some questions I'm too lazy to check the code to answer:
> 
> - Does this need to be a preference? Is there a strong reason for not
> wanting to maintain the IM history that isn't solved by Ctrl-L? If it
> is necessary, can we think of a better wording than "Close IMs
> immediately when the tab is closed"? It makes it sound like there's
> some arbitrary delay between hitting the close button and seeing it
> closed. I'd sugggest "Clear IM history when closing window" or
> something like that
> 

That sounds better, I think. (I am really bad with strings)

> - What's the timeout for eventually clearing the backlog?
> 

10 minutes. Non-configurable.

> - When does the prpl's conv_close callback get called? When the window
> is closed, or when the backlog is cleared? I'd prefer the former
> 

It's the latter, now. Closing the tab/window unsets the ui-data,
triggering the destroy-callback for Pidgin, but it doesn't actually
destroy the conversation itself. (so in gmail, the other end won't see
a 'Ihateyou has left the conversation.') I think I like this more. Is
there a reaons you prefer the former?

> - Do we need an auto-join and a persistent flag? I'd be just as happy
> with only one or even neither.
> 

I don't have any opinion on this. I added a new flag mostly so someone
doesn't get surprised by the new (and unexpected) behaviour.

> - Related: does persistant imply auto-join? It seems that a persistant
> chat should join on start, but not show the conversation window for
> it.

Or may be we can look at the 'hide conversation' preference before
deciding whether to show it or not.

> -s.

-s2