pidgin at unreliablesource.net
Fri Jun 22 13:54:05 EDT 2007
On Fri, Jun 22, 2007 at 10:24:52AM -0700, Sean Egan wrote:
> On 6/22/07, Etan Reisner <pidgin at unreliablesource.net> wrote:
> >Right, and I'm trying to understand for what purpose the ui needs to know
> >that information. It would seem to me that neither filling in fields
> >correctly or automatically connecting would need that (though auto-filling
> >might be made easier that way).
> You don't think the UI should know if clicking "Register" succeeded or not?
I'm unsure why the UI needs to know that explicitely, as opposed to the
current purple_notify_info method of just letting the user know. The two
cases that I believe I've seen being claimed to be solved here are both
solvable in other ways, as indicated by the fact that pidgin does the
former and my emailed code should do the latter.
> >Wouldn't it be just as effective to add a
> >purple_account_registered(PurpleAccount *account, gboolean success)
> >function to account.h and a corresponding ui_op? Which would be a minor
> >bump, right? Also adding a corresping account-registered signal, so that
> >plugins and other places can respond to it.
> All three of those options were discussed. Andy's original plan was to
> use a ui_op (or maybe it was a prpl_op, but a ui_op would be better).
> I was originally hesitant to use up one of those precious
> purple_reserved padding pointers, but upon reflection, I think that's
> a better idea now.
Ok, I definitely agree we shouldn't use up our padding willy-nilly, but
this seems like a case that is tailor made for this sort of thing, and we
can easily just go with the signal for now and only do the ui_op later if
we find we need it (and are bumping to 3.0.0 so we can add more padding).
More information about the Devel