[PATCH] Add serv_got_private_alias

[Felipe Contreras]

On Sun, Feb 24, 2008 at 11:57 PM, John Bailey <rekkanoryo at rekkanoryo.org> wrote:
> On Sun, Feb 24, 2008 at 09:24:47PM +0200, Felipe Contreras wrote:
>  > >  So here is an example patch to use the new serv_got_private_alias.
>  > >  Note that prpl->alias_buddy should also be renamed to
>  > >  set_private_alias or something like that.
>  > >
>  > >  Best regards.
>  >
>  > So?
>  >
>  > Any chance of this getting applied?
>
>  Funny that you should mention this; I was looking at it just a while
>  ago.  Currently the patch is failing to apply due to conflicts in
>  oscar.c and server.c because the code has changed too much since 2.3.1.
>  I was looking at merging it by hand, but I'm not comfortable touching
>  oscar at all.  Any chance I could convince you to resubmit the patch
>  against current monotone?  I have no objections to the patch, for the
>  record.

I only sent the changes to other prpls because I was asked for
examples of this new API. I guess the only important part are the
changes in the core, the rest can be ignored, and actually I recommend
you to do that since I'm not familiar with that code.

>  Out of curiosity, where would this alias stuff fit if we were to
>  completely kill off server.c for 3.0.0?  Some of it feels like a good
>  fit for blist.[ch], but it might be confusing if we split some functions
>  up because they seem like a better fit in, say, prpl.[ch] and their
>  companion functions are in other files because they seem to fit best
>  there.

I don't know how much are you planning to change things for the mythical 3.0.0.

If you kill server.c then part should go to blist (got_*_alias) and
part should go to prpl (set_*_buddy_alias).

If you also use GObjects then you will probably be passing around
buddy objects instead of char *. Then it might become clear that all
the buddy prpl functions don't make sense since each protocol can have
specific implementations for the buddy stuff:

purple_buddy_set_property (buddy, "private_alias", value);
or
g_object_set (buddy, "private_alias", value, NULL);

In this case "buddy" can be a PurpleBuddyMsn, or PurpleBuddyYahoo, etc.

Then "got_private_alias" becomes the GObject signal "private_alias_changed".

Of course, that's assuming the buddy stops being a TreeModel equivalent.

Best regards.

-- 
Felipe Contreras