pidgin: ae37fe6b: Added "chat-join-failed" signal, emitted...

[John Bailey]

Evan Schoenberg wrote:
>> This addition of API and a signal forces us to version 2.5.0. 
> 
> Are we actively avoiding minor version bumps?

Not specifically, no.

> Also, just to be clear: Is the reason that this forces to 2.5.0 that
> 2.4.0 libpurple must be able to work perfectly with plugins compiled for
> a theoretical 2.4.2... and those plugins might depend upon this new signal?

The addition of *anything* to the public API requires a minor version bump
according to the rules.

> In any case, this is needed because previously there was no way for a UI
> to know that the attempt to join a chat failed.  A notification is
> posted to tell the user, but this clearly is not the needed information
> as it is a string.

Ok, I imagine resiak's telepathy-haze will appreciate this as well.

> Pidgin and Finch may not care about this fact, as they may not create a
> GtkChat or GntChat (or whatever the correct struct name is) until the
> conversation-created signal is posted.  Other UIs which are less tied to
> libpurple's internal organization - Adium, for example - may create the
> equivalent object (AIChat for Adium) when the request to join a chat is
> made rather than once the chat is successfully joined.  Undoing this -
> releasing the memory, removing it from tracking dictionaries, etc. -
> should occur if the chat isn't joined because of an error.  (An easy
> error to generate is the 409 Conflict error from trying to join an XMPP
> conference with a handle which is already in use).

What specifically is the problem with waiting to create UI elements until the
chat creation is completed successfully?

John

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: OpenPGP digital signature
URL: <http://pidgin.im/pipermail/devel/attachments/20080428/bd17ec40/attachment.sig>