Re: Cancel kinds, implicit and explicit

[Sergio Villar Senin <svillar igalia com>]

Philip Van Hoof escribiu:
> On Tue, 2008-05-20 at 12:43 +0200, Sergio Villar Senin wrote:
>> Philip Van Hoof escribiu:
>>> It would be better if this were two kinds of cancels and that certain
>>> operations, like tny_folder_transfer_msgs_* don't subscribe as being a
>>> automatic-cancellable, but only as a explicitly-cancellable.
>> What it'd be nice as well is the possibility to add a parameter that
>> could allow the developer to specify if they want the operation to be
>> cancelable or not.
>>
> 
> For that I think I rather want to migrate to GIO's asynchronous API,
> with its GCancellable type. 
> 
> We could then provide a GCancellable that is not doing anything when
> doing g_cancellable_cancel.
> 
> I think this kind of API change is better than just-before-1.0 adding an
> extra bool or flag to all _async APIs. Because a Tny 2.0 will most
> likely migrate to GIO's asynchronous model.
> 
> I don't think for the sake of Modest we want to do this with Tny 1.0
> already. Right?

Right, I was talking about 2.0.

Br