Re: [HC Evolution] Some ideas

[NotZed <notzed helixcode com>]

> > The send routines work on an already-connected transport.
> >
> > ie, the app would do something like:
> >
> > smtp = camel_session_get_transport_for_protocol (session, "SMTP", ex);
> > camel_service_connect_with_url (CAMEL_SERVICE (smtp),
> >                             "smtp://trna.helixcode.com", ex);
> > for (msg = g_list_first (queued_messages); msg; msg = msg->next)
> >     camel_transport_send (smtp, CAMEL_MEDIUM (msg->data), ex);
> > camel_service_disconnect (CAMEL_SERVICE (smtp), ex)
> >
> > (except checking for errors, of course :-)
> Here I have to desagree if we do something like that we will have a prety lock application whenever we try 
> to send mail. 
> I was thinking more in doing something using IO (don't remember the name of the funtion) wach using gnet to 
> make all the socket conections and the receiving and sending stuff.

thats really an implementaiton issue, the api itself above will
not/need to contain such information.

> What you think about that?

I think the decision has finally been made to use threads - async
api on async api all routed through the main loop is getting too
painful.

But yes, that above api is a synchronous one, although it wouldn't
be very hard to make it asynchronous (infact, given what it does,
it could probably be made asynchronous transparently).

 Michael