Re: [Evolution-hackers] ESourceExtension and backends... "claimed" generic property ?

[Milan Crha <mcrha redhat com>]

On Tue, 2013-02-05 at 12:54 -0500, Matthew Barnes wrote:
> Sure.  It makes sense for backends to continue advertising their own
> capabilities -- static and non-static like "browsable".
> 
> I'm just suggesting a way to peek at a backend's known capabilities
> ahead of time without having to instantiate it.

	Hi,
I do not know, maybe I misunderstand your suggestion, but if there will
be two different ways of getting backend capabilities, and each will
return different set of them, then it'll be confusing. We are still
talking about capabilities, thus whatever way the result should be
consistent. And I do not see a point in "get capabilities without
creating actual instance of a backend in the factory", creating backend
instance and connecting to its destination (aka opening it) are two
different tasks. 

>From what I understand then Tristan will always work with an instance,
with an EClient descendant, asking "do you support xyz?" and if so, then
he'll (pre)configure the xyz based on his needs. Right? Thus there will
be nothing like "does generic backend support xyz?", if no, skip the
whole backend/source creation with nothing given to a user.
	Bye,
	Milan