[g-a-devel]Re: Woah ... atkobjectfactory ...

["Padraig O'Briain" <Padraig Obriain sun com>]

> Subject: Woah ... atkobjectfactory ...
> To: "Padraig O'Briain" <Padraig Obriain sun com>, Bill Haneman 
<bill haneman sun com>
> Cc: accessibility mailing list <gnome-accessibility-devel gnome org>
> Content-Transfer-Encoding: 7bit
> Mime-Version: 1.0
> 
> Hello,
> 
> 	I'm just starting to implement my first custom AtkObject, and I must
> say I'm slightly concerned;
> 
> 	Surely it can't be the case that every AtkObject has it's own
> AtkObjectFactory ? surely it must be possible (for efficiency reasons)
> to have a single AtkObjectFactory that can manufacture a whole range of
> different AtkObject types ?

It would be possible to use a single AtkObjectFactory to manufacture different 
AtkObject types but for the fact that I failed to add a GObject parameter to 
atk_object_factory_get_accessible_type().

Until that is sorted out, we are stuck if we want to derive the AtkObject for a 
custom widget from the AtkObject of the GTK widget from which the custom widget 
is derived.

Padraig

> 
> 	I seem to be seeing just a gigantic swathe of cut and paste coding here
> - I'm really hoping that this is just me being dim. Surely this is not
> one of those alleged 'programming patterns' [ code phrase for unthinking
> cut and paste ] - that I see ?
> 
> 	Looking at the diffs between some of the 40 factories in gail I see
> nothing but renamed structures - all of the same size - without any
> complex inheritance, new data members - anything substantially different
> between them all. Oh, and the object files are ~150Kb each ~6Mb of disk
> space accounting for perhaps 20% of the stripped size of libgail.
> 
> 	Could this not have been a single factory instance ? resulting in 1/2
> as many files in gail/ more icache coherency, smaller in-core size, more
> speed etc.
> 
> 	Please, tell me I'm wrong,
> 
> 	Regards,
> 
> 		Michael.
> 
> -- 
>  mmeeks gnu org  <><, Pseudo Engineer, itinerant idiot
>