Hi Dirk,
Dirk Reiners wrote:
> Hi Carsten,
>
> Carsten Neumann wrote:
>> Agreed. The only case I can see that right now is more efficient than
>> with the changed interface is, if the user calls the MaxType
>> get/setValue methods from the GeoVectorProperty level (i.e. not
>> TypedGeoVectorProperty), because it avoids the second conversion done in
>> the template member functions.
>> We can get rid of that by making some changes to the GeoConvert structs
>> (to some degree these are needed anyway, see below), using partial
>> template specialization (or do we need to support some crummy compiler
>> that still can not handle that ? [1]
>
> I don't think so. Our baseline is VS2003 and gcc 3.4, so those should be ok.
ok, good.
> ). I've attached a draft of what I
>> mean.
>
> I like the idea of specializing for identical types, that is neat.
I'm not sure if this turns into a case of premature optimization,
because the compiler would get it right anyway, but its fun to play
around with it a little :)
> Looks
> good to me, do you want to get it to the point that it compiles? And
> maybe add some unittests for it? ;)
I'm on it, hope to have results by the weekend.
Cheers,
Carsten
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Opensg-core mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-core
