Re: g_value_ types ...

On 3 May 2001, Owen Taylor wrote:

> Can we back up here and get some context here about what you
> want these types for?
> I think all of these could be done with G_TYPE_BOXED - using
> a new fundemental type is just a small efficiency win and
> for a lot of things will be lost in the noise.
> [
>  Actually, I'm quite opposed to the idea of reserving type IDs, it's
>  setting us up for pain, and if people really register their ideas, we
>  could could easily run out of space in the registry; I think the right
>  way to handle this is to simply have a g_type_register_fundemental().

to put this into perspective, we have had a dynamically working
g_type_register_fundemental() right from the start.

registering an ID just provides the additional benefit of having a
constant that you can switch (G_TYPE_FUNDAMENTAL (type)) on with
case statements which require constant expressions.

>  Yes, it's a tiny bit slower to use a global variable to store
>  your fundemental type than a constant. I don't think
>  this is worth worrying about
> ]
> Regards,
>                                         Owen


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]