Re: virtual signal system ?

[Bo Lorentsen <bl lue dk>]

Murray Cumming wrote:

> I'm not aware of significant performance problems on regular desktop
> PCs. For embedded hardware, this API is optional at build time.
Hmm, I found some discussion on the net, of the cost of gtkmm (i need to 
find the link, sorry) and the overhead of the library. And I have been 
using the lib for some minor projects and was a bit worried about the 
cost of "empty signal function" if some of the projects scaled up.

Just think about projects like k3d, the number of empty signal functions 
must me enormous !

> Anyway, we would need a very good reason to do a
> API-break-plus-parallel-install.
Yes, I know that, and for now I just want to here if this model was 
worth looking into, as this removed the forced use of the signal vtables 
(on_* functions).

/BL