Re: [sigc] Interesting minor libsigc++ 2 problems

[Martin Schulze <martin-ml hippogriff de>]

Am 2004.02.14 18:11 schrieb(en) Murray Cumming:
> These are not big problems, but they are interesting when porting to
> libsigc++ 2:
>
> 1. In libsigc++ 1.2 you can do  button.signal_clicked().connect 
> (*this,
> &Gtk::Widget::hide)
> but in libsigc++ 2 you need to write a separate intermediate method  
> to
> handle the signal and then call hide().

Hm, I'm afraid you have to do button.signal_clicked().connect(
 *(Gtk::Widget*)this, &Gtk::Widget::hide);

I have no idea why Karl originally added this restriction to
libsigc++ 2. It seems stupid. Shall we try to lift it again?
(I think I had a look at boost at some time and they have
this restriction but I don't have any problems with this...)

> 2. In libsigc++ 1.2 you can override a method like so:
>   void do_something(const Slot0<void>& slot);
>   void do_something(const Slot1<void, int>& slot);
> but in libsigc++ 2 that is ambiguous when you try to call
> do_something().

We could get around this, too, by reintroducing slot# templates
besides slot (like signal# and signal). It seems that this
would also be worth it, given that you already ran into troubles ...

> Sorry for not giving actual compiler errors - I'm rushed.

No problem :)

Regards,

 Martin