Isn't the XfceColorButton typdef and type macro wrong

[Brian J. Tarricone]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Benedikt Meurer wrote:
> Brian J. Tarricone wrote:
> 
>>Jeff Franks wrote:
>>
>>
>>>>Brain,
>>>>
>>>>Aren't the XfceColorButton typdef and 'get_type' macro wrong. Shouldn't
>>>>
>>>>#define XFCE_TYPE_COLOR_BUTTON  (gtk_button_get_type())
>>>>typedef GtkButton XfceColorButton;
>>>>typedef GtkButtonClass XfceColorButtonClass;
>>>>
>>>>actually be:
>>>>
>>>>#define XFCE_TYPE_COLOR_BUTTON  (gtk_color_button_get_type())
>>>>typedef GtkColorButton XfceColorButton;
>>>>typedef GtkColorButtonClass XfceColorButtonClass;
>>
>>
>>Nope.  The whole point of XfceColorButton was to work with gtk 2.2,
>>which doesn't have GtkColorButton.  I also didn't want to force
>>recompiles if people change their gtk version from underneath
>>libxfcegui4.  It's kinda ugly, yes, because it makes XfceColorButton not
>>a real widget.  At any rate, it's deprecated and you don't need to bind it.
> 
> 
> I'd suggest to define XFCE_TYPE_COLOR_BUTTON to
> xfce_color_button_get_type () and implement xfce_color_button_get_type
> () like this:
> 
> GType
> xfce_color_button_get_type (void)
> {
>   return gtk_color_button_get_type ();
> }
> 
> This way you maintain full source and binary compatibility. And while
> you are at it, you can ditch the Gtk2.2 compat code from the
> implementation and turn all xfce_color_button_* () calls into proxies.

Why bother?  No one should be using XfceColorButton anymore.  Seems like
a waste of time to me, but if someone else wants to do it, feel free.

	-b

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)

iD8DBQFDGZWJ6XyW6VEeAnsRAr1eAKCk8+xR9B/xGw9nRdYuczpOXmYpWgCgjIbZ
Swx8krwBYIAyg09bEoZUi/E=
=v0y7
-----END PGP SIGNATURE-----