On Tue, 12 Jul 2011 07:41:23 you wrote:
> Neil, sure, that's my whole point :-) My point is that the following
> change gives a lot of code to refactor:
>
> TeaCup::TeaCup() {
> - _full_prop = new RWProperty<bool> ();
> + _full_prop = new ROProperty<bool> ();
> }
Aside: you don't use new with properties.
>
> You need to refactor all Set() calls and make sure you manually emit
> signals where appropriate. I was just arguing for a solution where the
> above diff would be all there was to it.
There are no Set calls for a ROProperty. If you are changing a RWProperty (or
Property) into a ROProperty, then yes you should refactor as the underlying
semantics have changed, and the usage sites should be looked at and evaluated.
On Tue, 12 Jul 2011 07:41:23 you wrote:
> Neil, sure, that's my whole point :-) My point is that the following
> change gives a lot of code to refactor:
>
> TeaCup::TeaCup() {
> - _full_prop = new RWProperty<bool> ();
> + _full_prop = new ROProperty<bool> ();
> }
Aside: you don't use new with properties.
>
> You need to refactor all Set() calls and make sure you manually emit
> signals where appropriate. I was just arguing for a solution where the
> above diff would be all there was to it.
There are no Set calls for a ROProperty. If you are changing a RWProperty (or
Property) into a ROProperty, then yes you should refactor as the underlying
semantics have changed, and the usage sites should be looked at and evaluated.
> Anyway. Maybe I am just bikeshedding this :-)
Perhaps :-)