kdebase/kcontrol

[Simon Hausmann]

On Thu, Sep 18, 2003 at 12:37:46PM +0200, Matthias Kretz wrote:
> Thanks, for fixing these KCMs. But still the problem remains that there might  
> be other (third party) KCMs out there with the same problem, that's why I 
> need to know what's breaking your KControl.
> 
> I did not break src compat, I kind of broke behaviour compat (and I tried to 
> explain that before). Since I fixed both KCMultiDialog and ProxyWidget (and 
> Lubos fixed KCDialog) this should be no issue.
> 
> I'll try to explain once more what's causing this:
> KControl creates the KCM and only after that it can connect to the 
> changed(bool) signal. If a KCM is broken and calls setChanged(true) in the 
> ctor the changed(true) signal will not be emitted again until 
> setChanged(false) and setChanged(true) is called. To fix this all classes 
> using KCModule need to check for KCModule::changed() after the KCM ctor was 
> called and enable/disable buttons accordingly.

A hack to work around this would be to connect the signal to a
parent widget's slot in the kcmodule constructor, if the slot exists
(one kcontrol would provide, or whatever the container is) .

If one doesn't want to hardcode the slot name it could be provided
as Qt property by the parent.

Simon