D24494: Add convenience for defaults/dirty states to KCoreConfigSkeleton

[Kevin Ottens]

ervin added inline comments.

INLINE COMMENTS

> davidedmundson wrote in kcoreconfigskeleton.cpp:140
> Do we need to make this
> 
>   if (d->mIsDefaultImpl){
>      return d->mIsDefaultImpl();
>   }
>   return false;
> 
> and initialize mIsDefaultImpl to nullptr
> 
> so that it doesn't crash if someone subclasses KConfigSkeletonItem directly and doesn't implement this?

Initializing to nullptr wouldn't help, you would get the same behavior than with the default constructor.

It means it doesn't crash but raises a std::bad_function_call exception, which I think is fine... it's the next best thing to a pure virtual, but it's caught at runtime. I don't think we can do better than this in the current situation.

> davidedmundson wrote in kcoreconfigskeleton.h:218
> Is this an ABI break?
> 
> :/

*gasp* you're right, how stupid of me... we can't merge that...

REPOSITORY
  R237 KConfig

REVISION DETAIL
  https://phabricator.kde.org/D24494

To: ervin, #plasma, #frameworks, dfaure, mart
Cc: kossebau, davidedmundson, kde-frameworks-devel, LeGast00n, GB_2, michaelh, ngraham, bruns
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20191008/55ba3677/attachment.html>