[Differential] [Commented On] D4604: KConfig: stop exporting and installing KConfigBackend.

[Matthew Dawson]

mdawson added inline comments.

INLINE COMMENTS

> dfaure wrote in kconfigbackend.h:210
> In general I agree that leaving dead code is bad.
> 
> In this particular case, though, the other side of the plugin code is still there
> ("#if 0 // TODO port to Qt5 plugin loading" in kconfigbackend.cpp), and actually this is the bit that *was* ported to Qt5 plugin loading, so it will be useful if anyone decides to finally implement plugin based backends. Maybe I can make it
> #if 0 // TODO re-enable if the plugin loading code is re-enabled
> instead of just "//" which is indeed uglier.
> 
> Alternatively, we decide that plugin-based backends are not even on the far-future radar and we clean up both sides completely. Works for me too. The whole idea was crazy anyway (apps make assumptions about config files being files, if one day we change that for e.g. something like gconf then it won't be as transparent as just switching to a different backend).

Sure, #if 0 is best since it has the code, and they will at least match.

I'd actually like to try and bring this back sometime, but as you say it would be very difficult.  I don't know if it will ever come to pass, so I don't want to hold up actual improvements for a nice to have.  I definitely don't know if it would ever be plugin based, or just built into KConfig.

REPOSITORY
  R237 KConfig

REVISION DETAIL
  https://phabricator.kde.org/D4604

EMAIL PREFERENCES
  https://phabricator.kde.org/settings/panel/emailpreferences/

To: dfaure, mdawson
Cc: #frameworks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20170220/fe6da921/attachment.html>