Is this a kdelibs regression?

[Dominik Haumann]

On Tuesday, February 12, 2013 00:56:05 Albert Astals Cid wrote:
> Someone pointed me to
> http://quickgit.kde.org/?p=kate.git&a=commitdiff&h=3adf3cc5473c533f382ae8a48
> 1e452289a56bab3&hp=5addc104e5c181c25a3f16d596602d2755db2128
> 
> I have not had time to look at it but " fix plugin loading in KDE SC 4.10.0
> " as commit message looks a bit scary, anyone has time to see if that was
> bad coding or is a real regression on the libs side?

The bug for that is: https://bugs.kde.org/show_bug.cgi?id=314530

It is certainly bad coding in the sense that the <= was probably never
tested in the Kate code. It obviously does not behave as wanted, but
that probably could be expected.

The same bug probably exists when loading Kate app plugins in
kate/app/katepluginmanager.cpp in KatePluginManager::setupPluginList
Here, X-Kate-Version is used through

  // idea: in version 2.8, the Kate application interfaces changed (BIC)
  double pVersion = ptr->property("X-Kate-Version").toDouble();
  if ((pVersion >= 2.8) && (pVersion <= KateApp::kateVersion(false).toDouble()))

It should be done through properly by comparing major/minor versions, I guess.

Greetings
Dominik