On Sunday, 2014-02-16, 06:38:35, Michael Stemle wrote: > So, I'm still getting used to the dpointer paradigm. Is anybody familiar > with how to use QTest to simulate code paths dependent on values in the > dpointer which have no exposed mutators? > > For instance, I have QApt::Transaction which has its dpointer, and its > dpointer has a property "error". This property is exposed via a Q_PROPERTY > macro in the public class, but it's read-only. I'm adding error handling > and I'd like to add some unit tests for the error handling, but I'm having > trouble figuring out how to force an instance of QApt::Transaction into > executing that error handling without violating encapsulation or adding a > klugy method just for my test.
How about this:
Put the declaration of the Private in a header, say yourclass_p.h
This header does not get installed but can be included by the test.
Add a protected constructor to your class that takes an instance of the
Private
class MyClass
{
protected:
explicit MyClass(MyClassPrivate *d) : d_ptr(d) {} // maybe assert on !d
};
Your test can the do something like
MyClassPrivate *dTest = new MyClassPrivate;
class MyClassTest : public MyClass
{
public:
MyClassTest() : MyClass(dTest) {}
};
Cheers,
Kevin
--
Kevin Krammer, KDE developer, xdg-utils developer
KDE user support, developer mentoring
signature.asc
Description: This is a digitally signed message part.
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
