Q_INCLUDE_MOC vs Q_DECLARE_OPAQUE_POINTER

[Redman]

Hi,

whats the difference between Q_INCLUDE_MOC and Q_DECLARE_OPAQUE_POINTER?
Maybe let me phrase the question in a different way:
Which one of the above is more appropriate for my usecase?
I have fordward declared classes which get used in Q_PORPERTY and signal/slots. For that to work the types must be fully-defined, as stated in qmetatype.h

static_assert(is_complete<Pointed, void>::value,
                    "Pointer Meta Types must either point to fully-defined types "
                    "or be declared with Q_DECLARE_OPAQUE_POINTER(T *)");

I can get my forwarddeclared classes to be accepted by Q_PROPERTY and signals/slots by using either one of the above mentioned macros.
So which one should I use?

[GrecKo]

What's Q_INCLUDE_MOC?
Q_DECLARE_OPAQUE_POINTER is definitely well suited for your use case.
Its doc says:

This macro enables pointers to forward-declared types (PointerType) to be registered with QMetaType [...]

[Redman]

@GrecKo said in Q_INCLUDE_MOC vs Q_DECLARE_OPAQUE_POINTER:

What's Q_INCLUDE_MOC?

I meant Q_MOC_INCLUDE
which says in the documentation

This is useful if the types you use as properties or signal/slots arguments are forward declared.

That's why I am not sure which one to use.
Since they both achieve the same for me right now, I 'd like to know what happens under the hood and which approach is more lightweight?

[felixonmars]

I got into similar situation today, but I had to use Q_MOC_INCLUDE because the class is a Q_OBJECT in my case. The Qt docs suggest that Q_DECLARE_OPAQUE_POINTER is incompatible in this case.

Hope this helps!