Isn't it legal to put a class declaration in the same .cpp file where you 
define its member functions and the main() that's the only routine that uses it?

When I did that, I got errors about not being able to find the vtable for 
the class.

Is #include supposed to work just like a textual include? Is this something 
funky with the Qt preprocessing stuff?

-- 
   Darren New, San Diego CA, USA (PST)
   Insanity is a small city on the western
   border of the State of Mind.

Post a reply to this message

On Fri, 12 Jun 2009 20:22:16 +0200, Darren New <dne### [at] sanrrcom> wrote:
> Isn't it legal to put a class declaration in the same .cpp file where  
> you define its member functions and the main() that's the only routine  
> that uses it?

You can put the entire program in a single file if you want.



> When I did that, I got errors about not being able to find the vtable  
> for the class.

Possibly some internal compiler problem (which might be triggered by a  
syntactical problem in your code).



> Is #include supposed to work just like a textual include?

Yes.



-- 
FE

Post a reply to this message

Fredrik Eriksson wrote:
>> Is #include supposed to work just like a textual include?
> Yes.

OK. I think the problem is that Qt runs the code through a preprocessor for 
various reasons, and it was probably not running the .cpp stuff thru, just 
the .h stuff.

-- 
   Darren New, San Diego CA, USA (PST)
   Insanity is a small city on the western
   border of the State of Mind.

Post a reply to this message

Fredrik Eriksson <fe79}--at--{yahoo}--dot--{com> wrote:
> > Is #include supposed to work just like a textual include?

> Yes.

  With two minor exceptions:

  1) Certain compiler-dependent #pragmas may change the behavior of #include
slightly. The most common one is "#pragma once".

  2) __FILE__ will expand to the name of the header file if it appears in
it, rather than the source file where it's #included. Likewise for __LINE__.
Thus if either identifier is used in a header file, its behavior will be
different than if you copy-pasted the contents of the header file to the
source file.
  (The exception is if either identifier is used in a preprocessor macro, in
which case they are interpreted only after expanding the macro instantiation.)

-- 
                                                          - Warp

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> When I did that, I got errors about not being able to find the vtable for 
> the class.

  A minimal example code would help a lot answering the question.

-- 
                                                          - Warp

Post a reply to this message

Warp wrote:
> Fredrik Eriksson <fe79}--at--{yahoo}--dot--{com> wrote:
>>> Is #include supposed to work just like a textual include?
> 
>> Yes.
> 
>   With two minor exceptions:

Good point, thanks!  Neither of which was the problem, which seems to be 
specific to Qt, which runs its own preprocessors. But thanks for the 
clairifications.

-- 
   Darren New, San Diego CA, USA (PST)
   Insanity is a small city on the western
   border of the State of Mind.

Post a reply to this message

Warp wrote:
> Darren New <dne### [at] sanrrcom> wrote:
>> When I did that, I got errors about not being able to find the vtable for 
>> the class.
> 
>   A minimal example code would help a lot answering the question.

Yeah. I solved it before I asked the question. I was just trying to figure 
out if it was normal C++ stuff or QT stuff. Once I got confirmed that 
#include in C++ was essentially the same as in C (modulo leaving off .h), I 
went grubbing around QT and figured out there's a macro it only expands in 
.H files with the "moc" processor.

-- 
   Darren New, San Diego CA, USA (PST)
   Insanity is a small city on the western
   border of the State of Mind.

Post a reply to this message

Darren New wrote:
> Isn't it legal to put a class declaration in the same .cpp file where 
> you define its member functions and the main() that's the only routine 
> that uses it?
> 
> When I did that, I got errors about not being able to find the vtable 
> for the class.
> 
> Is #include supposed to work just like a textual include? Is this 
> something funky with the Qt preprocessing stuff?
> 

Perfectly legal.

As far as Qt is concerned... are you running moc by yourself?
You known, qmake uses the HEADERS variable to look for Q_OBJECT classes.

If your private class contains that macro, just try and include the 
generated .moc directly.

Something like:

source.cpp:

class my_private_implementation
{
	Q_OBJECT

	/* neat stuff */
};
#include "source.moc"

--

Jonathan.

Post a reply to this message

JRG wrote:
> As far as Qt is concerned... are you running moc by yourself?
> You known, qmake uses the HEADERS variable to look for Q_OBJECT classes.

I see. No, I'm a Qt newbie. I just worked out this weekend how to compile 
something outside of the Qt subdirectories. :-)  {To be fair, it's 
cross-compiled from a VM to a piece of hardware that doesn't have any 
keyboard or mouse, so it's not like it's trivial.}

> If your private class contains that macro, just try and include the 
> generated .moc directly.

Cool. It's easy enough to put it in a separate .h file. It just surprised 
me. Looking at it, I thought Q_OBJECT was a #define, not some tag for a 
separate preprocessor.


You know something else that surprises me? That I can screw up the 
compilation enough to get G++ outputting assembler that the assembler 
doesn't understand. I can't imagine why I can invoke G++ with arguments 
that, in one execution, would compile C++ to assembly, then pass that 
assembly to the assembler for the wrong architecture. Weird.

-- 
   Darren New, San Diego CA, USA (PST)
   Insanity is a small city on the western
   border of the State of Mind.

Post a reply to this message

Darren New wrote:
> Cool. It's easy enough to put it in a separate .h file. It just surprised
> me. Looking at it, I thought Q_OBJECT was a #define, not some tag for a
> separate preprocessor.

I believe it's both.

Post a reply to this message