RE: [Dev-C++] Newbie linking errors: 'undefined reference'

[Ioannis V. <no...@ya...>]

> -----Original Message-----
> From: dev...@li... 
> [mailto:dev...@li...] On Behalf 
> Of Devin Carless
> Sent: Saturday, February 09, 2002 11:55 PM
> To: Ioannis Vranos
> Cc: dev...@li...
> Subject: Re: [Dev-C++] Newbie linking errors: 'undefined reference'
> 
>   That's roughly what I  thought.  The function is, indeed, 
> mine; I'm not using a graphics library.  (The game is that 
> simple.)  I do have a definition in BaseBook.cpp.  The 
> definition refers to an object contained in the private 
> section of BaseBook, MoveParch, which has its own Lookup 
> function (I do it this way because subclasses of BaseBook may 
> want to redefine BaseBook::Lookup... WomanQSPrototype.cpp is 
> one of these subclasses.)
>   So plainly the linker isn't /reaching/ the definition.  The 
> method I use does require a few daisy-chained functions 
> (BaseBook::Lookup calling MoveParch::Lookup, and there are 
> other, three-step ones.) I don't know if this could make a 
> difference.  It does not seem to depend upon the legality of the code.
> 
>   Ah.  Genius lightbulb.  Here's an example of something 
> similar from a few days back:
> 
> bool ManeuverList::SetManeuver( short int ManI, char* NameA, short int
> BaseModA)
> // Inlining this gives linker error 'Undefined reference
> // to ManeuverList::Setmaneuver'.  Gotcha!
> // Same with the Index typedefl.
> {
>     // Maneuver::Set checks to see if we can
>     // set this maneuver.
>     bool b;
>     b = ManVec[ManI].Set( NameA, BaseModA );
>     return b;
> }
> 
>   Index was typedefed as 'unsigned short int.'  It worked 
> perfectly well elsewhere.  That, and making this function 
> inline caused the undefined reference linker errors.  


It may be for 1000 reasons, e.g. wrong or absense of definition of
ManVec, wrong operator[] definition for ManVec e.t.c..


> Previously, I was having trouble with size(): a while loop 
> explicitly told me both A and B were false (where B included 
> a size() function,) and A && B was false, /but while(A && B) 
> would still run through./
>   I don't know if that's related.  But it does show the level 
> of black-voodoo nbugs I'm running into.  Two functions 
> stopped giving errors when the definitions were moved from 
> the .cpp to the .h. 


Had you included the .cpp file in the project? Also if you have the
definition of a class in a .cpp file you must include a declaration of
the class in the scope you use it. For example:


in the a.cpp file:

class A
{
 // definition
};


in the file you use it:


class A;

A a;
//..


 An undefined reference to the 
> deconstructor went away when I moved the deconstructor up a 
> few lines!  So whatever is wrong, it's subtle - too subtle for me.


Yes i am sure it is noy error from the compiler... May i suggest a great
C++ book? "The C++ Programming Language" 3rd Edition or Special Edition
by Bjarne Stroustrup, the creator of C++. And he doesn't teaches only
C++ but very useful techniques and software engineering. Definitely a
great book!



Ioannis
 
* Ioannis Vranos
* Programming pages: http://www.noicys.d2g.com
* Alternative URL: http://run.to/noicys