[TriLUG] OT: C++ Question

[Brian Henning]

Ooo, good catch, Tanner.  I bet if we saw a "stack trace" of the syntax 
error, it'd be something along the lines of..
"In file included from point3d.h (vector3d.h):
   In file included from vector3d.h (point3d.h):
    In file point3d.h:
     Syntax error, expected identifier yadda yadda yadda"

I believe Tanner is right about the forward declaration being one 
solution, as well as it being a good indication that you may want to 
rethink your class design/heirarchy.

Of course, my geometry skills are very faded at this point, but can you 
really subtract a point from a vector?  I wouldn't know how to do that. 
  In my ignorance of geometric operations, I'd suggest removing the 
operator-(point3d) function altogether.  That's the only member of 
vector3d (such as it is now) that has a reference to point3d.  If you 
really need the function, maybe make a reverse-logic member function of 
point3d such as vector3d subtractFrom(vector3d) which would subtract 
itself from a vector and return the resultant vector.

Cheers,
~Brian



Tanner Lovelace wrote:
> On 1/31/06, Randy Barlow <rpbarlow at ncsu.edu> wrote:
> 
>> Thanks Brian (and Tanner!)  This was definitely an issue in my code, but
>>I'm still having problems with the vector3d identifier in my point3d.h file.
>> Here is point3d.h:
>>
>> #pragma once
>> #include "vector3d.h"
> 
> 
> [...snip...]
> 
> 
>>line in the code above.  Here is vector3d.h:
>>
>> #pragma once
>> #include "point3d.h"
> 
> 
> There's your problems.  You've told the compiler to only include
> things once and then gone and included them in each other.
> That's most definitely *not* going to work because, although
> the #pragma stops the infinite recursion of includes, you reference
> them from each class so one of the classes will be defined before
> the other and won't know anything about the other.  You could try
> adding a forward declaration, but I really think you should just
> bit the bullet and rethink your code.
> 
> Cheers,
> Tanner
> 
> --
> Tanner Lovelace
> clubjuggler at gmail dot com
> http://wtl.wayfarer.org/
> (fieldless) In fess two roundels in pale, a billet fesswise and an
> increscent, all sable.

-- 
----------------
Brian A. Henning
strutmasters.com
336.597.2397x238
----------------