|
Setting up your first class
The division I did already in the simple examplatory class Ratio
can be followed more closely in the first "realistic" case of an
n-dimensional vector class. There, the class definition with the
methods can be found in an
["vecNd.H"-file], the realization in the
["vecNd.C"-file],
and some test routines are to be found in
["Main.C"-file].
To compile and run the code, please use the
["Makefile"].
There are a couple of things to be noted and explained. Let us start
with the H-file:
- The full file is surrounded by the following lines:
#ifndefine vecNd_h
#define vecNd_h
body of the file
#endif
|
The idea is that wherever in the code you need the class vecNd,
wherevere you wish to instantiate such an object, you have to
include the vecNd.H file to define this particular class and to
make all methods etc. known to the code.
However, C++ copes really bad with multiple definitions of
one and the same object. You are allowed to definee the full
class only once. This is, what the syntax above does. Wherever
the vecNd.H file is called via
it is checked wether a marker vecNd_h has been defined
already. If this is not the case, it will be defined,
and the rest of the class definition will be done.
- The class vecNd includes the Standard library math.h
to have access to functions like squareroot (sqrt) and the like.
It also includes the iostream.h. The main syntax difference
between files from the Standard library and your own class
definitions is exemplified below:
#include < math.h>
#include "vecNd.H"
|
- You see two constructors, both named vecNd, bit with different
arguments. Constructors (as well as destructors, see below) must
have the class name as name. However, the two construcotrs here
are either with parameters or with another vecNd (passed by
reference) as argument. The latter is called "copy constructor"
for obvious reasons, check the C-file.
- There's a thing called destructor, denoted by ~vecNd().
If you check in the C-file, you will see that this destructor
deletes the memebers of the vector array. Note that this is
urgently needed, whenever you allocate memory within your
constructor, i.e. whenever you use the command "new". The reason
is that all classes are defined to some extend within the
framework of a method, a class, a function etc. . Whenever the
flow of the executable program leaves this framework all
intrinsic variables are deleted automatically. However, for
arrays allocated via "new" this does not hold true and you have
to delete them "by hand". Copnsequently, you have to define a
specific "delete method" for classes that incorporate such
arrays. This method is called destructor.
- You will see that some methods are realized within the H-file and
with the prefix inline . Of course, the realization is
outside the class definition. The keyword inline has the
effect that the corresponding executable code is included
withinhg the larger executable instead of some look-up within the
class definition plus additional links. This is for reasons of
efficiency. Basically, inlines should be short, to ensure that
the executable chunk of code related to them are shorter than the
lookup procedure. Let me stress that inlines have to be in
H-files.
[prev]
[top]
|