defining new types

[Joe Laughlin]

I'm sure there's a fairly easy answer for this... but how can I define a new
type with range checking?

Example: I want to define a new type that's like a double, except that you
can only give it values from 0.0 to 100.0. I'd also like it to act like a
double as much as possible, except that an exception is thrown when it's set
to an invalid number.

Ideas?

Thanks,
Joe

[Alf P. Steinbach]

* Joe Laughlin:

I'm sure there's a fairly easy answer for this... but how can I define a new
type with range checking?

Example: I want to define a new type that's like a double, except that you
can only give it values from 0.0 to 100.0. I'd also like it to act like a
double as much as possible, except that an exception is thrown when it's set
to an invalid number.

Ideas?

The difference between original C++ and C was that C++ had classes.

A class is a type.

To define a new type, define a class.

Supply the operations you want the type to have.

Get yourself a good C++ book, e.g. "Accelerate d C++".

--
A: Because it messes up the order in which people normally read text.
Q: Why is it such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?

[Gianni Mariani]

Joe Laughlin wrote:

I'm sure there's a fairly easy answer for this... but how can I define a new
type with range checking?

Example: I want to define a new type that's like a double, except that you
can only give it values from 0.0 to 100.0. I'd also like it to act like a
double as much as possible, except that an exception is thrown when it's set
to an invalid number.

Ideas?

This google groups link points to a recent discussion on comp.std.c++.

http://tinyurl.com/6zyg9

[JKop]

Joe Laughlin posted:

I'm sure there's a fairly easy answer for this... but how can I define
a new type with range checking?

Example: I want to define a new type that's like a double, except that
you can only give it values from 0.0 to 100.0. I'd also like it to act
like a double as much as possible, except that an exception is thrown
when it's set to an invalid number.

Ideas?

Thanks,
Joe

Use your brain:

class RestrictiveDoub le
{
public: class bad_proposal {};

private:

double data;

Set(double const proposed)
{
if (propose > 100 || proposed < 0) throw bad_proposal;

data = proposed;
}

public:

RestrictiveDoub le& operator=(doubl e const proposed)
{
Set(proposed);
}

//Copy constructor is unnecessary

//Put a constructor here

//Put an operator double here
};

I would've finished it for you, but then half way through I thought it may
have been homework for you.

-JKop

[Joe Laughlin]

JKop wrote:

Joe Laughlin posted:

I'm sure there's a fairly easy answer for this... but
how can I define a new type with range checking?

Example: I want to define a new type that's like a
double, except that you can only give it values from 0.0
to 100.0. I'd also like it to act like a double as much
as possible, except that an exception is thrown when
it's set to an invalid number.

Ideas?

Thanks,
Joe

Use your brain:

class RestrictiveDoub le
{
public: class bad_proposal {};

private:

double data;

Set(double const proposed)
{
if (propose > 100 || proposed < 0) throw
bad_proposal;

data = proposed;
}

public:

RestrictiveDoub le& operator=(doubl e const proposed)
{
Set(proposed);
}

//Copy constructor is unnecessary

//Put a constructor here

//Put an operator double here
};

I would've finished it for you, but then half way through
I thought it may have been homework for you.

-JKop

What's an "operator double"? And I'm confused why you are defining a class
bad_proposal inside of RestrictiveDoub le, and then throwing it.

[JKop]

What's an "operator double"? And I'm confused why you are defining a
class bad_proposal inside of RestrictiveDoub le, and then throwing it.

Sorry, I'd like to help you, but I still suspect that this is some sort of
homework question.

If you have a decent book on C++, then go to the chapter on "operator
overloading", the conversion operators will be in there with it.

As regards defining one class within another: All it means is that, instead
of the class's name being "bad_propos al", its name is
"RestrictiveDou ble::bad_propos al". Also, if I were to define the
"bad_propos al" class within the private section of the "RestrictiveDou ble"
class definition, then it would be inaccessible from outside of the class's
own code. (Also I wouldn't be able to throw it as the caller wouldn't be
able to play with it).
-JKop

[Joe Laughlin]

JKop wrote:

What's an "operator double"? And I'm confused why you
are defining a class bad_proposal inside of
RestrictiveDoub le, and then throwing it.
Sorry, I'd like to help you, but I still suspect that
this is some sort of homework question.

Not homework.
If you have a decent book on C++, then go to the chapter
on "operator overloading", the conversion operators will
be in there with it.
I understand about operator overloading, just never heard of "operator
double". I've only heard of the usual operator==, operator>>, etc.

As regards defining one class within another: All it
means is that, instead of the class's name being
"bad_propos al", its name is
"RestrictiveDou ble::bad_propos al". Also, if I were to
define the "bad_propos al" class within the private
section of the "RestrictiveDou ble" class definition, then
it would be inaccessible from outside of the class's own
code. (Also I wouldn't be able to throw it as the caller
wouldn't be able to play with it).
-JKop

[JKop]

I understand about operator overloading, just never heard of "operator
double". I've only heard of the usual operator==, operator>>, etc.

Here's the jist of it:
class Blah
{
public:

operator bool() const
{
return true;
}
};

void SomeFunc(bool const monkey)
{

}
intm main()
{
Blah blah_object;

SomeFunc(blah_o bject);
}

[Joe Laughlin]

JKop wrote:

I understand about operator overloading, just never
heard of "operator double". I've only heard of the
usual operator==, operator>>, etc.

Here's the jist of it:
class Blah
{
public:

operator bool() const
{
return true;
}
};

void SomeFunc(bool const monkey)
{

}
intm main()
{
Blah blah_object;

SomeFunc(blah_o bject);
}

So, you can use Blah anywhere you can use a bool?

[JKop]

Joe Laughlin posted:

JKop wrote:

I understand about operator overloading, just never
heard of "operator double". I've only heard of the usual operator==,
operator>>, etc.

Here's the jist of it:
class Blah
{
public:

operator bool() const {
return true; } };

void SomeFunc(bool const monkey)
{

}
intm main()
{
Blah blah_object;

SomeFunc(blah_o bject); }

So, you can use Blah anywhere you can use a bool?

Yep, for instance:

while (blah_object)
{
;
}

if (blah_object) ;

-JKop