Question about using namespace std

[Schizoid Man]

Hi,

I'm a novice whose just about migrating to C++ after cutting my teeth on
C for a few years.

To start with I'm using Microsoft's out-of-the-box Visual C++ Express
Edition compiler and I had a couple of questions:

1. If I am using namespace std, is there a specific reason for me to
place std:: before all the manipulators - cout, endl, etc? I've noticed
that without the std:: prefix, all the methods work as expected.

2. I've been skimming Deitel's How to Program in C++ and noticed
something very strange. In the chapter on functions, the author's code
contains the function method AFTER the main() method, even though the
main() method calls the function. This would throw a compile error in C,
so is this valid syntax in C++?

3. Another question about bridging C and C++: I do know that an implicit
conversion from void* to double* is completely legal in C, but (so I'm
told) illegal in C++. Is there a reason for the latter?

Thanks in advance,
Schiz

 

[Alf P. Steinbach]

* Schizoid Man:

I'm a novice whose just about migrating to C++ after cutting my teeth on
C for a few years.

To start with I'm using Microsoft's out-of-the-box Visual C++ Express
Edition compiler and I had a couple of questions:

1. If I am using namespace std, is there a specific reason for me to
place std:: before all the manipulators - cout, endl, etc? I've noticed
that without the std:: prefix, all the methods work as expected.

Namespace qualification (writing "std::") is required unless you have a
"using" declaration or directive.

See the FAQ item "Should I use using namespace std in my code?"
currently at <url:
http://www.parashift.com/c++-faq-lite/coding-standards.html#faq-27.5>.

2. I've been skimming Deitel's How to Program in C++ and noticed
something very strange. In the chapter on functions, the author's code
contains the function method AFTER the main() method, even though the
main() method calls the function. This would throw a compile error in C,
so is this valid syntax in C++?

AFAIK in C you /can/ use an undeclared function, and that's part of the
reason why it's not a good idea to cast the result of malloc in C. In
C++ you /can't/ use a so far undeclared free-standing function. So if
you're talking about free-standing functions, then the assertion in your
last sentence has it exactly backwards.

However, it seems you're talking about a class with a member function
'main'. Generally a class that "does" is a Bad Idea (TM), except for
functor and functoid classes. See the FAQ item "] What the heck is a
functionoid, and why would I use one?" currently at <url:
http://www.parashift.com/c++-faq-lite/pointers-to-members.html#faq-33.10>.
Another exception: sometimes a class includes a "doer" function or
functions for purposes of automated unit testing. But in general, steer
away from thinking of a class as "doing" something: that's a function.

Anyway, when you write a class with inline function bodies like

struct S
{
void foo() { bar(); }
void bar() { say( "Heh" ); }
};

it's as if you wrote (it's a shorthand notation for)

struct S
{
void foo();
void bar();
};

inline void S::foo() { bar(); }

inline void S::bar() { say( "Heh" ); }

and here you can see that S::foo doesn't actually call a function S::bar
that hasn't yet been declared: it's just the shorthand notation for
defining inline member functions, where you place the function body in
the class definition, that makes it seem that way.

3. Another question about bridging C and C++: I do know that an implicit
conversion from void* to double* is completely legal in C, but (so I'm
told) illegal in C++. Is there a reason for the latter?

C++ is more type-oriented than C, attempting to have somewhat stronger
static type checking.

 

[Mark P]

Hi,

2. I've been skimming Deitel's How to Program in C++ and noticed
something very strange. In the chapter on functions, the author's code
contains the function method AFTER the main() method, even though the
main() method calls the function. This would throw a compile error in C,
so is this valid syntax in C++?

I haven't seen the book, but perhaps the author has declared the
functions called in main somewhere earlier. From the sounds of it, I
would guess that they're declared in an included header file.

This is legal:



// foo.h

void foo(); // no definition, just a declaration




// foo.cpp

#include "foo.h"

int main ()
{
foo();
}

void foo()
{
// do stuff
}

 

[Jim Langston]

Hi,

I'm a novice whose just about migrating to C++ after cutting my teeth on C
for a few years.

To start with I'm using Microsoft's out-of-the-box Visual C++ Express
Edition compiler and I had a couple of questions:

1. If I am using namespace std, is there a specific reason for me to place
std:: before all the manipulators - cout, endl, etc? I've noticed that
without the std:: prefix, all the methods work as expected.

No, there is no need to use std:: before manipulators if you are using
namespace std. But you can, as you have found, use them anyway.

It is my personal opionion, and many others, never to use using namespace
std as it brings everything into the unnamed namespace, defeating the
purpose of namespaces in the first place. A better way is not to use using
at all but prefix with std:: always, or you can bring in the specific things
you want.

using std::cout;
using std::endl;
etc..

2. I've been skimming Deitel's How to Program in C++ and noticed something
very strange. In the chapter on functions, the author's code contains the
function method AFTER the main() method, even though the main() method
calls the function. This would throw a compile error in C, so is this
valid syntax in C++?

I'm guessing this is for methods in a class?

3. Another question about bridging C and C++: I do know that an implicit
conversion from void* to double* is completely legal in C, but (so I'm
told) illegal in C++. Is there a reason for the latter?

C++ is much more strict on types than C. Makes it better so that we don't
make as many stupid mistakes.