string::npos

what is it ?

Code:

---

if (phrase. find("eggplant") == string::npos)
  cout << "'eggplant' is not in the phrase.\n\n";

---

explain briefly please

string::npos is a position value which means "not found." Think "no position."

whats the actual value of it ?

can't we do the same thing without npos ?

Quote:

---

Originally Posted by manzoor

whats the actual value of it ?

---

You're not supposed to care.

Quote:

---

Originally Posted by manzoor

whats the actual value of it ?

---

Who knows. The whole point of having it is so you don't care.

Quote:

---

can't we do the same thing without npos ?

---

No.

>> You're not supposed to care.
Agreed.

If you're curious, though, the actual value of it is defined as -1 converted to the type string::size_type which is an unsigned integral type.

Without even opening a book, you can find things out like this yourself with a little thinking...

printf("The value is string::npos is %s\n", string::npos) ;

oops - that compiles, but crashes... try something else

printf("The value is string::npos is %d\n", string::npos) ;

ok, that compiles, now run and see. ;)

Todd

Quote:

---

Originally Posted by Daved

>> You're not supposed to care.
Agreed.

If you're curious, though, the actual value of it is defined as -1 converted to the type string::size_type which is an unsigned integral type.

---

Err, I believe it has to be signed to be -1. ;)

What? You've never used -1U? :p

Quote:

---

Originally Posted by Todd Burch

Err, I believe it has to be signed to be -1. ;)

---

I think that's what was meant by:

Quote:

---

converted to the type string::size_type which is an unsigned integral type

---

Basically something like this:

Code:

---

typedef  static_cast<size_type>( -1 )  npos;

---

Quote:

---

Originally Posted by Todd Burch

Err, I believe it has to be signed to be -1. ;)

---

That's where a typecast comes into play. Here is a bit from Visual Studio 2008's (and 2005's) xstring header file:

Code:

---

// STATIC npos OBJECT
template<class _Elem,
        class _Traits,
        class _Alloc>
        _PGLOBAL const typename basic_string<_Elem, _Traits, _Alloc>::size_type
                basic_string<_Elem, _Traits, _Alloc>::npos =
                        (typename basic_string<_Elem, _Traits, _Alloc>::size_type)(-1);

---

MSVC 6's xstring header file:

Code:

---

template<class _E, class _Tr, class _A>
        const basic_string<_E,_Tr,_A>::size_type
                basic_string<_E,_Tr,_A>::npos = -1;

---

>> Without even opening a book, you can find things out like this yourself with a little thinking...
True, although there is a difference between the value that one particular implementation uses and a value that is required to be used by all standards conforming implementations.

That is correct. It should be treated as an opaque value. The doc I have on it goes on for a paragraph about the typical value (-1) and the string::size_type declaration, and to not treat it like an int, blah blah blah.

Hopefully the OP's curiosity has been satisfied!

>Hopefully the OP's curiosity has been satisfied!
If not we can probably go into more detail. ;)

Quote:

---

True, although there is a difference between the value that one particular implementation uses and a value that is required to be used by all standards conforming implementations.

---

That would be:

Code:

---

static const size_type npos = -1;

---

in the scope of std::basic_string.