CODEWITHSUNDEEP
c++string
J S
J S

Reputation: 3496

How does string::npos know which string I am referring to?

I am working through a book on C++ and it just covered using string::npos to check if a character position exists in a string. I don't understand how this mechanism could possibly know which string I'm referring to, though! This specific code is counting the number of occurrences of a substring.

Code:

for (int i=cats.find("cat",0);i!=string::npos;i=cats.find("cat",i)) {
    ++catCount;
    ++i;
}

I understand that it is starting the loop at the first occurrence of the word, incrementing the counter every pass through to avoid counting the same substring twice, and then at the end of each loop the counter is jumping to the position of the next occurrence of the substring. The loop stops when the counter does not exist as a character index for the string.

The string is called cats though, and "cats" is nowhere to be found in "string::npos" so how in the heck does it know that's the variable I'm even searching? Is it simply because that was the last variable to call .find()?

Thanks!

Upvotes: 5

Views: 2690

Answers (3)

Yakk - Adam Nevraumont
Yakk - Adam Nevraumont

Reputation: 275820

find returns a std::string::size_type. std::string::npos is a constant of that type which is returned when the value cannot be found.

Note that std::string::size_type is an unsigned value, and int is signed. If std::string::npos cannot be represented as an int, then the conversion from std::string::npos to int is undefined behavior.

So you really shouldn't store the return value of std::string::find in an int. Instead, you should store it in a std::string::size_type, or in C++11 use auto. baisc_string<char> size_type is std::size_t, as is most other specializations.

Upvotes: 5

Shafik Yaghmour
Shafik Yaghmour

Reputation: 158599

It does not know what string you are referring to, npos is just a static const member that represents the maximum value representable and in this case represents and eror, if we look at what cppreference says about std::basic_string::npos:

static const size_type npos = -1;

This is a special value equal to the maximum value representable by the type size_type. The exact meaning depends on context, but it is generally used either as end of string indicator by the functions that expect a string index or as the error indicator by the functions that return a string index.

which matches the definition for npos in the draft C++ standard in section 21.4 Class template basic_string paragraph 5:

static const size_type npos = -1;

which is a bit odd since since size_type is unsigned but is due to the integral conversion rules in section 4.7 Integral conversions which says:

If the destination type is unsigned, the resulting value is the least unsigned integer congruent to the source integer (modulo 2n where n is the number of bits used to represent the unsigned type).[...]

guarantees that -1 will be converted to the largest unsigned value. It may be easier to see using the wording from the draft C99 standard which says:

Otherwise, if the new type is unsigned, the value is converted by repeatedly adding or subtracting one more than the maximum value that can be represented in the new type until the value is in the range of the new type.

which gives us MAX + 1 -1 which is MAX.

Upvotes: 3

Ed Swangren
Ed Swangren

Reputation: 124760

If find cannot find what you are looking for, it returns a sentinel value, std::string::npos. There is no need to know anything about the string itself. All that is needed is to return a value which cannot be a valid index.

For example, it could be implemented as:

static const size_t npos = std::numeric_limits<size_t>::max();

size_t string::find(...)
{
    // if we didn't find it...
    return npos;
}

Also, you should not be using an int to store the return value as it is not what find returns. What if you have a very long string and the index returned is > numeric_limits<int>.max()? Well now you have invoked undefined behavior.

Upvotes: 9

Related Questions