Mastering C++ Programming

4.2 (12 reviews total)
By Jeganathan Swaminathan
  • Instant online access to over 7,500+ books and videos
  • Constantly updated with 100+ new titles each month
  • Breadth and depth in over 1,000+ technologies
  1. C++17 Features

About this book

C++ has come a long way and has now been adopted in several contexts. Its key strengths are its software infrastructure and resource-constrained applications. The C++ 17 release will change the way developers write code, and this book will help you master your developing skills with C++. With real-world, practical examples explaining each concept, the book will begin by introducing you to the latest features in C++ 17. It encourages clean code practices in C++ in general, and demonstrates the GUI app-development options in C++. You’ll get tips on avoiding memory leaks using smart-pointers. Next, you’ll see how multi-threaded programming can help you achieve concurrency in your applications. Moving on, you’ll get an in-depth understanding of the C++ Standard Template Library. We show you the concepts of implementing TDD and BDD in your C++ programs, and explore template-based generic programming, giving you the expertise to build powerful applications. Finally, we’ll round up with debugging techniques and best practices.By the end of the book, you’ll have an in-depth understanding of the language and its various facets.

Publication date:
September 2017
Publisher
Packt
Pages
384
ISBN
9781786461629

 

Chapter 1. C++17 Features

In this chapter, you will be learning the following concepts:

  • C++17 background
  • What is new in C++17?
  • What features are deprecated or removed in C++17?
  • Key features in C++17 
 

C++17 background


As you know, the C++ language is the brain child of Bjarne Stroustrup, who developed C++ in 1979. The C++ programming language is standardized by International Organization for Standardization (ISO).

The initial standardization was published in 1998, commonly referred to as C++98, and the next standardization C++03 was published in 2003, which was primarily a bug fix release with just one language feature for value initialization. In August 2011, the C++11 standard was published with several additions to the core language, including several significant interesting changes to the Standard Template Library (STL); C++11 basically replaced the C++03 standard. C++14 was published in December, 2014 with some new features, and later, the C++17 standard was published on July 31, 2017.  

At the time of writing this book, C++17 is the latest revision of the ISO/IEC standard for the C++ programming language.

This chapter requires a compiler that supports C++17 features: gcc version 7 or later. As gcc version 7 is the latest version at the time of writing this book, I'll be using gcc version 7.1.0 in this chapter.

Note

In case you haven't installed g++ 7 that supports C++17 features, you can install it with the following commands:sudo add-apt-repository ppa:jonathonf/gcc-7.1  sudo apt-get update  sudo apt-get install gcc-7 g++-7

 

What's new in C++17?


The complete list of C++17 features can be found at http://en.cppreference.com/w/cpp/compiler_support#C.2B.2B17_features.

To give a high-level idea, the following are some of the new C++17 features:

  • New auto rules for direct-list-initialization
  • static_assert with no messages
  • Nested namespace definition
  • Inline variables
  • Attributes for namespaces and enumerators
  • C++ exceptions specifications are part of the type system
  • Improved lambda capabilities that give performance benefits on servers
  • NUMA architecture
  • Using attribute namespaces
  • Dynamic memory allocation for over-aligned data
  • Template argument deduction for class templates
  • Non-type template parameters with auto type
  • Guaranteed copy elision
  • New specifications for inheriting constructors
  • Direct-list-initialization of enumerations
  • Stricter expression evaluation order
  • shared_mutex 
  • String conversions

Otherwise, there are many new interesting features that were added to the core C++ language: STL, lambadas, and so on. The new features give a facelift to C++, and starting from C++17, as a C++ developer, you will feel that you are working in a modern programming language, such as Java or C#.

What features are deprecated or removed in C++17?

The following features are now removed in C++17:

  • The register keyword was deprecated in C++11 and got removed in C++17
  • The ++ operator for bool was deprecated in C++98 and got removed in C++17
  • The dynamic exception specifications were deprecated in C++11 and and got removed in C++17
 

Key features in C++17


Let's explore the following C++17 key features one by one in the following sections:

  • Easier nested namespace
  • New rules for type detection from the braced initializer list
  • Simplified static_assert
  • std::invoke
  • Structured binding
  • The if and switch local-scoped variables
  • Template type auto-detection for class templates
  • Inline variables

Easier nested namespace syntax

Until the C++14 standard, the syntax supported for a nested namespace in C++ was as follows:

#include <iostream>
using namespace std;

namespace org {
    namespace tektutor {
        namespace application {
             namespace internals {
                  int x;
             }
        }
    }
}

int main ( ) {
    org::tektutor::application::internals::x = 100;
    cout << "\nValue of x is " << org::tektutor::application::internals::x << endl;

    return 0;
}

The preceding code can be compiled and the output can be viewed with the following commands:

g++-7 main.cpp -std=c++17
./a.out

The output of the preceding program is as follows:

Value of x is 100

Every namespace level starts and ends with curly brackets, which makes it difficult to use nested namespaces in large applications. C++17 nested namespace syntax is really cool; just take a look at the following code and you will readily agree with me:

#include <iostream>
using namespace std;

namespace org::tektutor::application::internals {
    int x;
}

int main ( ) {
    org::tektutor::application::internals::x = 100;
    cout << "\nValue of x is " << org::tektutor::application::internals::x << endl;

    return 0;
}

The preceding code can be compiled and the output can be viewed with the following commands:

g++-7 main.cpp -std=c++17
./a.out

The output remains the same as the previous program:

Value of x is 100

New rules for type auto-detection from braced initializer list 

C++17 introduced new rules for auto-detection of the initializer list, which complements C++14 rules. The C++17 rule insists that the program is ill-formed if an explicit or partial specialization of std::initializer_list is declared:

#include <iostream>
using namespace std;

template <typename T1, typename T2>
class MyClass {
     private:
          T1 t1;
          T2 t2;
     public:
          MyClass( T1 t1 = T1(), T2 t2 = T2() ) { }

          void printSizeOfDataTypes() {
               cout << "\nSize of t1 is " << sizeof ( t1 ) << " bytes." << endl;
               cout << "\nSize of t2 is " << sizeof ( t2 ) << " bytes." << endl;
     }
};

int main ( ) {

    //Until C++14
    MyClass<int, double> obj1;
    obj1.printSizeOfDataTypes( );

    //New syntax in C++17
    MyClass obj2( 1, 10.56 );

    return 0;
}

The preceding code can be compiled and the output can be viewed with the following commands:

g++-7 main.cpp -std=c++17
./a.out

The output of the preceding program is as follows:

Values in integer vectors are ...
1 2 3 4 5 

Values in double vectors are ...
1.5 2.5 3.5

Simplified static_assert 

The static_assert macro helps identify assert failures during compile time. This feature has been supported since C++11; however, the static_assert macro used to take a mandatory assertion failure message till, which is now made optional in C++17.

The following example demonstrates the use of static_assert with and without the message:

#include <iostream>
#include <type_traits>
using namespace std;

int main ( ) {

        const int x = 5, y = 5;

        static_assert ( 1 == 0, "Assertion failed" );
        static_assert ( 1 == 0 );
        static_assert ( x == y );

        return 0;
}

The output of the preceding program is as follows:

g++-7 staticassert.cpp -std=c++17
staticassert.cpp: In function ‘int main()’:
staticassert.cpp:7:2: error: static assertion failed: Assertion failed
  static_assert ( 1 == 0, "Assertion failed" );

staticassert.cpp:8:2: error: static assertion failed
  static_assert ( 1 == 0 );

From the preceding output, you can see that the message, Assertion failed, appears as part of the compilation error, while in the second compilation the default compiler error message appears, as we didn't supply an assertion failure message. When there is no assertion failure, the assertion error message will not appear as demonstrated in static_assert ( x == y ).  This feature is inspired by the C++ community from the BOOST C++ library.

The std::invoke( ) method

The std::invoke() method can be used to call functions, function pointers, and member pointers with the same syntax:

#include <iostream>
#include <functional>
using namespace std;

void globalFunction( ) {
     cout << "globalFunction ..." << endl;
}

class MyClass {
    public:
        void memberFunction ( int data ) {
             std::cout << "\nMyClass memberFunction ..." << std::endl;
        }

        static void staticFunction ( int data ) {
             std::cout << "MyClass staticFunction ..." << std::endl;
        }
};

int main ( ) {

    MyClass obj;

    std::invoke ( &MyClass::memberFunction, obj, 100 );
    std::invoke ( &MyClass::staticFunction, 200 );
    std::invoke ( globalFunction );

    return 0;
}

The preceding code can be compiled and the output can be viewed with the following commands:

g++-7 main.cpp -std=c++17
./a.out

The output of the preceding program is as follows:

MyClass memberFunction ...
MyClass staticFunction ...
globalFunction ...

The std::invoke( ) method is a template function that helps you seamlessly invoke callable objects, both built-in and user-defined.

Structured binding

You can now initialize multiple variables with a return value with a really cool syntax, as shown in the following code sample:

#include <iostream>
#include <tuple>
using namespace std;

int main ( ) {

    tuple<string,int> student("Sriram", 10);
auto [name, age] = student;

    cout << "\nName of the student is " << name << endl;
    cout << "Age of the student is " << age << endl;

    return 0;
}

In the preceding program, the code highlighted in bold is the structured binding feature introduced in C++17. Interestingly, we have not declared the string name and int age variables. These are deduced automatically by the C++ compiler as string and int, which makes the C++ syntax just like any modern programming language, without losing its performance and system programming benefits. 

The preceding code can be compiled and the output can be viewed with the following commands:

g++-7 main.cpp -std=c++17
./a.out

The output of the preceding program is as follows:

Name of the student is Sriram
Age of the student is 10

If and Switch local scoped variables

There is an interesting new feature that allows you to declare a local variable bound to the if and switch statements' block of code. The scope of the variable used in the if and switch statements will go out of scope outside the respective blocks. It can be better understood with an easy to understand example, as follows:

#include <iostream>
using namespace std;

bool isGoodToProceed( ) {
    return true;
}

bool isGood( ) {
     return true;
}

void functionWithSwitchStatement( ) {

     switch ( auto status = isGood( ) ) {
          case true:
                 cout << "\nAll good!" << endl;
          break;

          case false:
                 cout << "\nSomething gone bad" << endl;
          break;
     } 

}

int main ( ) {

    if ( auto flag = isGoodToProceed( ) ) {
         cout << "flag is a local variable and it loses its scope outside the if block" << endl;
    }

     functionWithSwitchStatement();

     return 0;
}

The preceding code can be compiled and the output can be viewed with the following commands:

g++-7 main.cpp -std=c++17
./a.out

The output of the preceding program is as follows:

flag is a local variable and it loses its scope outside the if block
All good!

Template type auto-deduction for class templates

I'm sure you will love what you are about to see in the sample code.  Though templates are quite useful, a lot of people don't like it due to its tough and weird syntax. But you don't have to worry anymore; take a look at the following code snippet:

#include <iostream>
using namespace std;

template <typename T1, typename T2>
class MyClass {
     private:
          T1 t1;
          T2 t2;
     public:
          MyClass( T1 t1 = T1(), T2 t2 = T2() ) { }

          void printSizeOfDataTypes() {
               cout << "\nSize of t1 is " << sizeof ( t1 ) << " bytes." << endl;
               cout << "\nSize of t2 is " << sizeof ( t2 ) << " bytes." << endl;
     }
};

int main ( ) {

    //Until C++14
    MyClass<int, double> obj1;
    obj1.printSizeOfDataTypes( );

    //New syntax in C++17
    MyClass obj2( 1, 10.56 );

    return 0;
}

The preceding code can be compiled and the output can be viewed with the following commands:

g++-7 main.cpp -std=c++17
./a.out

The output of the program is as follows:

Size of t1 is 4 bytes.
Size of t2 is 8 bytes.

Inline variables

Just like the inline function in C++, you could now use inline variable definitions. This comes in handy to initialize static variables, as shown in the following sample code:

#include <iostream>
using namespace std;

class MyClass {
    private:
        static inline int count = 0;
    public:
        MyClass() { 
              ++count;
        }

    public:
         void printCount( ) {
              cout << "\nCount value is " << count << endl;
         } 
};

int main ( ) {

    MyClass obj;

    obj.printCount( ) ;

    return 0;
}

The preceding code can be compiled and the output can be viewed with the following commands:

g++-7 main.cpp -std=c++17
./a.out

The output of the preceding code is as follows:

Count value is 1
 

Summary


In this chapter, you got to know interesting new features introduced in C++17. You learned the super simple C++17 nested namespace syntax. You also learned datatype detection with a braced initializer list and the new rule imposed in the C++17 standard.

You also noticed that static_assert can be done without assert failure messages. Also, using std::invoke(), you can now invoke global functions, function pointers, member functions, and static class member functions. And, using structured binding, you could now initialize multiple variables with a return value.

You also learned that the if and switch statements can have a local-scoped variable right before the if condition and switch statements. You learned about auto type detection of class templates. Lastly, you used inline variables.

There are many more C++17 features, but this chapter attempts to cover the most useful features that might be required for most of the developers.  In the next chapter, you will be learning about the Standard Template Library.

About the Author

  • Jeganathan Swaminathan

    Jeganathan Swaminathan, Jegan for short, is a freelance software consultant and founder of TekTutor, with over 17 years of IT industry experience. In the past, he has worked for AMD, Oracle, Siemens, Genisys Software, Global Edge Software Ltd, and PSI Data Systems. He has consulted for Samsung WTD (South Korea) and National Semiconductor (Bengaluru). He now works as a freelance external consultant for Amdocs (India). He works as freelance software consultant and freelance corporate trainer. He holds CSM, CSPO, CSD, and CSP certifications from Scrum Alliance. He is a polyglot software professional and his areas of interest include a wide range of C++, C#, Python, Ruby, AngularJS, Node.js, Kubernetes, Ansible, Puppet, Chef, and Java technologies. He is well known for JUnit, Mockito, PowerMock, gtest, gmock, CppUnit, Cucumber, SpecFlow, Qt, QML, POSIX – Pthreads, TDD, BDD, ATDD, NoSQL databases (MongoDB and Cassandra), Apache Spark, Apache Kafka, Apache Camel, Dockers, Continuous Integration (CI), Continuous Delivery (CD), Maven, Git, cloud computing, and DevOps. You can reach him for any C++, Java, Qt, QML, TDD, BDD, and DevOps-related training or consulting assignments. Jegan is a regular speaker at various technical conferences.

    Browse publications by this author

Latest Reviews

(12 reviews total)
It's help me to enhace my knolwledge in modern C++.
I gave up on C++ twenty years ago--it added a lot of complexity but almost no additional function I couldn't get more efficiently using straight C. A bit ago, I decided to take another look at C++ and discovered it had gotten a whole lot better over those twenty years--Mastering C++ has been a big help in getting me up to speed on it.
Be aware that is based on C++17 standard, and it doesn't have the option to export to Kindle

Recommended For You

C++ Fundamentals

Write high-level abstractions while retaining full control of the hardware, performances, and maintainability.

By Antonio Mallia and 1 more
Extreme C

Push the limits of what C - and you - can do, with this high-intensity guide to the most advanced capabilities of C

By Kamran Amini
Hands-On Network Programming with C

A comprehensive guide to programming with network sockets, implementing internet protocols, designing IoT devices, and much more with C

By Lewis Van Winkle
C Programming Cookbook

A comprehensive guide with curated recipes to help you gain a deeper understanding of modern C.

By B. M. Harwani