The C++ Standard Library A Tutorial and Reference (Nicolai M. Josuttis) (Z-Library)

The C++ Standard Library 2 C++ Standard Library, The: A Tutorial and Reference Many of the designations used by manufacturers and sellers to distinguish their products are claimed as trademarks. Where those designations appear in this book and Addison Wesley Longman Inc., was aware of a trademark claim, the designations have been printed in initial caps or all caps. The authors and publisher have taken care in the preparation of this book, but make no expressed or implied warranty of any kind and assume no responsibility for errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of the use of the information or programs contained herein. The publisher offers discounts on this book when ordered in quantity for special sales. For more information, please contact: AWL Direct Sales Addison Wesley Longman, Inc One Jacob Way Reading, Massachusetts 01867 (781) 944-3700 Visit AW on the Web: www.awl.com/cseng/ Library of Congress Cataloging-in-Publication Data Josuttis, Nicolai M. The C++ standard library: a tutorial and reference / Nicolai M. Josuttis. p. cm. Includes bibliographical references and index. 1. C++ (Computer program language) I. Title. QA76.73.C153J69 1999 005.13'3--dc21 99-24977 CIP Copyright © 1999 by Addison Wesley Longman, Inc. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form, or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior consent of the publisher. Printed in the United States of America. Published simultaneously in Canada. 1 2 3 4 5 6 7 8 9 -CRW- 0302010099 First printing, July 1999

The C++ Standard Library 3 Table of Contents Preface Acknowledgments 1. About this Book 1.1 Why this Book 1.2 What You Should Know Before Reading this Book 1.3 Style and Structure of the Book 1.4 How to Read this Book 1.5 State of the Art 1.6 Example Code and Additional Information 1.7 Feedback 2. Introduction to C++ and the Standard Library 2.1 History 2.2 New Language Features 2.3 Complexity and the Big-O Notation 3. General Concepts 3.1 Namespace std 3.2 Header Files 3.3 Error and Exception Handling 3.4 Allocators 4. Utilities 4.1 Pairs 4.1.1 Convenience Function make_pair() 4.1.2 Examples of Pair Usage 4.2 Class auto_ptr 4.3 Numeric Limits 4.4 Auxiliary Functions 4.5 Supplementary Comparison Operators 4.6 Header Files <cstddef> and <cstdlib> 5. The Standard Template Library 5.1 STL Components 5.2 Containers 5.3 Iterators 5.4 Algorithms 5.5 Iterator Adapters 5.6 Manipulating Algorithms 5.7 User-Defined Generic Functions 5.8 Functions as Algorithm Arguments 5.9 Function Objects 5.10 Container Elements 5.11 Errors and Exceptions Inside the STL 5.12 Extending the STL 6. STL Containers 6.1 Common Container Abilities and Operations 6.2 Vectors 6.3 Deques

The C++ Standard Library 4 6.4 Lists 6.5 Sets and Multisets 6.6 Maps and Multimaps 6.7 Other STL Containers 6.8 Implementing Reference Semantics 6.9 When to Use which Container 6.10 Container Types and Members in Detail 7. STL Iterators 7.1 Header Files for Iterators 7.2 Iterator Categories 7.3 Auxiliary Iterator Functions 7.4 Iterator Adapters 7.5 Iterator Traits 8. STL Function Objects 8.1 The Concept of Function Objects 8.2 Predefined Function Objects 8.3 Supplementary Composing Function Objects 9. STL Algorithms 9.1 Algorithm Header Files 9.2 Algorithm Overview 9.3 Auxiliary Functions 9.4 The for_each() Algorithm 9.5 Nonmodifying Algorithms 9.6 Modifying Algorithms 9.7 Removing Algorithms 9.8 Mutating Algorithms 9.9 Sorting Algorithms 9.10 Sorted Range Algorithms 9.11 Numeric Algorithms 10. Special Containers 10.1 Stacks 10.2 Queues 10.3 Priority Queues 10.4 Bitsets 11. Strings 11.1 Motivation 11.2 Description of the String Classes 11.3 String Class in Detail 12. Numerics 12.1 Complex Numbers 12.2 Valarrays 12.3 Global Numeric Functions 13. Input/Output Using Stream Classes 13. Input/Output Using Stream Classes 13.1 Common Background of I/O Streams 13.2 Fundamental Stream Classes and Objects

The C++ Standard Library 5 13.3 Standard Stream Operators << and >> 13.4 State of Streams 13.5 Standard Input/Output Functions 13.6 Manipulators 13.7 Formatting 13.8 Internationalization 13.9 File Access 13.10 Connecting Input and Output Streams 13.11 Stream Classes for Strings 13.12 Input/Output Operators for User-Defined Types 13.13 The Stream Buffer Classes 13.14 Performance Issues 14. Internationalization 14.1 Different Character Encodings 14.2 The Concept of Locales 14.3 Locales in Detail 14.4 Facets in Detail 15. Allocators 15.1 Using Allocators as an Application Programmer 15.2 Using Allocators as a Library Programmer 15.3 The Default Allocator 15.4 A User-Defined Allocator 15.5 Allocators in Detail 15.6 Utilities for Uninitialized Memory in Detail Internet Resources Where You Can Get the Standard Internet Addresses/URLs Bibliography

The C++ Standard Library 6 Preface In the beginning, I only planned to write a small German book (400 pages or so) about the C++ standard library. That was in 1993. Now, in 1999 you see the result — an English book with more than 800 pages of facts, figures, and examples. My goal is to describe the C++ standard library so that all (or almost all) your programming questions are answered before you think of the question. Note, however, that this is not a complete description of all aspects of the C++ standard library. Instead, I present the most important topics necessary for learning and programming in C++ by using its standard library. Each topic is described based on the general concepts; this discussion then leads to the specific details needed to support every-day programming tasks. Specific code examples are provided to help you understand the concepts and the details. That's it — in a nutshell. I hope you get as much pleasure from reading this book as I did from writing it. Enjoy!

The C++ Standard Library 7 Acknowledgments This book presents ideas, concepts, solutions, and examples from many sources. In a way it does not seem fair that my name is the only name on the cover. Thus, I'd like to thank all the people and companies who helped and supported me during the past few years. First, I'd like to thank Dietmar Kühl. Dietmar is an expert on C++, especially on input/output streams and internationalization (he implemented an I/O stream library just for fun). He not only translated major parts of this book from German to English, he also wrote sections of this book using his expertise. In addition, he provided me with invaluable feedback over the years. Second, I'd like to thank all the reviewers and everyone else who gave me their opinion. These people endow the book with a quality it would never have had without their input. (Because the list is extensive, please fogive me for any oversight.) The reviewers for the English version of this book included Chuck Allison, Greg Comeau, James A. Crotinger, Gabriel Dos Reis, Alan Ezust, Nathan Meyers, Werner Mossner, Todd Veldhuizen, Chichiang Wan, Judy Ward, and Thomas Wikehult. The German reviewers included Ralf Boecker, Dirk Herrmann, Dietmar Kühl, Edda Lörke, Herbert Scheubner, Dominik Strasser, and Martin Weitzel. Additional input was provided by Matt Austern, Valentin Bonnard, Greg Colvin, Beman Dawes, Bill Gibbons, Lois Goldthwaite, Andrew Koenig, Steve Rumbsby, Bjarne Stroustrup, and David Vandevoorde. Special thanks to Dave Abrahams, Janet Cocker, Catherine Ohala, and Maureen Willard who reviewed and edited the whole book very carefully. Their feedback was an incredible contribution to the quality of this book. A special thanks goes to my "personal living dictionary" — Herb Sutter — the author of the famous "Guru of the Week" (a regular series of C++ programming problems that is published on the comp.std.C++.moderated Internet newsgroup). I'd also like to thank all the people and companies who gave me the opportunity to test my examples on different platforms with different compilers. Many thanks to Steve Adamczyk, Mike Anderson, and John Spicer from EDG for their great compiler and their support. It was a big help during the standardization process and the writing of this book. Many thanks to P. J. Plauger and Dinkumware, Ltd, for their early standard-conforming implementation of the C++ standard library. Many thanks to Andreas Hommel and Metrowerks for an evaluative version of their Code Warrior Programming Environment. Many thanks to all the developers of the free GNU and egcs compilers. Many thanks to Microsoft for an evaluative version of Visual C++. Many thanks to Roland Hartinger from Siemens Nixdorf Informations Systems AG for a test version of their C++ compiler. Many thanks to Topjects GmbH for an evaluative version of the ObjectSpace library implementation. Many thanks to everyone from Addison Wesley Longman who worked with me. Among others this includes Janet Cocker, Mike Hendrickson, Debbie Lafferty, Marina Lang, Chanda Leary, Catherine Ohala, Marty Rabinowitz, Susanne Spitzer, and Maureen Willard. It was fun. In addition, I'd like to thank the people at BREDEX GmbH and all the people in the C++ community, particularly those involved with the standardization process, for their support and patience (sometimes I ask really silly questions). Last but not least, many thanks and kisses for my family: Ulli, Lucas, Anica, and Frederic. I definitely did not have enough time for them due to the writing of this book. Have fun and be human!

The C++ Standard Library 8 Chapter 1. About this Book 1.1 Why this Book Soon after its introduction, C++ became a de facto standard in object-oriented programming. This led to the goal of standardization. Only by having a standard, could programs be written that would run on different platforms — from PCs to mainframes. Furthermore, a standard library would enable programmers to use general components and a higher level of abstraction without losing portability, rather than having to develop all code from scratch. The standardization process was started in 1989 by an international ANSI/ISO committee. It developed the standard based on Bjarne Stroustrup's books The C++ Programming Language and The Annotated C++ Reference Manual. After the standard was completed in 1997, several formal motions by different countries made it an international ISO and ANSI standard in 1998. The standardization process included the development of a C++ standard library. The library extends the core language to provide some general components. By using C++'s ability to program new abstract and generic types, the library provides a set of common classes and interfaces. This gives programmers a higher level of abstraction. The library provides the ability to use • String types • Different data structures (such as dynamic arrays, linked lists, and binary trees) • Different algorithms (such as different sorting algorithms) • Numeric classes • Input/output (I/O) classes • Classes for internationalization support All of these are supported by a fairly simple programming interface. These components are very important for many programs. These days, data processing often means inputting, computing, processing, and outputting large amounts of data, which are often strings. The library is not self-explanatory. To use these components and to benefit from their power, you need a good introduction that explains the concepts and the important details instead of simply listing the classes and their functions. This book is written exactly for that purpose. First, it introduces the library and all of its components from a conceptional point of view. Next, it describes the details needed for practical programming. Examples are included to demonstrate the exact usage of the components. Thus, this book is a detailed introduction to the C++ library for both the beginner and the practical programmer. Armed with the data provided herein, you should be able to take full advantage of the C++ standard library. Caveat: I don't promise that everything described is easy and self-explanatory. The library provides a lot of flexibility, but flexibility for nontrivial purposes has a price. Beware that the library has traps and pitfalls, which I point out when we encounter them and suggest ways of avoiding them. 1.2 What You Should Know Before Reading this Book To get the most from this book you should already know C++. (The book describes the standard components of C++, but not the language itself.) You should be familiar with the concepts of classes, inheritance, templates, and exception handling. However, you don't have to know all of the minor details about the language. The important details are described in the book (the minor

The C++ Standard Library 9 details about the language are more important for people who want to implement the library rather than use it). Note that the language has changed during the standardization process, so your knowledge might not be up to date. Section 2.2, provides a brief overview and introduction of the latest language features that are important for using the library. You should read this section if you are not sure whether you know all the new features of C++ (such as the keyword typename and the concept of namespaces). 1.3 Style and Structure of the Book The C++ standard library provides different components that are somewhat but not totally independent of each other, so there is no easy way to describe each part without mentioning others. I considered several different approaches for presenting the contents of this book. One was on the order of the C++ standard. However, this is not the best way to explain the components of the C++ standard library from scratch. Another was to start with an overview of all components followed by chapters that provided more details. Alternatively, I could have sorted the components, trying to find an order that had a minimum of cross-references to other sections. My solution was to use a mixture of all three approaches. I start with a brief introduction of the general concepts and the utilities that are used by the library. Then, I describe all the components, each in one or more chapters. The first component is the standard template library (STL). There is no doubt that the STL is the most powerful, most complex, and most exciting part of the library. Its design influences other components heavily. Then I describe the more self- explanatory components, such as special containers, strings, and numeric classes. The next component discussed is one you probably know and use already: the IOStream library. It is followed by a discussion of internationalization, which had some influence on the IOStream library. Each component description begins with the component's purpose, design, and some examples. Next, a detailed description follows that begins with different ways to use the component, as well as any traps and pitfalls associated with it. The description usually ends with a reference section, in which you can find the exact signature and definition of a component's classes and its functions. The following is a description of the book's contents. The first four chapters introduce this book and the C++ standard library in general: • Chapter 1: About this Book This chapter (which you are reading right now) introduces the book's subject and describes its contents. • Chapter 2: Introduction to C++ and the Standard Library This chapter provides a brief overview of the history of the C++ standard library and the context of its standardization. It also contains some general hints regarding the technical background for this book and the library, such as new language features and the concept of complexity. • Chapter 3: General Concepts This chapter describes the fundamental concepts of the library that you need to understand to work with all the components. In particular, it introduces the namespace std, the format of header files, and the general support of error and exception handling. • Chapter 4: Utilities

The C++ Standard Library 10 This chapter describes several small utilities provided for the user of the library and for the library itself. In particular, it describes auxiliary functions such as max(), min(), and swap(), types pair and auto_ptr, as well as numeric_limits, which provide more information about implementation-specific details of numeric data types. Chapters 5 through 9 describe all aspects of the STL: • Chapter 5: The Standard Template Library This chapter presents a detailed introduction to the concept of the STL, which provides container classes and algorithms that are used to process collections of data. It explains step-by-step the concept, the problems, and the special programming techniques of the STL, as well as the roles of its parts. • Chapter 6: STL Containers This chapter explains the concepts and describes the abilities of the STL's container classes. First it describes the differences between vectors, deques, lists, sets, and maps, then their common abilities, and all with typical examples. Lastly it lists and describes all container functions in form of a handy reference. • Chapter 7: STL Iterators This chapter deals in detail with the STL's iterator classes. In particular, it explains the different iterator categories, the auxiliary functions for iterators, and the iterator adapters, such as stream iterators, reverse iterators, and insert iterators. • Chapter 8: STL Function Objects This chapter details the STL's function object classes. • Chapter 9: STL Algorithms This chapter lists and describes the STL's algorithms. After a brief introduction and comparison of the algorithms, each algorithm is described in detail followed by one or more example programs. Chapters 10 through 12 describe "simple" individual standard classes: • Chapter 10: Special Containers This chapter describes the different special container classes of the C++ standard library. It covers the container adapters for queues and stacks, as well as the class bitset, which manages a bitfield with an arbitrary number of bits or flags. • Chapter 11: Strings This chapter describes the string types of the C++ standard library (yes, there are more than one). The standard provides strings as kind of "self-explanatory" fundamental data types with the ability to use different types of characters.

The C++ Standard Library 11 • Chapter 12: Numerics This chapter describes the numeric components of the C++ standard library. In particular, it covers types for complex numbers and classes for the processing of arrays of numeric values (the latter may be used for matrices, vectors, and equations). Chapters 13 and 14 deal with I/O and internationalization (two closely related subjects): • Chapter 13: Input/Output Using Stream Classes This chapter covers the I/O component of C++. This component is the standardized form of the commonly known IOStream library. The chapter also describes details that may be important to programmers but are typically not so well known. For example, it describes the correct way to define and integrate special I/O channels, which are often implemented incorrectly in practice. • Chapter 14: Internationalization This chapter covers the concepts and classes for the internationalization of programs. In particular, it describes the handling of different character sets, as well as the use of different formats for such values as floating-point numbers and dates. The rest of the book contains: • Chapter 15: Allocators This chapter describes the concept of different memory models in the C++ standard library. • An appendix with o Internet Resources o Bibliography o Index 1.4 How to Read this Book This book is a mix of introductory user's guide and structured reference manual regarding the C++ standard library. The individual components of the C++ standard library are independent of each other, to some extent, so after reading Chapters 2 through 4 you could read the chapters that discuss the individual components in any order. Bear in mind, that Chapter 5 through Chapter 9 all describe the same component. To understand the other STL chapters, you should start with the introduction to the STL in Chapter 5. If you are a C++ programmer who wants to know, in general, the concepts and all parts of the library, you could simply read the book from the beginning to the end. However, you should skip the reference sections. To program with certain components of the C++ standard library, the best way to find something is to use the index. I have tried to make the index very comprehensive to save you time when you are looking for something. In my experience, the best way to learn something new is to look at examples. Therefore, you'll find a lot of examples throughout the book. They may be a few lines of code or complete programs. In the latter case, you'll find the name of the file containing the program as the first

The C++ Standard Library 12 comment line. You can find the files on the Internet at my Web site at http://www.josuttis.com/libbook/. 1.5 State of the Art While I was writing this book, the C++ standard was completed. Please bear in mind that some compilers might not yet confirm to it. This will most likely change in the near future. As a consequence, you might discover that not all things covered in this book work as described on your system, and you may have to change example programs to fit your specific environment. I can compile almost all example programs with version 2.8 or higher of the EGCS compiler, which is free for almost all platforms and available on the Internet (see http://egcs.cygnus.com/) and on several software CDs. 1.6 Example Code and Additional Information You can access all example programs and acquire more informations about this book and the C++ standard library from my Web site at http://www.josuttis.com/libbook/. Also, you can find a lot of additional information about this topic on the Internet. See Internet Resources for details. 1.7 Feedback I welcome your feedback (good and bad) on this book. I tried to prepare it carefully; however, I'm human, and at some time I have to stop writing and tweaking. So, you may find some errors, inconsistencies, or subjects that could be described better. Your feedback will give me the chance to improve later editions. The best way to reach me is by Email: libbook@josuttis.com You can also reach me by phone, fax, or "snail" mail: Nicolai M. Josuttis Berggarten 9 D-38108 Braunschweig Germany Phone: +49 5309 5747 Fax: +49 5309 5774 Many thanks.

The C++ Standard Library 13 Chapter 2. Introduction to C++ and the Standard Library 2.1 History The standardization of C++ was started in 1989 and finished at the end of 1997, although some formal motions delayed the final publication until September 1998. The result was a reference manual with approximately 750 pages, published by the International Standards Organization (ISO). The standard has the title "Information Technology — Programming Languages — C++." Its document number is ISO/IEC 14882-1998, and it is distributed by the national bodies of the ISO, such as the ANSI in the United States.[1] [1] At the time this book was written, you could get the C++ standard at the ANSI Electronics Standard Store for $ 18.00 (US; see http://www.ansi.org/). The standard was an important milestone for C++. Because it defines the exact contents and behavior of C++, it makes it easier to teach C++, to use C++ in applications, and to port C++ programs to different platforms. It also gives users greater freedom of choice regarding different C++ implementations. Its stability and portability help library providers and tool providers as well as implementers. Thus, the standard helps C++ application developers build better applications faster, and maintain them with less cost and effort. Part of the standard is a standard library. This library provides core components for I/O, strings, containers (data structures), algorithms (such as sort, search, and merge), support for numeric computation, and (as could be expected from an international standard) support for internationalization (such as different character sets). You may wonder why the standardization process took almost 10 years, and if you know some details about the standard you might wonder why after all this time it is still not perfect. Ten years, in fact, was not enough time! Although, according to the history and the context of the standardization process, a lot was accomplished. The result is usable in practice, but it is not perfect (nothing ever is). The standard is not the result of a company with a big budget and a lot of time. Standards organizations pay nothing or almost nothing to the people who work on developing standards. So, if a participant doesn't work for a company that has a special interest in the standard, the work is done for fun. Thank goodness there were a lot of dedicated people who had the time and the money to do just that. The C++ standard was not developed from scratch. It was based on the language as described by Bjarne Stroustrup, the creator of C++. The standard library, however, was not based on a book or on an existing library. Instead, different, existing classes were integrated.[2] Thus, the result is not very homogeneous. You will find different design principles for different components. A good example is the difference between the string class and the STL, which is a framework for data structures and algorithms: [2] You may wonder why the standardization process did not design a new library from scratch. The major purpose of standardization is not to invent or to develop something; it is to harmonize an existing practice. String classes are designed as a safe and convenient component. Thus, they provide an almost self-explanatory interface and check for many errors in the interface.

The C++ Standard Library 14 The STL was designed to combine different data structures with different algorithms while achieving the best performance. Thus, the STL is not very convenient and it is not required to check for many logical errors. To benefit from the powerful framework and great performance of the STL, you must know the concepts and apply them carefully. Both of these components are part of the same library. They were harmonized a bit, but they still follow their individual, fundamental design philosophies. One component of the library existed as a de facto standard before standardization began: the IOStream library. Developed in 1984, it was reimplemented and partially redesigned in 1989. Because many programs were using it already, the general concept of the IOStream library was not changed, thus keeping it backward compatible. In general, the whole standard (language and library) is the result of a lot of discussions and influence from hundreds of people all over the world. For example, the Japanese came up with important support for internationalization. Of course, mistakes were made, minds were changed, and people had different opinions. Then, in 1994, when people thought the standard was close to being finished, the STL was incorporated, which changed the whole library radically. However, to get finished, the thinking about major extensions was eventually stopped, regardless of how useful the extension would be. Thus, hash tables are not part of the standard, although they should be a part of the STL as a common data structure. The current standard is not the end of the road. There will be fixes of bugs and inconsistencies, and there likely will be a next version of the standard in five years or so. However for the next few years, C++ programmers have a standard and the chance to write powerful code that is portable to very different platforms. 2.2 New Language Features The core language and the library of C++ were standardized in parallel. In this way, the library could benefit from improvements in the language and the language could benefit from experiences of library implementation. In fact, during the standardization process the library often used special language features that were not yet available. C++ is not the same language it was five years ago. If you didn't follow its evolution, you may be surprised with the new language features used by the library. This section gives you a brief overview of those new features. For details, refer to books on the language in question. While I was writing this book (in 1998), not all compilers were able to provide all of the new language features. I hope (and expect) that this will change very soon (most compiler vendors were part of the standardization process). Thus, you may be restricted in your use of the library. Portable implementations of the library typically consider whether features are present in the environment they use (they usually have some test programs to check which language features are present, and then set preprocessor directives according to the result of the check). I'll mention any restrictions that are typical and important throughout the book by using footnotes. The following subsections describe the most important new language features that are relevant for the C++ standard library. 2.2.1 Templates Almost all parts of the library are written as templates. Without template support, you can't use the standard library. Moreover, the library needed new special template features, which I introduce after a short overview of templates.

The C++ Standard Library 15 Templates are functions or classes that are written for one or more types not yet specified. When you use a template, you pass the types as arguments, explicitly or implicitly. The following is a typical example — a function that returns the maximum of two values: template <class T> inline const T& max (const T& a, const T& b) { // if a <b then use b else use a return a < b ? b : a; } Here, the first line defines T as an arbitrary data type that is specified by the caller when the caller calls the function. You can use any identifier as a parameter name, but using T is very common, if not a de facto convention. The type is classified by class, although it does not have to be a class. You can use any data type as long as it provides the operations that the template uses.[3] [3] class was used here to avoid the introduction of a new keyword when templates were introduced. However, now there is a new keyword, typename, that you can also use here (see page 11). Following the same principle, you can "parameterize" classes on arbitrary types. This is useful for container classes. You can implement the container operations for an arbitrary element type. The C++ standard library provides many template container classes (for example, see Chapter 6 or Chapter 10). It also uses template classes for many other reasons. For example, the string classes are parameterized on the type of the characters and the properties of the character set (see Chapter 11). A template is not compiled once to generate code usable for any type; instead, it is compiled for each type or combination of types for which it is used. This leads to an important problem in the handling of templates in practice: You must have the implementation of a template function available when you call it, so that you can compile the function for your specific type. Therefore, the only portable way of using templates at the moment is to implement them in header files by using inline functions.[4] [4] To avoid the problem of templates having to be present in header files, the standard introduced a template compilation model with the keyword export. However, I have not seen it implemented yet. The full functionality of the C++ standard library requires not only the support of templates in general, but also many new standardized template features, including those discussed in the following paragraphs. Nontype Template Parameters In addition to type parameters, it is also possible to use nontype parameters. A nontype parameter is then considered as part of the type. For example, for the standard class bitset<> (class bitset<> is introduced in Section 10.4,) you can pass the number of bits as the template argument. The following statements define two bitfields, one with 32 bits and one with 50 bits: bitset<32> fIags32; // bitset with 32 bits bitset<50> flags50; // bitset with 50 bits These bitsets have different types because they use different template arguments. Thus, you can't assign or compare them (except if a corresponding type conversion is provided). Default Template Parameters

The C++ Standard Library 16 Templates classes may have default arguments. For example, the following declaration allows one to declare objects of class MyClass with one or two template arguments[5] : [5] Note that you have to put a space between the two ">" characters. ">>" would be parsed as shift operator, which would result in a syntax error. template <class T, class container = vector<T> > class MyClass; If you pass only one argument, the default parameter is used as second argument: MyClass<int> x1; // equivalent to: MyClass<int,vector<int> > Note that default template arguments may be defined in terms of previous arguments. Keyword typename The keyword typename was introduced to specify that the identifier that follows is a type. Consider the following example: template <class T> Class MyClass { typename T::SubType * ptr; ... }; Here, typename is used to clarify that SubType is a type of class T. Thus, ptr is a pointer to the type T::SubType. Without typename, SubType would be considered a static member. Thus T::SubType * ptr would be a multiplication of value SubType of type T with ptr. According to the qualification of SubType being a type, any type that is used in place of T must provide an inner type SubType. For example, the use of type Q as a template argument MyClass<Q> x; is possible only if type Q has an inner type definition such as the following: class Q { typedef int SubType; ... }; In this case, the ptr member of MyClass<Q> would be a pointer to type int. However, the subtype could also be an abstract data type (such as a class): class Q { class SubType; ... }; Note that typename is always necessary to qualify an identifier of a template as being a type, even if an interpretation that is not a type would make no sense. Thus, the general rule in C++ is that any identifier of a template is considered to be a value, except it is qualified by typename. Apart from this, typename can also be used instead of class in a template declaration: template <typename T> class MyClass;

The C++ Standard Library 17 Member Templates Member functions of classes may be templates. However, member templates may not be virtual, nor may they have default parameters. For example: class MyClass { ... template <class T> void f(T); }; Here, MyClass::f declares a set of member functions for parameters of any type. You can pass any argument as long as its type provides all operations used by f(). This feature is often used to support automatic type conversions for members in template classes. For example, in the following definition the argument x of assign() must have exactly the same type as the object it is called for: template <class T> class MyClass { private: T value; public: void assign(const MyClass<T>& x) { // x must have same type as *this value = x.value; } ... }; It would be an error to use different template types for the objects of the assign() operation even if an automatic type conversion from one type to the other is provided: void f() { MyClass<double> d; MyClass<int> i; d.assign(d); //OK d.assign(i); //ERROR: i is MyClass<int> // but MyClass<double> is required } By providing a different template type for the member function, you relax the rule of exact match. The member template function argument may have any template type, then as long as the types are assignable: template <class T> class MyClass<T> { private: T value; public template <class X> // member template void assign(const MyClass<X>& x) {// allows different template types value = x.getValue(); } T getValue() const {

The C++ Standard Library 18 return value; } ... }; void f() { MyClass<double> d; MyClass<int> i; d.assign(d); // OK d.assign(i); // OK (int is assignable to double) } Note that the argument x of assign() now differs from the type of *this. Thus, you can't access private and protected members of MyClass<> directly. Instead, you have to use something like getValue() in this example. A special form of a member template is a template constructor. Template constructors are usually provided to enable implicit type conversions when objects are copied. Note that a template constructor does not hide the implicit copy constructor. If the type matches exactly, the implicit copy constructor is generated and called. For example: template <class T> class MyClass<T> { public: //copy constructor with implicit type conversion //- does not hide implicit copy constructor template <class U> MyClass(const MyClass<U>& x); ... }; void f() { MyClass<double> xd; ... MyClass<double> xd2(xd); // calls built-in copy constructor MyClass<int> xi (xd); // calls template constructor ... } Here, the type of xd2 is the same as the type of xd, so it is initialized via the built-in copy constructor. The type of xi differs from the type of xd, so it is initialized by using the template constructor. Thus, if you write a template constructor, don't forget to provide a copy constructor, if the default copy constructor does not fit your needs. See Section 4.1, for another example of member templates. Nested Template Classes Nested classes may also be templates: template <class T> class MyClass { ...

The C++ Standard Library 19 template <class T2> class NestedClass; ... }; 2.2.2 Explicit Initialization for Fundamental Types If you use the syntax of an explicit constructor call without arguments, fundamental types are initialized with zero: int i1; // undefined value int i2 = int(); // initialized with zero This feature is provided to enable you to write template code that ensures that values of any type have a certain default value. For example, in the following function the initialization guarantees that x is initialized with zero for fundamental types: template <class T> void f() { T x = T(); ... } 2.2.3 Exception Handling The C++ standard library uses exception handling. Using this feature, you can handle exceptions without "polluting" your function interfaces: arguments and return values. If you encounter an unexpected situation, you can stop the usual data processing by "throwing an exception": class Error; void f() { ... if (excetion-condition) { throw Error(); // create object of class Error and throw it as exception } ... } The throw statement starts a process called stack unwinding; that is, any block or function is left as if there was a return statement. However, the program does not jump anywhere. For all local objects that are declared in the blocks that the program leaves due to the exception their destructors are called. Stack unwinding continues until main() is left, which ends the program, or until a catch clause "catches" and handles the exception: int main() { try { ... f(); ... } catch (const Error&) {