Book home pages

Java Methods
Math and Python
Be Prepared Comp Sci
Be Prepared Calculus
800 Questions in Calculus

  Download files

Students
Teachers

  Math and Python

Test Package
Getting started/help
Papers/projects
What teachers say...

  Java Methods

Test Package
PowerPoints
Syllabi
Getting started/help
Papers/projects
Workshops
What teachers say...

  AP^* test prep

Be Prepared for the AP CS Exam
250 MC Questions in Java
Be Prepared for the AP Calculus Exam
800 Questions in Calculus
Calculus Calculator Labs
What teachers say...

  C++

100 MC Questions in C++

  Ordering Information

Shop online
Print order form
Contact us

Skylight home

apvector, apstring, etc:

Frequently Asked Questions

Description of the C++ classes defined for use in the Advanced Placement^* Computer Science courses and inclusion of the link to the College Board web page does not constitute endorsement of the other material in this document by the College Board, Educational Testing Service, or the AP^* Computer Science Development Committee. The description of the C++ classes defined for use in the AP Computer Science courses included in this document were accurate as of 9/01/98. Revisions to the classes may have been made since that time.

• What are the AP classes?
• Are the AP classes part of the C++ standard?
• Why do we need the AP classes?
• Where can I get the AP classes?
• Trouble downloading the AP classes?
• Why do we have two files, something.h and something.cpp?
• What do I have to do to use the apvector class?
• What do I have to do to use the apstring class?
• What is the difference between the "capacity" and the "length" of a vector?
• Error messages:
  • Unable to open include file 'APVECTOR.H'
  • APVECTOR.CPP ...: Templates must be classes or functions
  • Variable bool initialized more than once
  • 'bool' redefinition; different basic type
  • Linker error: Undefined symbol apstring::~apstring() in module ...
  • Illegal vector index: 0 max index = -1
  • Illegal structure operation

How to set up include paths and projects: specific instructions for your compiler.

Please e-mail other questions to support@skylit.com

Q: What are the AP classes?

A: In C++ and a few other programming languages, a class is an entity that combines data elements and functions that operate on them. A class may implement a particular type of an object or a functional module. The College Board's Ad Hoc Development Committee on C++ defined five classes, apvector, apmatrix, apstring, apstack and apqueue to be used in the Advanced Placement program and exams. The A-level exam requires only the apvector, apmatrix and apstring classes; the AB-level exam requires all five classes.

The apvector, apmatrix and apstring classes implement "safe" one- and two-dimensional arrays and character strings, respectively. The apstack and apqueue classes implement the stack and queue ADTs.

Q: Are the AP classes part of the C++ standard?

A: No, they are not included in standard C++. The AP classes are separate modules, developed by the College Board, that provide additional functionality. The apstring class is a subset of a typical string class that handles character strings -- different versions of such classes are included in different class libraries. The other four classes are loosely based on similar classes from the STL -- Standard Template Library.

Q: Why do we need the AP classes?

A: The question whether the AP classes are helpful in the AP curriculum has been discussed at length on the ap-compsci listserve. The Development Committee felt that standard arrays in C++ -- the Committee refers to them as "C-style arrays" -- are "unsafe." Indeed, in C and C++ an array is treated pretty much as a pointer to its first element and neither the compiler, nor the executable program checks that a subscript refers to an element inside the array. Therefore, a bug in the program can easily cause overwriting a memory location outside an array, and such program may hang. The "C-style" null-terminated strings are unsafe, too. The library functions that handle them, such as strcpy(...) for copying and strcat() for concatenation, assume that the array that receives the result has enough space to accommodate the resulting string. If the programmer is not careful, these functions may overwrite vital segments of the program's or system's memory.

The apstack and apqueue classes provide implementations of stacks and queues. The Committee defined these classes so that they could be shared by everyone involved in the AP program and used in the AP exam questions and solutions.

Q: Where can I get the AP classes?

A: You can download the AP classes from the College Board's AP CS web page. A slightly edited version of the AP classes is also included on the C++ for You++ Teacher's Disk.

Q: I can display apvector.h, apvector.cpp etc. in my web browser, but how do I "download" them?

A: Once you display a file in your browser, click on the File menu and choose Save As.... Navigate to the folder on your disk where you want to save the file, and click OK.

Q: Why do we have two files, a .h and a .cpp file (e.g. apvector.h and apvector.cpp), for each of the classes?

A: This is an important question. Strictly speaking we need two separate files, apstring.h and apstring.cpp, only for the apstring class. A .h file only would suffice for the other four classes. The best thing to do is to pretend that apvector.cpp, apmatrix.cpp, apstack.cpp, and apqueue.cpp do not exist, at least until you are ready to find out how these classes are implemented.

A full explanation of this requires some understanding of C++ modules, header files, and templates (See, for example, C++ for You++, Chapter 14). A C++ module (e.g. a class) is usually implemented as a header file, something.h, and a source file something.cpp. The source file contains the actual code for the functions. The header file contains only structure/class definitions and function prototypes and serves as an interface between the module and a client program that uses it. something.h is included into both something.cpp and into the client program using the #include directive. A source file can be compiled separately, then linked with other modules to form one executable program.

That is how apstring.h and apstring.cpp are used.

The situation is different for the other four classes, because they are templated classes, that is classes that work with elements of different data types. Such classes cannot be compiled separately because the compiler needs to see first which data type they are used with. It is customary to place all the definitions and code for such classes into one header file, something.h. The Development Committee decided to split each of these four classes into a "header" (.h) and an "implementation" (.cpp) file. But in this case the .cpp file serves simply as the continuation of the respective .h file. For example, apvector.h has the directive

Q: What do I have to do to use the apvector class?

A: Add the line

Q: I have downloaded the apvector.h and apvector.cpp files, and added #include "apvector.h", but I am getting an error message:

A: There are three ways of doing it: an "ugly" way, an "expedient" way, and a "correct" way.

The "ugly" way is to include the complete pathname for the folder where apvector.h resides. For example:

It works, but no self-respecting programmer would do that (because you would need to edit and rebuild all your programs if the location of your AP classes changes).

The "expedient" way is to copy all the AP classes into your compiler's "include" folder, that is the folder that contains iostream.h and other standard header files. For example, if you have Borland C++ 5.0 installed in c:\bc5, you can copy apvector.h, apvector.cpp, etc. into the c:\bc5\include folder. This works, but you "contaminate" your compiler installation with "foreign matter." If you have to reinstall your compiler or install a new version of the compiler, or if the new version of the AP classes arrives, you have to remember to copy them again into the compiler's include folder.

The "correct" way (i.e. the way we recommend) is to put all the AP classes into a separate folder (for example, c:\apclass) and to add that folder to the list of the "include paths" searched by your compiler. Follow the appropriate link for your compiler to learn the specific steps needed for using apvector, apstring, and the other AP classes.

Q: I am getting an error

A: The data type bool was added to the C++ standard fairly recently and older compilers do not understand it. To get around this problem, the Development Committee (and book authors) use a special header file bool.h. It contains the code that redefines the bool type as int and defines the constants false and true:

This file may be included at the top of a module that uses the bool data type:

In particular apstring.h, apstack.h, and apqueue.h have this line.

This directive may be enabled or commented out depending on whether your compiler recognizes the bool type. The problems begin when you use bool.h with compilers that have the built-in bool type (e.g. Visual C++ 6.0, Borland's 5.0, Borland's Builder, CodeWarrior) or forget to include it with compilers that do not understand bool (e.g. Borland's 4.5 or older). For example, the error message above comes from Turbo C++ 4.5 when you try to declare a bool variable but forgot to include or enable bool.h. Another message --

-- comes from VC++ 6.0 if you forget to disable bool.h. We recommend the following method of dealing with bool:

• leave all #include "bool.h" directives active (i.e. do not comment them out) in all programs, in particular, in apstring.h, apstack.h and apqueue.h;
• If your compiler has the built-in bool type, comment out the definitions inside bool.h: // typedef int bool; // const int false = 0; // const int true = 1;

This will make your programs portable between compilers: all you have to do is to use the appropriate (active or inactive) version of bool.h. (In a more recent version of bool.h, the definitions inside bool.h are automatically disabled for Visual C++ because they are placed inside the conditional compilation directive #ifndef _MSC_VER.)

Q: I am getting an error message:

What have I done wrong?

A: You are trying to compile apvector.h or apvector.cpp. As explained above, neither of these files can be compiled separately and should never be added as an item to projects.

Q: What do I have to do to use the apstring class in my program?

A: You need to do two things:

1. Add the line #include "apstring.h"

   at the top of your program.
   
   

2. Set up a project with two items: your program and apstring.cpp. Follow the appropriate link for your compiler to learn the specific steps needed for using the apstring class.

Q: I have included apstring.h into my program but I am still getting a long list of errors:

etc. What is missing?

A: Your program compiled correctly, but it failed at the linking stage because the code for the apstring functions is missing. apstring.h contains only function prototypes for the apstring functions, but not their code.

There are three ways to take care of this problem: the "ugly," the "expedient" and the "correct" way. The "ugly" way is to add all the source code for the apstring class to your program using #include "apstring.cpp" instead of (or in addition to) #include "apstring.h". A more subtle way of achieving the same result is to leave #include "apstring.h" in your program, but add #include "apstring.cpp" at the bottom of apstring.h. This will work for a while -- as long as you use single-module projects -- but it will fail when you start using multiple modules (e.g. the Marine Biology case study). Then the linker will complain that the apstring functions are defined more than once.

The "expedient" way is to add the compiled apstring module to the standard compiler libraries.

The "correct" way (i.e. the one we recommend) is to set up a project that includes your program and apstring.cpp.

Q: What is the difference between the "capacity" and the "length" of a vector?

A: The implementation of the apvector class is such that there is no explicit distinction between the length of the vector (meaning the number of elements currently stored in the vector) and its capacity (meaning the number of elements that can fit in its currently allocated array). These two notions are somewhat confused. For example, the apvector class definition contains the line:

There are two points of view on how vectors should be used. One approach is that the actual length of the vector may be smaller than its capacity. In this approach your program has to maintain the length of the vector in a separate variable. If you pass a vector to a function, you have to also pass its length. For example:

In this approach it makes sense to say that the length() member function returns the capacity of the vector, and therefore it is not very useful. The subscript checking functionality will be limited: the apvector class will check that a subscript falls within the capacity of the vector, but not that it actually refers to a valid element. Another advantage of the apvector class -- the capability of allocating exactly the required amount of space -- will not be always utilized. This approach is closer to the way built-in arrays are used.

The second approach (the one we recommend) is to make sure that the capacity of the vector and the length of the vector are almost always the same. There may be a need to sometimes resize a vector to achieve that (using the resize(...) member function). In this approach, the length() function returns the length of the vector. There is no need to maintain the length of the vector in a separate variable and pass it to functions: it can be obtained inside a function by calling length(). For example:

Past AP exam materials seem to lean toward the second approach, but AP students will be safer if they are familiar with both approaches and understand the difference between them.

Q: I wrote a simple program:

When I run it I get a message: "Illegal vector index: 0 max index = -1" and the program is aborted. What is the meaning of this message?

A: The declaration

constructs an empty vector because you have not specified its size. The run- time message is a service of the apvector class: it indicates that a subscript (index) is out of bounds. The apvector class calculates the maximum allowed index as size - 1, and in this case the size is 0, so it reports "max index = -1."

Q: I am trying to use the apvector class in a simple program:

The compiler gives me an error message "Illegal structure operation" for line 11. Why?

A: You used brackets instead of parentheses in the declaration

It should be

Brackets are used with built-in arrays. Your declaration, while syntactically valid, declares an array of three empty vectors instead of one vector with three elements.

The compiler detects a problem only when you try to add a[0] + a[1] + a[2]: each addend here is an apvector, and there is no valid addition operator for apvector variables.

Skylight Publishing Home Page