Introduction To C

Course Notes

Software Engineering
Technician /
Technology
Version 2014.1
 Preface

Many hours were spent writing and making extensive changes to these course notes.
This document was originally created by Ignac Kolenko and then expanded and revised
by Carlo Sgro.
If you find any errors, inconsistencies, outdated information, formatting problems, or
simply things that don't make sense, please e-mail with details to
[email protected].
Also, many references are made to a textbook called Simple Program Design by Lesley
Anne Robertson (Course Technology, Cambridge, MA, 2000, ISBN: 0-619-01590-X). If
you do not have much programming background in other languages, you should purchase
that book.
 Simple Table of Contents
Chapter 1. C Programming Overview ...................................................................................... 1
Chapter 2. Source Code Structure ............................................................................................ 6
Chapter 3. Variables and Data Types ..................................................................................... 19
Chapter 4. Operators ............................................................................................................... 33
Chapter 5. Flow of Control ..................................................................................................... 40
Chapter 6. Pseudocode ........................................................................................................... 58
Chapter 7. Designing for Flow Of Control ............................................................................. 61
Chapter 8. Writing Custom Functions .................................................................................... 66
Chapter 9. Rules about Variables ........................................................................................... 81
Chapter 10. Other Control Statements ...................................................................................... 87
Chapter 11. Standard Library Functions................................................................................... 97
Chapter 12. Simple Pointers ................................................................................................... 108
Chapter 13. Arrays .................................................................................................................. 114
Chapter 14. Strings ................................................................................................................. 123
Chapter 15. More Pointers ...................................................................................................... 129
Chapter 16. The C Preprocessor ............................................................................................. 143
Chapter 17. Structures ............................................................................................................ 148
Chapter 18. Unions, Enumerations and Type Definitions ...................................................... 159
Chapter 19. Multi-file Projects ............................................................................................... 167
Chapter 20. Dynamic Memory ............................................................................................... 174
Chapter 21. File Input and Output .......................................................................................... 179
Chapter 22. Binary File Input and Output .............................................................................. 188
Chapter 23. Parameters To main() .......................................................................................... 194
Chapter 24. Operating System Functionality ......................................................................... 196
Chapter 25. Recursion ............................................................................................................ 202
Chapter 26. Advanced Pointer Usage ..................................................................................... 204
Chapter 27. Now what? .......................................................................................................... 215
Appendix A. Setting Up A Visual Studio Project .................................................................... 219
Appendix B. Counting Systems ............................................................................................... 222
Appendix C. Recap of Course Requirements Mentioned In This Document .......................... 228
 Expanded Table of Contents
Chapter 1. C Programming Overview ...................................................................................... 1
1.1. The History of C ............................................................................................................... 1
1.2. Why C? ............................................................................................................................. 1
1.3. No, Really … Why C and not C++ or Java or C#? .......................................................... 2
1.4. Compilers ......................................................................................................................... 3
1.5. Design Cycle .................................................................................................................... 4
1.6. Common Beginner Errors ................................................................................................ 4
1.7. Summary .......................................................................................................................... 4
Chapter 2. Source Code Structure ............................................................................................ 6
2.1. A Typical C Program ....................................................................................................... 6
2.2. Program Structure ............................................................................................................ 6
2.3. Functions .......................................................................................................................... 9
2.4. Another Sample Program ............................................................................................... 15
2.5. Common Beginner Errors .............................................................................................. 17
2.6. Summary ........................................................................................................................ 17
Chapter 3. Variables and Data Types ..................................................................................... 19
3.1. Data Objects in your CPU .............................................................................................. 19
3.2. Signed vs. Unsigned ....................................................................................................... 20
3.3. Floating Point ................................................................................................................. 20
3.4. Compiler's View of Data Objects ................................................................................... 21
3.5. Notation .......................................................................................................................... 22
3.6. Initialized Variables ....................................................................................................... 23
3.7. Global Variables ............................................................................................................. 23
3.8. Local Variables .............................................................................................................. 24
3.9. Constants in C ................................................................................................................ 26
3.10. Character Constants .................................................................................................... 27
3.11. How to Output Variables ............................................................................................ 29
3.12. Common Beginner Errors ........................................................................................... 32
3.13. Summary..................................................................................................................... 32
Chapter 4. Operators ............................................................................................................... 33
4.1. What is an Operator? ...................................................................................................... 33
4.2. C Operators .................................................................................................................... 33
 4.3. Binary Operators ............................................................................................................ 34
4.4. Unary Operators ............................................................................................................. 35
4.5. Operator Examples ......................................................................................................... 35
4.6. Precedence Rules............................................................................................................ 36
4.7. Assignment Operators .................................................................................................... 37
4.8. Increment and Decrement .............................................................................................. 37
4.9. Putting it Together.......................................................................................................... 38
4.10. Common Beginner Errors ........................................................................................... 39
4.11. Summary..................................................................................................................... 39
Chapter 5. Flow of Control ..................................................................................................... 40
5.1. Introduction .................................................................................................................... 40
5.2. if Statement .................................................................................................................... 40
5.3. Logical Operators ........................................................................................................... 42
5.4. Truth Table ..................................................................................................................... 43
5.5. Putting It Together ......................................................................................................... 45
5.6. Chained if Statements..................................................................................................... 48
5.7. Compound Statements.................................................................................................... 50
5.8. Nesting if Statements ..................................................................................................... 51
5.9. Loops .............................................................................................................................. 51
5.10. The while Statement ................................................................................................... 52
5.11. Pitfalls of Flow of Control .......................................................................................... 54
5.12. Indentation Revisited .................................................................................................. 56
5.13. Common Beginner Errors ........................................................................................... 56
5.14. Summary..................................................................................................................... 56
Chapter 6. Pseudocode ........................................................................................................... 58
6.1. Sequence......................................................................................................................... 58
6.2. Selection ......................................................................................................................... 59
6.3. Repetition ....................................................................................................................... 59
6.4. Example of Pseudocode ................................................................................................. 59
6.5. Why Pseudocode? .......................................................................................................... 60
6.6. Common Beginner Errors .............................................................................................. 60
6.7. Summary ........................................................................................................................ 60
Chapter 7. Designing for Flow Of Control ............................................................................. 61
7.1. Some Help With Flow of Control .................................................................................. 61
 7.2. More General Flow Of Control Design Principles......................................................... 61
7.3. Some General Flow Of Control Design Examples ........................................................ 61
Chapter 8. Writing Custom Functions .................................................................................... 66
8.1. Review of Functions....................................................................................................... 66
8.2. Why Write a Function? .................................................................................................. 66
8.3. Creating a Function ........................................................................................................ 67
8.4. Function Declaration Notation ....................................................................................... 68
8.5. Returning a Value........................................................................................................... 68
8.6. Parameters ...................................................................................................................... 69
8.7. Function Comments ....................................................................................................... 71
8.8. Prototyping Functions .................................................................................................... 73
8.9. Functions That Do Not Return Any Values ................................................................... 76
8.10. Designing Using Functions ........................................................................................ 76
8.11. Naming Conventions, Revisited ................................................................................. 79
8.12. Common Beginner Errors ........................................................................................... 79
8.13. Summary..................................................................................................................... 79
Chapter 9. Rules about Variables ........................................................................................... 81
9.1. Scope .............................................................................................................................. 81
9.2. Lifetime Rules ................................................................................................................ 83
9.3. Static Variables .............................................................................................................. 83
9.4. The Processor Stack ....................................................................................................... 84
9.5. Memory Layout .............................................................................................................. 85
9.6. Common Beginner Errors .............................................................................................. 86
9.7. Summary ........................................................................................................................ 86
Chapter 10. Other Control Statements ...................................................................................... 87
10.1. do - while Statement ................................................................................................... 87
10.2. The for Statement ....................................................................................................... 89
10.3. Putting It Together ...................................................................................................... 90
10.4. Other Features of Loops ............................................................................................. 91
10.5. The switch Statement.................................................................................................. 92
10.6. The Evil goto Statement ............................................................................................. 95
10.7. Common Beginner Errors ........................................................................................... 95
10.8. Summary..................................................................................................................... 96
Chapter 11. Standard Library Functions................................................................................... 97
 11.1. Mathematical Functions ........................................................................................... 101
11.2. String Handling......................................................................................................... 102
11.3. Character Classification Functions ........................................................................... 104
11.4. Miscellaneous Functions .......................................................................................... 107
11.5. Common Beginner Errors ......................................................................................... 107
11.6. Summary................................................................................................................... 107
Chapter 12. Simple Pointers ................................................................................................... 108
12.1. Pointer Syntax........................................................................................................... 108
12.2. Addressing and Dereferencing ................................................................................. 108
12.3. Pass-by-reference...................................................................................................... 109
12.4. Pass-by-reference Example ...................................................................................... 110
12.5. Another Example Of Pass-By-Reference ................................................................. 111
12.6. Casting ...................................................................................................................... 112
12.7. Common Beginner Errors ......................................................................................... 113
12.8. Summary................................................................................................................... 113
Chapter 13. Arrays .................................................................................................................. 114
13.1. Introduction .............................................................................................................. 114
13.2. Base Address ............................................................................................................ 115
13.3. Array Offset .............................................................................................................. 115
13.4. Array Offset Operator ............................................................................................... 115
13.5. Notation .................................................................................................................... 115
13.6. More About Array Offsets ........................................................................................ 118
13.7. Initialization of Arrays.............................................................................................. 118
13.8. Multi-Dimensional Arrays ........................................................................................ 121
13.9. Common Beginner Errors ......................................................................................... 121
13.10. Summary................................................................................................................... 121
Chapter 14. Strings ................................................................................................................. 123
14.1. C and Character Strings ............................................................................................ 123
14.2. The Innards of Strings .............................................................................................. 125
14.3. String Library Functions ........................................................................................... 126
14.4. Common Beginner Errors ......................................................................................... 128
14.5. Summary................................................................................................................... 128
Chapter 15. More Pointers ...................................................................................................... 129
15.1. Review of Array Concepts ....................................................................................... 129
 15.2. Addresses .................................................................................................................. 129
15.3. Initializing a Pointer Variable................................................................................... 130
15.4. Accessing Memory Using Pointers .......................................................................... 131
15.5. 'Contents of' Operator ............................................................................................... 132
15.6. Putting It Together .................................................................................................... 133
15.7. Scaling Of Pointer Math ........................................................................................... 135
15.8. Pointers and Arrays .................................................................................................. 136
15.9. The scanf() Family.................................................................................................... 138
15.10. Using fgets() and sscanf() for Getting Integer Input ................................................ 140
15.11. Common Beginner Errors ......................................................................................... 141
15.12. Summary................................................................................................................... 142
Chapter 16. The C Preprocessor ............................................................................................. 143
16.1. #define ...................................................................................................................... 143
16.2. Macros ...................................................................................................................... 144
16.3. Conditional Compilation .......................................................................................... 145
16.4. Common Beginner Errors ......................................................................................... 147
16.5. Summary................................................................................................................... 147
Chapter 17. Structures ............................................................................................................ 148
17.1. What is a Structure? .................................................................................................. 148
17.2. Structure On Paper.................................................................................................... 148
17.3. C Structures .............................................................................................................. 149
17.4. Structure Notation..................................................................................................... 150
17.5. Variable Declarations ............................................................................................... 151
17.6. Structure Copy .......................................................................................................... 151
17.7. Field Notation ........................................................................................................... 151
17.8. Aggregate Structures ................................................................................................ 152
17.9. Pointers and Field Notation ...................................................................................... 152
17.10. Pointer Arithmetic .................................................................................................... 153
17.11. Putting It Together .................................................................................................... 153
17.12. Initializing Structures ............................................................................................... 154
17.13. Using Structures from the Standard Library............................................................. 155
17.14. sizeof......................................................................................................................... 156
17.15. Bitfield Structures ..................................................................................................... 157
17.16. Summary................................................................................................................... 158
Chapter 18. Unions, Enumerations and Type Definitions ...................................................... 159
18.1. Unions ....................................................................................................................... 159
18.2. Union Notation ......................................................................................................... 159
18.3. Union Field Access ................................................................................................... 161
18.4. Enumerations ............................................................................................................ 162
18.5. Enumeration Notation............................................................................................... 163
18.6. typedef ...................................................................................................................... 164
18.7. Common Beginner Errors ......................................................................................... 165
18.8. Summary................................................................................................................... 166
Chapter 19. Multi-file Projects ............................................................................................... 167
19.1. Introductory Notes .................................................................................................... 167
19.2. Projects and Workspaces .......................................................................................... 167
19.3. Variable Declarations ............................................................................................... 167
19.4. Prototyping ............................................................................................................... 168
19.5. Common Constants................................................................................................... 169
19.6. Programmer Defined Header Files ........................................................................... 169
19.7. When Will A Compiler Compile A File? ................................................................. 171
19.8. Guidelines for Splitting Files .................................................................................... 172
19.9. Example of Splitting Files ........................................................................................ 172
19.10. Common Beginner Errors ......................................................................................... 173
19.11. Summary................................................................................................................... 173
Chapter 20. Dynamic Memory ............................................................................................... 174
20.1. Static Memory Allocation ........................................................................................ 174
20.2. Dynamic Memory Allocation ................................................................................... 174
20.3. malloc(), calloc() and realloc() ................................................................................. 175
20.4. NULL........................................................................................................................ 176
20.5. free() ......................................................................................................................... 177
20.6. Common Beginner Errors ......................................................................................... 178
20.7. Summary................................................................................................................... 178
Chapter 21. File Input and Output .......................................................................................... 179
21.1. Opening Notes .......................................................................................................... 179
21.2. Stream Based File I/O............................................................................................... 179
21.3. Pre-declared Streams ................................................................................................ 179
21.4. Standard File I/O Functions ...................................................................................... 179
 21.5. How to Open or Create a File ................................................................................... 181
21.6. Filename ................................................................................................................... 181
21.7. Access Types For Opening Files .............................................................................. 182
21.8. Closing an Opened File ............................................................................................ 183
21.9. Putting It Together .................................................................................................... 183
21.10. Common Beginner Errors ......................................................................................... 187
21.11. Summary................................................................................................................... 187
Chapter 22. Binary File Input and Output .............................................................................. 188
22.1. Binary Files............................................................................................................... 188
22.2. Stream Based Binary File Techniques...................................................................... 188
22.3. fread() and fwrite() ................................................................................................... 189
22.4. Putting It Together .................................................................................................... 190
22.5. Other Stream Based Functions ................................................................................. 192
22.6. Common Beginner Errors ......................................................................................... 193
22.7. Summary................................................................................................................... 193
Chapter 23. Parameters To main() .......................................................................................... 194
23.1. Introductory Notes .................................................................................................... 194
23.2. main() Parameter Example ....................................................................................... 194
23.3. Common Beginner Errors ......................................................................................... 195
23.4. Summary................................................................................................................... 195
Chapter 24. Operating System Functionality ......................................................................... 196
24.1. Introductory Notes .................................................................................................... 196
24.2. Listing of Files .......................................................................................................... 196
24.3. Windows 32-bit Search Functions ............................................................................ 197
24.4. Changing Directories ................................................................................................ 198
24.5. Renaming a File ........................................................................................................ 199
24.6. Deleting a File .......................................................................................................... 200
24.7. The system() Call...................................................................................................... 200
24.8. Common Beginner Errors ......................................................................................... 200
24.9. Summary................................................................................................................... 200
Chapter 25. Recursion ............................................................................................................ 202
25.1. Recursion Example ................................................................................................... 202
25.2. Recursion vs. Looping .............................................................................................. 203
25.3. Common Beginner Errors ......................................................................................... 203
 25.4. Summary................................................................................................................... 203
Chapter 26. Advanced Pointer Usage ..................................................................................... 204
26.1. Pointer Review ......................................................................................................... 204
26.2. Pointers to Pointers ................................................................................................... 204
26.3. Pointers and Functions.............................................................................................. 206
26.4. Pointer to a Function Notation.................................................................................. 207
26.5. Putting It Together .................................................................................................... 207
26.6. Arrays of Function Pointers ...................................................................................... 210
26.7. Putting It Together .................................................................................................... 210
26.8. Other Advanced Uses ............................................................................................... 213
26.9. Common Beginner Errors ......................................................................................... 213
26.10. Summary................................................................................................................... 213
Chapter 27. Now what? .......................................................................................................... 215
Appendix A. Setting Up A Visual Studio Project .................................................................... 219
Appendix B. Counting Systems ............................................................................................... 222
Appendix C. Recap of Course Requirements Mentioned In This Document .......................... 228
Chapter 1. C Programming Overview

1.1. The History of C

C is a language developed at Bell Labs in the early 1970's. Its early roots were closely
tied to the creation of the UNIX operating system for mini and mainframe computers, and
because of this history, it has historically been the standard language for the development
of most new operating systems for computers.
C was created to help solve the problem of designing and writing an efficient operating
system for a computer. The operating system of a computer is the software, which allows
the outside world to control the hardware within the computer system, much like
Windows allows a user to manipulate resources such as disks on a PC.
Operating systems are naturally required to be fast, efficient, have complete control over
the hardware resources of the computer, and most of all, be easy to upgrade and maintain
even years after they were written. C is a language that supports those ideals.
Over the years, standards committees have been created to help arrive at a version of the
C programming language that is common to all computing platforms throughout the
world. This version is commonly referred to as the ANSI Standard C language (ANSI is
an international standards committee), and you will find most C compilers are ANSI-
compatible (there are more modern standards that are not as widely used).

1.2. Why C?
C is a language which allows the author of a program to communicate with the machine.
The difference lies in how this language causes a computer to execute a desired function.
C is a compiled language, whereas many other languages (e.g. JavaScript) are interpreted.
"Interpreted" means the computer is running a program which will translate the program
instructions into something the computer understands. The interpreter program will do
this operation each and every time you run your JavaScript program. Consider a program
to generate a conversion table to convert degrees Fahrenheit to degrees Celsius in
JavaScript. If you ran this program one hundred times a day, every instruction in the
JavaScript program to handle this conversion will be translated each and every time the
JavaScript interpreter sees the instruction, even though it may have translated it just
seconds earlier. This act of constantly interpreting instructions cause interpreted
languages to be relatively slow compared to the full speed of today's computers.
C, on the other hand, is a compiled language. A compiled language means that you will
have to run a program, called the compiler, to translate your C language program, which
is normally called source code, into something called an object file. The object file
contains machine-readable code, but not in its final form – in other words, it is not fully
executable yet, as it is missing vital information to allow your code to link to existing
functionality. Another program, called the linker, will combine groups of object files into
an executable file, which is then a program you can execute on your computer, much like

1
 other commands in your operating system. The translation has been done exactly once for
you, and if you want to run your C program a thousand times each day, you only have to
run the executable version of your C program, not the compiler! It should be noted that
when programmers wish to reuse large numbers of object modules from application to
application, often a program called a librarian will be used to manage libraries of object
modules for code reuse. When testing out your executable files, another program, called a
debugger will be used to help test out your application, to see where problems may exist,
and to help trace through areas in your software that require special attention. This then
allows a programmer to return to the editor, make changes, recompile and hence, repeat
the cycle once again. Quite often, all of these tools are combined into an Integrated
Development Environment (IDE). As an example, Visual Studio contains an IDE for
Visual C++. You can do all of your editing, then compile, then link, and finally execute
your program, all within the IDE.
A compiled language such as C creates programs that are fast. The compiler will
translate the C program into instructions which are handled directly by your computer's
CPU, and once the translation has been done for you, the computer can execute these
instructions at full speed each and every time you run your compiled program.
As can be seen from our above discussion, C is a language that can be used to generate
fast, efficient programs which allow you to exploit the full speed of your computer,
something an interpreted language like JavaScript will never allow you to do. This is the
main reason why operating systems such as UNIX are written in a language like C.
A slight drawback of C, which is due to its compiled nature, is that you do not know what
your program will produce as its results until after you run the executable file produced
by the compiler. The compiler is only capable of translating the program into an
executable file. If there are any grammatical mistakes, commonly referred to as syntax
errors, in your source, they are found and reported, and the compilation is aborted to
give you a chance to fix the errors. If no such grammatical mistakes exist, an object file is
created. The linker will in turn generate an executable file on your computer's disk. At
this point, you are finally able to see the fruits of your labour. You may execute your
program, and if something is still incorrect (for example, a logic problem), you must fix
it, re-compile, re-link and re-execute your software once again. If more mistakes exist,
the whole process is repeated. The debugging cycle for a C program can become rather
lengthy!

1.3. No, Really … Why C and not C++ or Java or C#?

C is recognized as an excellent language for learning the fundamentals of programming
and still giving you the skills to pick up other programming languages very easily. It
doesn’t have a lot of the overhead of object-oriented languages and doesn’t have some of
the more advanced features of other languages like automatic garbage collection (you’ll
learn about this when you learn about pointers). If your first language has something like
automatic garbage collection, you will have problems getting used to situations where
you don’t have that feature (e.g. in an embedded system where you’re working with C in

2
 a very small code and data space). Since C is the basis (directly or indirectly) for C++,
Java, and C#, if you can learn C, you can learn anything based on it very easily.

1.4. Compilers
Conestoga College has made arrangements with Microsoft to provide their Visual Studio
development systems for free for you.
MSDN-AA allows you to access Visual Studio (which contains Visual C++), as well as
many other Microsoft products (such as Windows 8, Windows 7, Windows XP,
OneNote, Visio, and Virtual PC). In some cases, these products come in .ISO format,
which requires that you burn it onto a blank CD or DVD (or mount the .ISO using a
third-party utility). In other cases, a proprietary downloader will be used to download the
software.
Download of Visual Studio Professional is rather large (over 2 GB) and installation
requires over 3 GB of hard disk space. Typically, newer versions of Visual Studio (e.g.
2012, 2013) end up being larger. If the space requirements are too much for your
computer, Visual C++ Express is smaller and still acceptable. However, there is no
guarantee that Microsoft will continue to make Visual C++ Express available for free
download.
We will use Visual C++ (through Visual Studio) in class. There are also other compilers
available for other platforms. If you use other compilers, it is required that anything that
you hand in must be compilable under Visual C++ and runnable under Windows 7 in
console mode.

3
 1.5. Design Cycle
The following table will highlight the design cycle that a typical C program will follow:

Program Input Output Function

Editor keyboard input .c (source), .h (header) create and maintain

source code

Compiler .c .obj take source code and

translate it into an
object file

Linker .obj, .lib .exe combine multiple

object files into an
executable file

Debugger .exe n/a allows debugging of

your program
The Visual C++ IDE incorporates many elements of the design cycle mentioned above.
First and foremost, it has a project manager that allows you to set up the type of
program that you want to write. Secondly, it has an editor so that you can create the C
source files that you will work with. It also has a compiler and a linker to create the
program. At the bottom of the IDE, there is a Messages window that contains status and
error messages associated with the compile and link. Lastly, it contains the ability to run
and debug the program that you have created. The details on how to create a program
from the setting up of the project to execution is found in Appendix A. Debugging is
discussed later.

1.6. Common Beginner Errors

• Make sure that you set up a project appropriately with the correct defaults.
• If you are having problems with your project, make sure that you follow the
instructions found in the appendices of this document.

1.7. Summary
This chapter has highlighted the basic underlying concepts regarding the C programming
language. C is an ideal language to create and maintain computer operating systems, to
manipulate low level hardware within a computer system, and because of the data
structure features of the C language, it lends itself quite easily in database manipulations.
C is a well-rounded language in that it offers much to just about any application that can
be thought of. There may be other languages which offer better ways of achieving the

4
same goals, but C at least allows you to attack any programming problem with few
overriding concerns because of its flexibilities.
We've also looked at how to create a C program, using an IDE through separate elements
of the design cycle.

5
Chapter 2. Source Code Structure
2.1. A Typical C Program
As was alluded to in the first chapter, a C program begins its life as a source code file.
The source code is a text file, which contains C statements, grouped in such a manner as
to create a program that can be compiled by the compiler. The compiler must take the
text from the source code file and translate these instructions into code that the
microprocessor will execute. This will allow you to run the software you have created.
There is a fairly standard appearance to typical C code. Just about all software written in
C will follow this order:
• Header comments
• Include files
• Constant definitions
• Type definitions
• Function prototypes
• Global variables if permitted
• Function definitions
To help us understand the above concepts, a sample C program will now be introduced.

/*
* Filename: file.c
* Project: assignment1
* By: Joe Schmoe
* Date: August 13, 2000 (yes, this is old)
* Description: This program is a simple demonstration
* of a C program.
*/

#include <stdio.h>

int main (void)

{
printf ("wow. a C program\n");
return 0;
} /* end main() */

2.2. Program Structure

Header Comments
You should notice that prior to the actual statement which includes stdio.h, there are a
number of lines of code which begin with a slash-asterisk pair (/*) and end with an

6
asterisk-slash pair (*/). The two pairs of characters are called the comment delimiters,
and are used by the C compiler to designate when you are adding comments inside of
your code. The comments are meant to help you document what you are trying to do with
your code, and are invaluable in situations where someone else will be required to
maintain the software you have written. It's always easier to read a programmer's
comments rather than try to always figure out what the programmer did by leafing
through pages of code! Note that modern compiler packages are dual C and C++
compilers, and will thus accept the C++ comment style: // (a pair of slashes side by side).
Every C file that you write should have a comment at the top of the file (called the
header comment). It is not required in order for the program to work but it is required
so that you and others can tell a bit about the contents of the file. Typical contents
include:
• filename
• project name (if different from the base of the filename)
• programmer's name (the first person to work on the file, typically)
• date of the file's first useful version
• description of the contents (this is important)

You will see many boxes like this in the course notes. They indicate Course
Requirements, Useful Tips, and other important things you should know.
In the case that there is a conflict between a Course Requirement found in this
document and some other document provided to you, the other document will have
precedence unless your instructor tells you otherwise.
Course Requirement: Header Comments
You must have a header comment in each source code file that you create. It must
contain the information mentioned above. The format should be similar to that shown in
the example above.
Course Requirement: Header Comment Description
The description contained within the header comment should accurately describe the
intent of the contents of the file. This is most important in multi-file projects that contain
many, many .c and .h files. Thus, a description like “This is assignment #2.” is bad while
a description like “This program takes Fahrenheit temperatures from the user and
displays them as Kelvin and Celsius temperatures. The program loops until the user
enters an invalid temperature.” is good.

Besides the header comment, you will typically have comments associated with the code
itself. As a rule of thumb, always try to make your comments as plain and judicious as
possible. Don't write a Shakespearean soliloquy when a simple statement would suffice.
Also, try not to comment every single line of code, rather only those lines of code that
perform a function that are not immediately obvious. C coding is, by its very nature,
generally self documenting.
7
The examples in these course notes will not have many comments, as the text of the
paragraphs surrounding the examples should serve the purpose of the comments.

Include Files
Just about every C program has a section where special files called header or include
files are declared to be compiled along with the source code stored in the C file itself.
These header files, many of which are supplied by the compiler manufacturer, are
necessary because they contain important constants and declarations to allow your C
program to properly access the standard library of functions. In Program 1, you will
notice we have used the statement:
#include <stdio.h>
to tell the C compiler to include a file called stdio.h (pronounced either "standard-i-o-
dot-h" or "stid-i-o-dot-h") when compiling Program 1. This header file contains
declarations for the standard input and output family of capabilities, which allow the C
programmer to obtain keyboard input, and send output to the computer console (monitor).
These input and output capabilities (usually called I/O functions) have been designed by
C compiler manufacturers to be common to all implementations of the C language, and
any software written to use this header file are (more-or-less) portable to any other
computing platform.
Whenever you add statements to tell the C compiler to include header files within your
own software, be aware that you are in reality asking the C compiler to temporarily
pretend that the statements within the header file actually belong to your source code.
Don't add header files to your source code unless you absolutely need the statements that
they contain. You will find that the standard header files for most C compilers are
designed in such a way that the statements they include are grouped by purpose. For
instance, if you want to manipulate mathematical equations, you can include math.h or,
if you want to work with character based strings, include string.h with your source code.

Useful Tip: Header Comments

Header comments are not comments about the header (or include) files. This is
especially important to know for exams!

Definitions, Globals, and Prototypes

After header file inclusion statements we have sections where we have various
definitions, global variables and function prototype declarations. As our first sample
program does not require these features, we will leave this section for the time being
since it is not a concept that is crucial to our understanding of C program format at this
moment. It suffices to know at this point that a global variable is a piece of memory that
can be used to store information your program will manipulate, and that a prototype
informs the compiler of a function before it is defined in your source code.

8
 Functions
What comes next are the function definitions. This is where the actual work your
program will accomplish resides. A function in C can be considered to be a block of
code which accomplishes a specific task. You will find that C functions are as diverse
as calculating the square root of a floating point number, or printing characters to the
screen or reading a file in from a disk. Each of the above actions are a specific task, and
as such, are easily handled by the creation of simple functions which are easy to use and
understand.

2.3. Functions

The main() function

The only function in Program 1 is the function called main(). (Note: When discussing
functions, C programmers usually denote the name of a function using the above
notation: the name, followed by a pair of round brackets. This easily distinguishes a
function name from a variable name when discussing source code.) This function is
mandatory in the sorts of C programs that you will be writing for this course. You cannot
fully compile and link a C program into an executable file without having a function
called main() somewhere in the source code you have just compiled. This is due to the
manner in which a C program executes once the operating system has transferred control
to your software - the operating system will in essence jump to the entry point inside your
application called main().
A C program always begins to execute on the first line of code within the main()
function. This line of code must exist after the opening brace (sometimes called a
"squiggly bracket" in informal C jargon) and before the closing brace. In Program 1, the
first line of code inside the main() function is the line which reads:
printf ("wow. a C program\n");

Standard Library Functions

This line of code is actually telling the computer to execute a task which is known as a
standard library function for virtually all C compilers: printf(). This function has the
job of displaying characters on the computer's monitor. The characters it will display are
placed inside of the pair of double quote (") characters, therefore the output we will see
on the screen is:
wow. a C program
The use of the double quotes and characters between them are an example of a construct
commonly found in C programs: a string literal. The entire string as seen above
constitutes a parameter for the function printf(). A parameter can be thought of as
information that flows into a particular function. The function itself will take the input
and process it and, at times, return some information back to the line of code which called
the function.

9
Useful Tip: Function Concept
A function can be thought of as a black box with information flowing into the box, and
information flowing out of the box. The black box will process the incoming information
and transform it to the output. Generally, from a programming point of view, one does
not need to know exactly what happens "in the box", just understand how your input is
transformed to the output.
Useful Tip: Arguments and Parameters
Parameters are typically called arguments when they are passed in the calling function. It
is not unusual, though, to refer to arguments as parameters as well.

Once printf() has printed to the screen the string parameter that it has been supplied, the
next line reached is the return 0 line. This returns control from the function. This will
be discussed later when we discuss functions in more detail.
Finally, we find that we come across the closing brace. This informs the C compiler that
this function is complete. Because we have just terminated the main() function, we are
actually terminating our C program as well. When main() completes, we are returning
control to the operating system or the compiler's development environment.
Although Program 1 is a very simple program, it is a perfectly valid example of how a C
program is constructed. All C programs begin to execute instructions contained in the
main() function. It is up to the programmer to decide what to place inside of the main()
function, thus main() is never pre-defined anywhere in any of the compiler's libraries.
Execution of the C program will stop as soon as there are no more instructions to process
within the main() function. A programmer may wish to put as little as 1 or 2 instructions
(as we have above) or millions (or more) statements within the braces of the main()
function. The latter example usually would prove to be far too unwieldy for most
programmers to handle, and as you will learn in later chapters, you are free to create your
own programmer-defined functions apart from main(). This feature allows streamlined,
modular C code to be produced with the minimum of fuss.

Course Requirement: main() Return Value

The main() function must return an int value. If you don't care what the value is (and
you typically don't), you can feel free to use return 0 before the closing brace in main().
Don’t return void! Don't leave out the return value!

Function Declaration Syntax and Style

The general format for a function can be described using the following declaration
notation:
returnDataType nameOfFunction (parameterList)
{
datatype firstLocalVariable;
. . .
datatype lastLocalVariable;
10
 firstLineOfCodeForFunction;
. . .
lastLineOfCodeForFunction;
} /* end function */
Notice the style of the sample code above: the return data type (to be discussed below),
function name, and parameter list are on the first line, against the margin. The opening
brace is on the second line against the margin. This layout is not mandatory but it is the
style that will be used in these course notes.
The body of the function (local variable declarations and code) is indented one tab stop.
The closing brace is back at the margin. This is a very common C programming style.
There are different styles that can be used (to be discussed later). The C compiler doesn't
care what style is used, but you should adopt a particular style, and stick with it.
Whitespace (blank lines, spaces, etc.) is highly encouraged when writing programs. It
breaks up a page of code into visually meaningful sections. Notice that in the sample
notation above, the local variable declarations (discussed in the next chapter) have a
blank line between them and the code of the function. This separates code from any
variable declarations.

Indentation and Bracing Style

For the most part, the C compiler does not care what indentation and bracing styles you
use. It's the people who you work with who have to look at and change your code that
care what your style looks like.
Why is style important? An absolutely incredible book called Code Complete: A
Practical Handbook of Software Construction by Steve McConnell (Redmond,
Washington: Microsoft Press, 2004) uses what it calls The Fundamental Theorem of
Formatting: "Good visual layout shows the logical structure of a program." To be
more specific, a good layout "should accurately and consistently represent the logical
structure of the code, improve readability of the code, and withstand modifications."
The primary feature that should be present in indentation and bracing style is consistency
within the code written in your functions. If you choose a particular style, then keep to
that style in that file. You will also find it best to be consistent across many files, so that
anyone looking at your code will have an idea of what to expect.
Here's some instructions on what to do:
• use white space to group related items together, using blank lines, spaces within
lines, and indentation
• indent every time you make a decision in your code (e.g. if, while, for, do-while,
switch) since you may or may not be executing the code in the corresponding block
• indent the bodies of functions so that they are not level with the left margin
• indenting one TAB for each code level is the best way to ensure consistent
indentation; you can usually adjust the number of spaces that each TAB stop
represents in the customization section of the editor you are using

11
Three bracing styles
Style 1: Braces level with code block
int main()
{
printf ("wow. a C program\n");
return 0;
}
Style 2: Braces level with control statement
int main()
{
printf ("wow. a C program\n");
return 0;
}
Style 3: Braces attached to control statement (K&R Style)
int main() {
printf ("wow. a C program\n");
return 0;
}

Flexibility in Indentation and Bracing

In certain cases, it does not make sense to require a style that might actually detract from
the understandability of the code being written. One of the primary places where
variation happens is variable declaration and initialization.
• If you're using style 2 or style 3, it's typically considered OK to indent a
variable declaration either level with the brace that surrounds the code block
or with the code block itself, like this:
int doIt(void)
{
int i = 0; /* level with brace */
printf("i is %d\n", i);
return i;
}
or this
int doIt(void)
{
int i = 0; /* level with block */
printf("i is %d\n", i);
return i;
}
Again, though, you must be consistent.
• If you have to use braces to initialize a variable (e.g. with a struct or array), it's
most desirable to use a bracing and indentation style that best helps with the

12
 understanding of the data. Quite often, that means simply having the data on
the same line as the variable declaration, like this:
int lotsOfStuff[5] = { 4, 54, 2, 3, 1 };
A typical question at this point is "why show us stuff like arrays that we haven't even
seen yet?" The answer is that it is important to not develop bad style habits so you're hit
with the right way to do it right up front.

Course Requirement: Style Consistency

You must use correct and consistent style in this course!

Data Type
The term data type may be unfamiliar to most novice programmers. A data type is
usually defined as the type, or format, a particular piece of information may have within a
program. For instance, the number pi (3.1415926535...) is said to be a floating point data
object since it is a real number (it has decimal digits following the whole number part and
decimal point). The letter A is usually said to be a character data object because the
letter A is definitely a character of the alphabet. The number 37 is an integer data object
since it is a number without a decimal point (integers contain no fractional portions).
For our main() function inside Program 1, you will notice that we have declared the
return data type and parameter list of main() to be int. This indicates that an integer value
will be returned to the caller of main(), which is usually the operating system. In the next
chapter, we will have a close look at the different data types that are standard in C.

Function Name
After defining the return data type, a function requires the name to be defined. The name
of a function can consist of any combination of alphanumeric characters as well as the
underscore character, although a name cannot begin with a number and at least one
alphabetical character must exist in the name. Valid function names include:
test1()
_123Execute()
a()
A word of warning: you must always define unique names for your functions (and
variables). Two functions with the same name in the same source code file will lead to
compilation errors, forcing you to rename one of the offending functions. Also, different
compilers have different restrictions on the maximum lengths of names. This is typically
not a problem, since the lengths are usually large enough.
Also, it should be noted that C is case-sensitive. That is, uppercase and lowercase letters
are deemed to be different. Thus, the following function names are all different:
doIt()
doit()
DOIT()
DoIt()
13
In this case, however, it would be a very bad decision to create more than one of the
above functions in a program. A great deal of communication in the workplace is verbal:
you don't want any ambiguity if someone tells you "just take a look at the do it function".
For the same reason, you should always be consistent in how you use case when you
name functions and variables. If you tell a co-worker "just take a look at the do it
function" and they know your style of programming, they should be able to immediately
know what the spelling of that function name is.
Start all of your function names with a lowercase letter. This will help you differentiate
functions from C++ methods when you program in C++. The language will let you use
either camel case (e.g. doIt()) or underscores (do_it()) to separate words in function and
variable names. That way, your names will be easier to read and understand. The use of
underscores is rapidly becoming obsolete, however.

Course Requirement: Distinguishing Between Words In Identifiers

Use camel case to distinguish between words in names, unless otherwise explicitly stated.
Course Requirement: Naming
Start your names with a lowercase letter, unless otherwise explicitly stated.

Although the standard syntax for a function suggests that we must define the names and
data types of the parameters for the function, you will find that not all functions require
input parameters. These functions will therefore take a void parameter list, as our
example in Program 1. void is used to indicate that no information will be flowing to the
function. If you are designing a database program, and a common help screen is required
from a number of different places within the program, you may want to put the help code
inside of a help function. Such a function does not need input: its only job is to display
help information on the screen, therefore the parameter list will also be void. It probably
will not return a value to the caller of the help function, therefore its return type would be
void as well (indicating that no information will be flowing back from the function to the
caller). We will see more examples of data types and variable names in the next chapter.
One final note: many statements within any function require the programmer to append a
semicolon to the end of the statement. This is the method the C compiler uses to
delineate when one statement is complete and another begins. This is an important
concept to understand and a useful concept since C does not require a complete statement
to fill one line. In fact, a valid example of a C program statement includes:
printf ("testing %d %d %d", variable1, variable2,
variable3);
Even though the statement above takes two lines of code, it is still only one single
statement to the C compiler, since the semicolon is at the end of the second line of code.

14
 Useful Tip: Semicolons
Figuring out when semicolons are required in a C program can seem like a black art.
Here's some general guidelines, some of which use terms that will be described later in
the notes:
Use a semicolon:
- variable or type declarations
- function prototypes
- function calls that are alone on a line
- normal executable code statements that do not have a code block after them
Don't use a semicolon:
- comments
- "#" statements
- function declarations
- "{" or "}"
- statements that usually have a code block after them (e.g. if, while, for, switch)
- statements within function calls or conditions
Useful Tip: Finding Compile Errors
If the C compiler generates an error which does not appear to make sense at the line
number where the error is reported at, generally it is advisable to look at statements just
prior to the reported line number. Often, a missing semicolon is the culprit prior to where
the compiler got confused.
A C compiler will attempt to recover from an error and make assumptions about what
you had intended. Often, the assumptions it makes lead it to other compiling errors. So,
do not be alarmed if a single missing semicolon causes hundreds of errors to appear - it
really isn't that serious!

2.4. Another Sample Program

Here's another sample program which will be discussed in class:
#include <stdio.h>

int main(void)
{
int number = 9;

/* print the number upon startup */

printf("Numbers are coming... ");
printf("The number is %d\n", number);

/* change the value of the number and

15
 * print it */
number = 8;
printf("The number is now %d, not %d\n",
number, 5);

return 0;
}

Useful Tip: Code Samples in this Document

You will find that your program runs too quickly and then disappears. If this happens,
you can choose to execute the program in a different way. If you choose Start Without
Debugging (which is, oddly enough, found in the Debug menu), your program will pause
at the end. This will, as the name suggests, not allow debugging, though.

Changing The Sample Program

Once you have compiled and run the program above, do the following to the program:
Make a minor change
Change int number to int fred.
Execute the program.
Note the error.
Why did the error occur?
Fix the error.
Execute again.
Syntax change
Put a semicolon after the int main(void).
Execute.
Note the errors.
Note that the errors aren't all that clear in terms of helping you solve the problem.
Undo the change.
Execute.
Comment change
Get rid of the */ at the end of the first comment. Note the change of colour.
Execute.
Note the change of result.
Undo the change.
Execute.

16
 String change
Delete the closing " in the first string.
Execute.
Note the error.
Undo the change.
Execute.
Bracing change
Delete the closing }.
Execute.
Note the error.
Undo the change.
Execute.
Review errors
Note the error wordings and how they related (or didn't relate) to the causes.
Note the error locations and how they related to the actual locations of the causes.

Useful Tip: Error List

Start compiling a list of common errors. Share the list with your friends.

Yes, I'm serious.

2.5. Common Beginner Errors

• Make sure that you remember your braces and semicolons.
• Close all comments using */ or use the // style of comment.
• Put all code that you want to execute within a function (for now, within main()).

2.6. Summary
In this chapter we have learned a number of useful concepts, of which the idea of a
function is the most important. All work that a C program will accomplish is through the
use of one or more functions, each performing a certain task. Every C program that we'll
be doing in this course must have at least one function called main() somewhere in the
source code. A C compiler cannot produce a complete executable file without main()
existing in the source code.
Other useful concepts include the use of header files in our source code. A header file
consists of definitions of important constants and functions which allow us to take
advantage of the features in the standard library of functions.

17
A function will begin execution after the opening brace of the function, and will cease
execution prior to the closing brace of the function. The main() function contains the first
and last lines of all C programs, which makes this function the most important function
of all.
Style is very important, as source code that is hard to follow is bad source code.
We also looked at some specific ways of working with C, compiling and making
corrections to fix errors.

18
Chapter 3. Variables and Data Types
In the previous chapter, the concept of variable declarations, although standard parts of
any C program, was left until this time. This was due to the fact that a fundamental
understanding of what a variable is in C, as well as the type of memory storage allocated
to each data type must come first before learning how to declare a variable. Firm
knowledge of data types and storage requirements will allow a competent C programmer
to write efficient code based upon the data requirements of the program (data driven
software) rather than basing code on what the language forces you to use (code driven
software).
In C, there are a number of fundamental data types already defined for the programmer to
use. These data types are based upon the standard building blocks the microprocessors
and co-processors of today use. The data types are thus oriented around what a processor
can access efficiently, and as such, they may appear to be strange to a budding C
programmer.
A variable is simply a piece of memory that has been allocated (obtained) from the
overall memory of the computer, for the purpose of storing values that a program may
wish to record. Variables are useful in storing values based upon computations a program
generates, and thus, some knowledge of the various types of data objects is necessary
before a C programmer can define any variables.

3.1. Data Objects in your CPU

Virtually every computer can access the memory inside of the central processing unit
(CPU) in the form of bytes. A byte is normally the smallest unit of memory the processor
can accesses at one time, and it consists of 8 bits, each bit representing the number 0 or 1.
This is the same type of information a light bulb stores - it's either on or off, zero or one.
With eight bits of information within each byte, it can easily be shown that there are 256
possible combinations of bits within a byte, thus the byte can store a number between 0
and 255 inclusive (a range of 256 values).
The next step up from a byte is the word. A word is a pair of bytes side by side in
memory on PCs, which gives us 16 bits of information at one time. With 16 bits, a word
can have 65536 possible combinations of values, thus the range of values that can be
represented within a word are zero to 65535 inclusive.
Next, computers work with long words. A long word consists of a group of four bytes
side by side in memory accessible at one time, a total of 32 bits of information to
manipulate at any given time. With 32 bits, values from 0 to well over 4 billion can be
represented. Some other microprocessors also use the quad word, with 64 bits of
information.
With the above three different data objects, an important fact must be recognized: bytes,
words and long words are all integer data objects. The term integer data object refers to
the fact that only whole numbers can be represented by the value stored in such a data
object, with no fractional part. Thus, 3.5 cannot be stored in any combination of bits since

19
 the fractional portion, .5, cannot be represented. Another data object (to be described
below) allows a C programmer to avoid this pitfall if decimal fractions are desired.

3.2. Signed vs. Unsigned

Another important fact about the integer data objects is that if the highest bit (or MSB -
most significant bit) of the particular data object is set to a value of one (on), a CPU has
the ability to work with that number as either a signed (both positive and negative
numbers) or unsigned (only positive numbers). In either case, the total number of
different values does not change, only the interpretation of the values that can be
represented. For instance, for a signed word, values between -32768 through 32767 can
be represented. This still gives us 65536 different possible values, only the meaning of
that high bit has changed. CPUs have the ability to work with both positive and negative
integer values using the standard byte, word and long word data objects.
It should be apparent to a budding programmer that, depending on the nature of the
software that must be created, a wealth of data objects can be manipulated, each with its
own natural range of values. Imagine a program that is required to maintain votes for a
national referendum in Canada. By studying the data requirements here, a programmer
should realize that it could be possible that approximately 15 million adults can vote in
total, thus to be able to store the count of all votes for a yes vote or a no vote, a long word
will do the trick. There could be 15 million yes votes or 15 million no votes, or a
combination of the two. By trying to save memory and using a word or byte, a
programmer will be inviting possible errors if more than 256 or 65536 people vote in the
referendum. A data object that is too small to hold the number will wrap around or be
truncated after its natural limit is reached. On the same token, to keep track of the speed
of a standard passenger automobile, a program can use a byte since there are virtually no
street-legal cars that travel more than 256 kilometres per hour!

3.3. Floating Point

If you can recall high school mathematics, a real number is a value that has both a whole
number portion and a possible fractional portion as well. The number (3.14159 or pi) is a
floating point, or real number, since it has a whole number portion (3) and a decimal
fraction portion (0.14159). We can't express most real numbers using integer values
because of that fractional portion. We can, however, deal with them as floating point
values.
These are special data objects for most microprocessors since they are not naturally
equipped to deal with such values. CPUs are designed to work with integer-based values
with high speed, and only deal with floating point values using special routines in
whatever software is running.
A floating point data object has the capability of storing values that far exceed the normal
range of values for a long word. In fact, the typical floating point number today is
represented as a sequence of 8 bytes, with the 8 bytes having special meaning to the
floating point co-processor. With such a large number of bytes used to store floating
point values, a programmer can count on the range of values to be quite large. In fact, the
20
 usual range for a floating point value is approximately 1.0 * 10308 to -1.0 * 10308. There
is a discontinuity near 0, as well, as the smallest positive value that can be expressed is
usually 1.0 * 10-308 (which is 1 divided by 10308). Naturally, the smallest negative value
that can be expressed is -1.0 * 10-308. Each compiler can have different values for the
endpoints of these ranges.
You should also note that a floating point number is always a signed value: there is no
unsigned counterpart as in the integer data objects.

3.4. Compiler's View of Data Objects

The C language has been designed to take advantage of a computer's internal architecture
so that efficient use of memory can be obtained. Thus, the standard data types available
to a C programmer, which are predefined by C, are based upon the data objects described
above.
The byte data object is called char in C. A char is by default a signed data type, thus a
variable of this data type can store values from -128 through 127. A C programmer can
also declare a variable to be the unsigned version of a byte by using the data type
unsigned char. In fact, placing the keyword unsigned in front of any of the integer data
types (the rest to be described below) can produce the unsigned or positive only versions
of the standard integer data objects. Both char and unsigned char take one byte of
storage.
The word data object is defined as short. It too is signed, allowing a variable of this type
to hold values from -32768 to 32767. Again, the unsigned version is simply unsigned
short. Both short and unsigned short take 2 bytes of memory.
The long word is defined as long, and is a signed data type. Its range is from less than -2
billion to over 2 billion. The unsigned equivalent is unsigned long. Both long and
unsigned long take 4 bytes of storage.
In C, the data type called int is defined to be the natural word size of your CPU. In other
words, the most natural processing data type that your CPU is geared to work with is
assigned to be an int data type in C. An unsigned int is therefore the positive-only
version of this data type. For some CPUs today, an int and an unsigned int are 32-bit
values. For other modern CPUs, they are 64-bit values. The current version of Visual
Studio uses 32-bit ints and unsigned ints, regardless of what the CPU is.
The floating point data object has two different representations in C. The first is called
float, which is a single precision floating point number representation. The term "single
precision" is used to denote how many significant digits the data object can hold, with
single precision typically being able to hold about 7 or 8 digits of precision, much like a
typical calculator display. This is more than enough for a typical household budget, but
nowhere near enough to calculate accurately where the next satellite will be placed into
orbit.
To alleviate this problem, C has another version of the floating point number called a
double, which, as you may have already guessed, is capable of storing twice as many
significant digits as the float, normally 14 to 16 digits of precision.
21
 Although both float and double can hold floating point values, it is advisable not to use
the float data type as most math co-processors use the double data type automatically.
Extra code must be created by a compiler to convert a float to a double prior to invoking
an instruction on the floating point co-processor. The only advantage a float has over its
double counterpart is that the former takes only four bytes of memory, while the latter
uses eight. Therefore, in cases where memory limitations dictate data type choices, a
float may be useful. For most applications, use the double as the standard floating point
data type. Again, different compilers will have different values for the sizes of the float
and double datatypes.
The following chart may help you to decide upon what data type is applicable in your
programs.

Data Type Bytes Range Values

(typical)
char 1 -128 to 127 256
unsigned char 1 0 to 255 256
short 2 -32768 to 32767 65536
unsigned short 2 0 to 65535 65536
long (or int) 4 -2147483648 to 2147483647 4294967296
unsigned long 4 0 to 4294967295 4294967296
float 4 -3.4 * 1038 to 3.4 * 1038 infinite
(with a discontinuity close to
0)
double 8 -1.7 * 10308 to 1.7 * 10308 infinite
(with a discontinuity close to
0)

3.5. Notation
To actually create a variable in C (called "declaring the variable"), the code must follow
the following syntax:
dataType variableName;
The first portion of this declaration informs the C compiler of the data type the variable is
composed of. The C compiler must know this information before it allows you to use any
variables in your program since the C compiler must generate code to allocate memory
resources for your software.
The second portion of the declaration is the name itself. The variable name follows the
same naming rules as does the name of a function. The following variable declarations
are valid in C:
int x;
double _fpSeg;
char keystroke;
long a, b, c; /* see next paragraph for a commentary on this */
int flag = 0;

22
 Notice the second last declaration. Here we have instructed the C compiler to reserve
three long words of memory, named a, b, and c. When you have more than one variable
with the same data type, the compiler will allow you to put them on the same line of
source code, separated with commas. This is often considered bad form and some other
languages do not allow this, though, so it is not allowed under our SET Coding
Standards.
Notice also that all variable name declarations, when complete, are terminated with a
semicolon, which must be present for the C compiler to accept the declaration as valid.

3.6. Initialized Variables

In the above examples, the final variable, flag, was followed by an equal sign and the
number zero. This is an example of initializing your variables at the time of its
declaration. This is a common (often required) programming practice, and is an indicator
that the programmer understands the default values of variables in their program. Many
compilers do not guarantee what the default values of variables will be, so it is up to the
programmer to initialize variables with their proper values as often as possible.

Course Requirement: Initialization of Variables

Always initialize your variables when you declare them, unless notified otherwise.
Failing to do so will cause you to lose marks and get quite upset about it.

3.7. Global Variables

In the previous chapter, it was stated that variables may be declared after the header file
inclusion statements and before any function declarations. Variables defined at this point
of a C program are termed global variables because the inventors of C deemed that
variables defined outside the scope, or sphere of execution of a function (outside the open
and close braces of a function), are allowed to be accessible within any function of a
program, hence the term global.
Global variables are very easy to use but are also extremely dangerous. Since they can be
changed anywhere within a program, they quite often make debugging logic problems
horrendous. Their use should be heavily restricted. It is a good idea to have global
variable use reviewed by other members of the programming team.

23
 Course Requirement: Global Variables Banned
Do not use global variables unless explicitly permitted by the instructor. Note that this
does not restrict the use of global constants (variables whose declarations are preceded
by the keyword const; this is to be discussed later).
There is a LARGE mark penalty that is associated with using global variables when not
allowed to.

3.8. Local Variables

It is often the case that a programmer may want to make a variable accessible only during
the execution of a particular function. This type of variable is called a local variable,
since its accessibility (called scope) and lifetime is limited to the execution of a single
function. Local variables are usually declared after the opening brace of a function and
before the first line of code and only that one function can access those variables.
It should be noted that a local variable can have the same name as a global variable, but
the local variable will take precedence over the global inside of that function. Having
local variables that have the same names as global variables is A Bad Idea. Having local
variables in several functions that have the same name is a common thing and is A Good
Idea.
Global variables can be useful for storing information of a overall nature in terms of
software execution. For instance, a word processor may want to globally store the state of
the Insert key on the keyboard since all portions of the word processing software must
know if it is acceptable to insert new characters instead of overwriting old characters.
Another example would be storing the default drive and path information for the loading
and saving of file oriented information from a program. However, in most normal cases,
global variables can be eliminated through proper design and choice of programming
language. You will see the exceptions to this in some second year courses.
Local variables are useful for storing intermediate results of computations during the
execution of a function. If a function was written to calculate the average of a series of
numbers, a local variable may be created to hold the running total of the series prior to
averaging.
To help clarify these topics, two short programs will be introduced. Both programs
declare the same variables, although Program 2 declares them globally, and Program 3
declares them locally. The aim of both programs is to calculate the average of two
numbers.

/*
* Program 2 - PROG2.C
* NOTE to student: to save space in this document, full
* header comments will not be used. You must still
* provide them in assignments, though!
*/

24
#include <stdio.h>

/* global variables */
int a = 0;
int b = 0;
int c = 0;

int main (void)

{
a = 5;
b = 3;
c = (a + b) / 2;

printf ("a: %d b: %d c: %d\n", a, b, c);

return 0;
} /* end main */

The global int variables declared in Program 2 will be accessible to all functions defined
in that program. The main() function has access to these global variables because of this
feature of the language.

/*
* Program 3 - PROG3.C
*/

#include <stdio.h>

int main (void)

{
/* local variables */
int a = 7;
int b = 2,
int c = 5;

c = (a + b) / 2;
printf ("a: %d b: %d c: %d\n", a, b, c);
return 0;
} /* end main */
In Program 3, because a, b, and c were declared as local to main(), no other function
within this program can access the values of these three variables. These three variables
are shielded from influence by external functions.

25
 Useful Tip: Finding Global Variables
Global variables are defined outside of functions.

3.9. Constants in C
Often, a programmer may want to assign a constant value into a variable. A constant is
usually just a number on the right hand side of an equal sign, or a number that is part of a
mathematical equation. In both PROG2.C and PROG3.C, we can see constants being
assigned into integer variables prior to a mathematical operation being performed upon
them.
For non-ANSI C compilers, there are two situations that require special handling of
constants, to avoid a problem where all numeric constants were treated as integers, unless
they were followed by specific modifiers.
For floating point constants, many non-ANSI C compilers require that the number be
written as the whole number portion, a decimal point, and a fractional portion, even in
situations where there is no fractional portion. For instance, the statements:
double x = 0.0;
. . .
x = 5;
may seem obvious to us that the floating point value 5 was being placed into the double
precision floating point variable x. But older compilers would try to assign the integer 5
into the double, leading to an incorrect copy of two bytes into eight. To avoid such
problems, the programmer would simply write:
x = 5.0;
for the assignment. To ensure no errors ever in floating point assignments or constant use,
make sure that a floating point value has the decimal point, and at least a zero following
if the number has no decimal digits.
For long integer constants, it is conceivable that you could run across a compiler that
does not understand the following assignment:
long x = 0L;
. . .
x = 25321;
Older compilers would again try to assign an integer to the long integer variable x. To
avoid this problem, again, a special notation was introduced:
x = 25321L;
By adding the letter L (uppercase or lowercase) to the end of the constant, we can be
assured that the compiler will compile code to assign the long integer version of the
number into the long integer variable.
In programming situations where constant values are altered many times during the
development of software, or if the programmer wants to make a piece of code more
readable, the following idea can be used:
26
 #define CONSTANT_NAME constantValue
The above generic syntax shows a constant being declared using the #define approach.
This mechanism allows a programmer to create a name for a constant, and have the name
associated with the constant's value. This idea does not allocate or create variables - it
simply allows the programmer to use a name instead of a constant number.
For example, study the following:
#define CANADA_POPULATION 30000000L
. . .
long pop = CANADA_POPULATION;
The above code shows a definition of a constant called CANADA_POPULATION, and
this long integer constant (value of 30000000) is being placed into a variable called pop.
The CANADA_POPULATION constant can be used anywhere a programmer wants to
reference a number that represents Canada's population, and makes the software much
more readable as a result! These constants are usually named with a totally-uppercase
name.
The same result can be achieved by using the const keyword preceding a variable
declaration, like this:
const long kCanadaPopulation = 30000000L;
These constants are usually named like variables, except for being preceded by the letter
'k' (short for "konstant" (yes, I know)).
Of the two methods, the const method is considered to be the best, as it incorporates a
desirable use of data type to make code more solid. However, it is not universally better
than #define (more will be discussed about this when we cover arrays).
Constant definitions are generally always placed after the header file include statements
and before the function prototypes. There is typically no taboo against "global
constants", as they cannot be changed and, thus, do not have the same pitfalls as global
variables.
For more detailed information on constants and their use, please see the section on Rules
about Variables.

Useful Tip: Constant Use

It's best to use constants defined using #define or const rather than hard-coded. That
way, if you need to change it, you change it in one place rather than many places.
Useful Tip: Constant Naming
If you use #define to create a constant, you can name it totally in UPPERCASE with
underscores (_) separating the words.

3.10. Character Constants

For character constants, it is common to try to put the letters of the alphabet, numbers,
spaces and punctuation into variables. C compilers allow you to directly assign such

27
characters by placing the character inside of single quotation marks. Study the following
lines of code:
char a = 'A';
char b = ' ';
char c = '\n';
The first assignment places the letter A into variable a. The letter A is actually
represented inside of memory using the ASCII (pronounced “ass-key”) equivalent of the
letter. ASCII (American Standards Committee for Information Interchange) values are
simply numbers which are common to all computers so that one computer can represent
the same set of letters as any other computer. This is the way two different computers can
exchange textual information without error.
The second example above shows how to assign a space character into a variable: simply
place the space inside of the single quotes. The third example shows how to put a newline
character (what causes the cursor to move down a line on the screen) into a variable. The
newline character is represented by a backslash followed by the character 'n', all
surrounded by single quotes.

(the above table is from

http://fr.wikipedia.org/wiki/American_Standard_Code_for_Information_Interchange#me
diaviewer/Fichier:ASCII-Table-wide.svg)
A programmer is able to use a character-based constant in their programming by putting
the codes listed above inside of single quotation marks.
For example, if you wish to set a variable equal to the letter Q, the character constant is
'Q'

28
 and for a beep (or bell), the character constant is
'\a'
The use of a two-character (or longer) long character sequence within single quotes and
starting with a backslash defines a single character. This is a notation that is used for
most of the codes below 32 (the space character). They are termed control codes and
were used to control ASCII based terminals in the earlier days of computers. There are
some useful codes in the first 32:

Common Name Character Meaning

CR '\r' carriage return

LF '\n' newline
TAB '\t' TAB
BELL '\a' beeps

The presence of '\x' indicates that there will be more characters that define a hexadecimal
value (see the Appendix on Counting Systems for more information on hexadecimal) for
the single character.
It is advisable to always use a character constant inside of single quotes rather than the
number that represents the character. This is especially important when dealing with
Unicode characters (which is beyond the scope of this course). If you are interested in
finding out more about internationalization through Unicode, there is a good short video
found at https://www.youtube.com/watch?v=MijmeoH9LT4 (if it's still there when you're
reading this),

3.11. How to Output Variables

We have been exposed to the standard library function printf() as a means of printing
textual information to the screen. This function is also capable of formatting our variables
to the screen as well, giving rise to the name of this function (printf <==> print
formatted).
The following lines of code illustrate how to use printf() to print values of variables:
int x = 10;
long y = 3L; /* use L to indicate a long value */
char z = 5; /* this can be a number or a character */
double w = 9.0; /* append the decimal place */

printf ("x is equal to: %d\n", x);

printf ("y: %ld; z: %c\n", y, z);
printf ("w is equal to: %f\n", w);

29
 Notice that each of the above printf() calls uses a percent (%) sign followed by one or
more characters. The percent sign is used to inform printf() that we want to output one of
the parameters following the quoted string of characters. For the first example, we have
%d, telling printf() to format the second parameter, x, as a signed integer. The second
example uses two percent formatting strings: %ld tells printf() to output the second (y)
parameter as a long integer and %c, which tells printf() to output the third parameter (z)
as a character. In the last example, we are using %f to print the double precision floating
point variable w using floating point notation.
The following table describes some of the more common formatting codes:
Specifier Output Example
d Signed decimal integer 392
u Unsigned decimal integer 7235
o Unsigned octal 610
x Unsigned hexadecimal integer 7fa
X Unsigned hexadecimal integer (uppercase) 7FA
f Decimal floating point, lowercase 392.65
F Decimal floating point, uppercase 392.65
e Scientific notation (mantissa/exponent), lowercase 3.9265e+2
E Scientific notation (mantissa/exponent), uppercase 3.9265E+2
g Use the shortest representation: %e or %f 392.65
G Use the shortest representation: %E or %F 392.65
c Character a
s String of characters sample
p Pointer address b8000000
% A % followed by another % character will write a single % to the stream. %

By using the above formatting strings, any data type discussed thus far can be displayed
on the screen using printf().
It is interesting to note that you may print out the numeric, or ASCII, value of a char
variable using %d, %u, or %x, and also print out the ASCII equivalent character of any
integer by using %c.
You may have also noted that in our earlier examples of printf(), we had the pair of
characters \n being used. The proper term for these pairs of characters is escape
sequence, and is used to embed special characters within our strings that we normally
could not represent or type from the keyboard. The most common of these is '\n', used to
go to the start of the next line. There are others, though.

30
Escape sequence Meaning

\n newline
\r carriage return
\b backspace
\x00, \x01, etc. hexadecimal values
\t tab
\" double quote
\\ backslash

Notice that if we wanted to place a backslash within the string for printf(), we need to
escape it itself using another backslash. It looks funny, but it is a valid escape sequence!
As an example, if you wanted to print a string which sends a beep, a tab and a carriage
return to the screen, the formatting string for printf would be:
printf ("beep: \a tab: \t carriage return: \r");
It should be apparent that escape sequences and variable formatting can be intermixed
anywhere within the formatting string for printf(), making this function extremely
versatile and powerful. In fact, many programmers like to think of the formatting string
as a method of specifying how the output will look, including the "slots" where
information will be formatted. The extra parameters will correspond to the slots that have
been specified.
To make the formatting string just a bit more flexible, the programmer can specify a
number in between the percent symbol and the formatting code to specify the padding of
the format. For instance:
printf ("x is equal to: %4d\n", x);
will print the value of the variable x, but because there is a 4 between the % and the d,
the variable will be padded out to fill 4 characters on the screen. printf() will pad such
format strings with extra spaces in front of the number to fill it to the appropriate number
of characters. Using such a method, a programmer can easily print columns of lined up
numbers simply by requesting a specified width or padding for the numbers in the table!
As well, by adding a zero in front of the number, you can pad a numeric format with
leading zeros! The following:
printf ("x: %04d\n", x);
will output to the screen (if x is currently equal to 10):
x: 0010
Finally:

31
 printf ("real number: %8.2f\n", a);
will format a floating point number, taking up eight spaces in total, with two of those
characters coming after the decimal point. It should be apparent that this is a convenient
method of printing currency (dollars and cents) values to the screen. All of these nifty
features of printf() clearly make this function a true workhorse!

3.12. Common Beginner Errors

• printf() always has a format string. This format string usually starts with a " and ends
with a ".
• Any formatting codes present in a printf() formatting string must have an expression
for the value to which it corresponds. That expression (which could be as simple as a
variable) must follow the format string, separated from the format string by a comma.
• Other common printf() error examples:
• printf("%d, number");
• printf("%d");
• printf(%d, number);
• printf("%d") number;
• printf("number");
• printf("%d", number, "\n");
• printf("The number is ", number, "\n");

3.13. Summary
In this chapter, we learned about the various data types that a computer is designed to
handle as well as the names C gives to these different data types. C contains equivalents
to the byte, word and long word data types, char, int and long, their unsigned
counterparts (by prepending the keyword unsigned in front of the data type in the
declaration), as well as two different floating point data types, float and double.
The syntax for a variable declaration was introduced, and the naming rules for variables
were found to be identical to the naming rules for a function, as shown earlier.
Next, a discussion concerning where to declare variables led to the definition of two
classes of variables: global and local.
We also discussed the methods of declaring constants of different types, and we had a
look at the ASCII table of values.
Finally, we looked at outputting text using printf(), how to format the output of
variables, and how to print characters which are normally not accessible within the
formatting string of printf().

32
Chapter 4. Operators
4.1. What is an Operator?
As a programmer, we know that once we've defined our variables for our program, we
will want to try to do something with these variables. Perhaps we will want to add two
variables together, and store the result in a third variable. This is an example of two
different operators in use: addition and assignment. It is through the use of the different
operators that the C language offers that meaningful work can be performed in our C
programs.
Addition is the simplest one to understand since addition is the same in math or in C
programming. Assignment is a bit harder to grasp. An assignment is simply telling the C
compiler to store a value inside of a variable. In our example above, the third variable is
assigned the result of adding our first two variables together.
From math, we know that there are many mathematical operators to use. We can
subtract, multiply, and divide. You will find that C naturally handles the standard
operators but forces the programmer to create functions which perform more complex
operations.
C will also allow you to perform bitwise operations, which allow the programmer to
manipulate the bits directly inside of a variable. As well, we will see the use of
comparison operators that allow the programmer to test if a value is equal, greater than,
not equal or less than, another value.

4.2. C Operators
Most of the commonly found operators in any computer language are found in C as well.
As in math, C contains two classes of operators: binary operators and unary operators.
Binary operators are used in between two values upon which the operator is to act upon.
Addition (+) is a binary operator since it will act upon two values on either side of the
addition sign. The negation operator (-) is a unary operator since its job is simply to
negate the value that follows the operator.

33
 4.3. Binary Operators
Here are some of the more common binary operators:

Operator Use

+ addition

- subtraction

* multiplication

/ division

% modulo

&& logical-AND

|| logical-OR

== equality comparison

!= inequality comparison

< less-than

> greater-than

<= less-than-or-equal-to

>= greater-than-or-equal-to

= assignment of a value to a variable

There are also bitwise operators, which are explained in the Appendix on Counting
Systems. It suffices to know that they are used to manipulate the bits within the various
data types we've been exposed to thus far. The same goes for the shift operators, since
they are used to move bits side to side within data objects.
The && and || operators work in terms of boolean logic, the logic of true or false. The
&& operator will produce a boolean true result when both sides of the operator evaluate

34
 to boolean true, and the || operator will evaluate to a boolean true if either one or the
other, or both sides of the operator evaluate to a boolean true value. Otherwise, the result
of either operation is boolean false. The comparator operators such as equality,
inequality, greater than, less than, greater than or equal and less than or equal will
produce a boolean result if the comparison succeeds (true) or fails (false). These boolean
operators will be used extensively when we explore flow of control later.
A curious operator is present in the above table: modulo. This operator is used to perform
a division between two integer values, and extract the remainder of the division. This is
the complement of the division operator, which will, in the case of the char, int and long
data types (and unsigned versions of these data types), produce just the whole number
result. For instance:
7 / 3 = 2.333333333
With modulo division, we obtain the remainder of 1, which is one third (1/3) of the way
to the next even division of a value by 3.

4.4. Unary Operators

Operator Use

- negation
! boolean inverse (a.k.a. not)

The negation operator should be a familiar operator (changing a value from positive to
negative and vice versa). The exclamation mark is the boolean inverse operator, which
flips a true result into a false result, and vice versa.

4.5. Operator Examples

Here are some examples of how to use the above operators:
x = 5; /* assignment */
y = x * 3; /* multiplication and assignment */
z = y % 2; /* modulo and assignment */
q = x && y; /* boolean AND and assignment */
a = -y; /* unary negation and assignment */
b = a * x / y + z; /* combinations! */
In looking at the above examples, it's possible that the one that you might find most
obscure is the third one, involving the modulo operator. However, this particular
example is extremely valuable, as it assigns z with a value of 0 if y is even and 1 if y is
odd. It's nice to be able to determine this in a single line of code!

35
4.6. Precedence Rules
It should be noted that in C, just like in math, there is an order of operations that the
compiler will follow in evaluating an expression. For instance, multiplication takes
precedence over subtraction, meaning that a multiplication will be done first before any
subtraction. As an example, look at the following statement:
a = b * c - d * e;
Because multiplication is more important than subtraction, the compiler will generate
code that will multiply variable b with c, then multiply d with e, and finally subtract the
second multiplication result from the first before assigning it to variable a.
The precedence rules for the operators we've seen so far can be summarized using the
following table:

Operators What They Do

*/% Multiplicative

+- Additive

< > <= >= Relational

== != Equality

&& Logical-AND

|| Logical-OR

= Assignment

Notice with the above table of precedence, the assignment operator will always occur last
when the C compiler generates code. Also notice the order of precedence for the logical
and comparator operators since this ordering makes writing comparison logic much
simpler in the long run.
Of course, just like in math, you may override the precedence order at any time by using
brackets. For instance:
a = b * (c - d) * e;
looks similar to our earlier example, but because of the brackets, the compiler will
generate code to subtract d from c before multiplying that result with b and e.
Note that adding brackets does not make code less efficient, it simply allows the
programmer to alter the order in which a statement will evaluate or to make the desired

36
 order clearer. Note as well that the brackets around an expression operate differently than
the brackets required for a function call. Study the following incorrect expression:
a = b (c - d) * e;
This statement is similar to our previous statement, but notice how the first multiplication
sign is left out. In math, this would be a valid way of expressing an implied
multiplication, but in C, the compiler would think that we are multiplying e with the
return value of a function called b!

Useful Tip: Brackets in Arithmetic

Use brackets rather than relying on your (or someone else's) knowledge of precedence in
all but the most trivial cases.

4.7. Assignment Operators

Another important concept is the assignment operator in C. An assignment operator is a
binary operator that is used to give a value to a single variable. This may sound odd, but
an example of an assignment operator will clear up any confusion quickly:
x = 5;
The above statement simply puts the integer 5 into the variable x.
Here’s a somewhat less familiar assignment operator:
x += 5;
The above statement uses the addition assignment operator, which will take the current
value of the variable x and add 5 to it, and store that result back in variable x. The
addition assignment operator will increment a variable by some amount. The above
statement is equivalent to:
x = x + 5;
There are also similar assignment-like operators for other binary operators. For example,
-= behaves like += except that subtraction is done instead of addition. Similarly, *=, /=,
%=, etc.

4.8. Increment and Decrement

The final two operators to look at are the increment and decrement operators. The
increment and decrement operators are used to alter a variable's contents upwards or
downwards by one in value. The operators are ++ for increment and -- for decrement.
The following lines illustrate how to use these operators:
x++; /* x = x + 1; */
y--; /* y = y - 1; */
z = --x; /* x = x - 1; z = x; */
a = x++; /* a = x; x = x + 1; */
The increment and decrement operators should always be used by programmers
whenever a change upwards or downwards by one is desired by a C programmer.

37
 Note however that a compiler will order increments and decrements different depending
on which side of the variable the operator is on! In the third example above, the
decrement is on the left-hand side of the variable x. Hence, we first subtract one from the
contents of x, and lastly assign the new value of x to variable z. This is called pre-
increment.
In the fourth example, variable x has an increment is on the right hand side. Here, we first
take the current value of x and assign it to a and lastly increment the value of x. This is
called post-increment.
Both statements are similar in nature, but totally different results will come out of either
statement.

4.9. Putting it Together

The following program, OPERATE.C, utilizes some of the concepts outlined in this
chapter.
#include <stdio.h>

int main (void)

{
int a = 5;
int b = 10;
int c = 0;
long d = 100L;
long e = 200L;
long f = 0;
double g = 3.14159
double h = 20.4;
double i = 0.0;
double j = 0.0;

c = a * b - 5 * (a + b);
c++;
a++;
c -= a;
printf ("a: %d b: %d c: %d\n", a, b, c);

f = e % d;
printf ("d: %ld e: %ld f: %ld\n", d, e, f);
f = (e + 1) % d;
printf ("new value of f: %ld\n", f);

i = g * g + h * h;
j = g * (g + h) * h;
printf ("g: %f h: %f i: %f j: %f\n", g, h, i, j);

return 0;
} /* end main */

38
4.10. Common Beginner Errors
• Watch out for the operators that have both single-occurrence and double-occurrence
uses (e.g. = and = =, & and &&, | and ||, < and <<, > and >>).
• Very important note if you're used to programming in another language: if you're
used to being able to declare variables anywhere in a code block in a language like
C++, you'll find that you can't do this in C. Your variables must be declared before
any non-variable-declaration line.

4.11. Summary
In this chapter, we learned about the different operators that a programmer in the C
language has at their disposal. The different operators allow code to be written which can
manipulate variables to generate the results a programmer desires. The operators include
the standard mathematical binary operators (addition, subtraction, multiplication and
division), along with unary operators (negation, boolean inversion), logical operators
(greater than, equality, inequality, less than, logical AND, logical OR).
We also had a look at the precedence order for C, and noticed that it was similar in nature
to the order of operations in math. Finally, we had a brief look at the assignment
operators and the increment and decrement operators.

39
Chapter 5. Flow of Control
5.1. Introduction
Through the course of the last few chapters, what should be apparent is that a function
provides an implied feature: it temporarily causes your program to alter its flow of
control, by stopping execution of the current function's code whenever a function is
called (e.g. printf()). The function will then take over and execute, and when complete, it
will return control back to where it was called. You can think of the flow of control of a
program as the path a program may take, much like the path a driver may make on a long
driving trip. You can get somewhere by driving along a highway (doing everything in the
main() function), but you usually find more interesting and exciting things if you take a
side road once or twice before getting back on the main highway. Likewise altering the
flow of control within a program allows a programmer to produce, not only more
interesting programs, but also more powerful and useful software.
Apart from calling functions, a C program has a number of other methods of altering the
path a program may take. If you are familiar with other programming languages, you may
be aware of certain concepts such as looping, as well as the ability of making a decision
before proceeding to perform some task. These allow a programmer to offer a choice in
how a program's path will be followed, thus creating the possibility of writing powerful,
useful software with little effort. In this chapter we will begin to explore the methods the
C programming language offers for changing the control flow within a program.

5.2. if Statement
As stated above, one method of altering how a program flows is through the use of a
comparison, or decision. In C, as is the case in many other languages, a comparison is
done by checking the boolean, or true/ false outcome of some expression, and if the
result is true, some section of code will execute, and if false, some other section can be
executed. The syntax for performing a decision in C is as follows:
if (boolean expression)
{
statement;
}

40
Notice that for the if statement in C, the boolean expression to be tested must be within a
pair of round brackets. The if statement in C looks much like a function call but should
not be confused with one.
A very important note: an if statement is not programmed fully until a complete statement
follows the closing bracket of the if portion of the statement. A complete if statement
includes the boolean expression test, as well as a statement to be executed if the
expression is true.
If the boolean expression turns out to be true, the statement following the closing round
bracket will be executed. If false, the statement will not be executed, causing your
program to move to the next line of code in your program.

Useful Tip: Semicolons After if Statements

Make sure that you don't overgeneralize and put a lone semicolon after the closing
bracket for an if statement. The reason for this is that a semicolon is considered a
statement that does nothing! Thus, the semicolon would be the body of code that is
executed when the if statement's condition is true and the real body would always be
executed, no matter what!

The following are a number of examples of valid comparisons in the C language:

if (x > 5)
{
y = 3;
}

if ((x == y) && (a != b))

{
z = 10;
}

The first example above does a check to see if the value of variable x is greater than the
5. If true, the statement y = 3; will be executed. If not, the statement y = 3; will be
skipped, as if it was never in your program at all. This complete if statement tells the
compiler that whenever the variable x contains a value greater than 5, the variable y will
be assigned the value 3.
The second example is little more complicated. You should be aware that even though
this example has a second line z = 10;, this line is still part of the same if statement.
Recall that an if is not complete until you terminate it with a complete statement.
The lines of code between the curly braces following the if statement is called the body
of the if statement. You must indent the body of an if statement by one tab stop to
visually indicate what will execute if the if boolean expression is true. Well, really the
compiler doesn't strictly require it but good programmers do :-).

41
 Course Requirement: Indentation After Decisions
You must indent the body of the if statement to show that a decision has to be made and
the indented code may or may not be executed. This also applies to while, do-while, and
for. The principle is to indent the body after a decision. This also implies that the
decision and body cannot be on the same line.
Useful Tip: Spacing Around Brackets
Use spaces near brackets. You can have one space before, one space after, or both.
Being consistent with this will make your code look more regular and easier to
understand. Being inconsistent will make you look amateurish.

5.3. Logical Operators

You should also notice that this second example if statement is composed of two boolean
expressions linked together using the && binary operator, which is the logical AND
operator. The way the logical AND operator works is that if the first operand evaluates to
true, and the second operand evaluates to true, then the overall expression will be
evaluated to true. If either one or the other expression evaluates to false, then the entire
expression is false.
Above, the first operand for the binary logical AND is the boolean expression x == y.
The operator == is the equality operator, which is another binary operator. The equality
operator (usually read by C programmers as "is equivalent to" or "matches") instructs the
C compiler to generate code to take the current value of the variable x, and compare it
against the current value of the variable y. If they match, a true result is generated, if not,
false is generated.

Useful Tip: Confusion About = and = =

Always remember that = does assignment and == compares equality. Unfortunately, it is
very easy to put the wrong one in and not get any kind of compile error.
You will also run across this confusion with ||, &&, <<, >>.

You should be aware that the limitations of the font used in this document makes this
operator look like two long lines, one above the other. Don't waste your time looking for
the long lines on your keyboard ... it's two consecutive equal signs. Changing font to
make it clearer gives us this: ==.
The second operand in the example is the expression a != b. This expression is an
inequality operation, (another binary operation) in which the value of variable a is
compared to the value of variable b, and if they don't match, true is generated, otherwise
false. The exclamation mark in C is the not symbol, and a C programmer normally reads
!= as "not equal".
With the evaluation of both operands of the && operator, the compiler will generate
code to check these two conditions. As stated above, if both sides are true, the entire

42
 expression evaluates as true, if not, the entire expression is false. If true, the statement z =
10; will be executed, otherwise, nothing will be executed.
As an exercise, let us assume the variables a, b, x, y, and z have the following values (the
2nd example is repeated for ease of discussion):
x = 20;
y = 20;
a = 10;
b = 10;
z = 99;

if ((x == y) && (a != b))

{
z = 10;
}
The logic of the second example above would generate these boolean results:
x == y since x is 20 and y is 20, the result is true
a != b since a has the same value as b, and since we are checking for the condition of
inequality, the result is false
true && false the overall result is false since both sides of the && operator are not true
results
since the overall result is false, the statement z = 10; is skipped, and z remains 99
Repeat the above exercise with the value of b set to 30. Prove to yourself that z will be
set to the value 10 in this case.

5.4. Truth Table

The following table may help you understand the operation of the logical operators, &&
(logical AND), || (logical OR) and ! (logical inversion).

Operator Usage Notes

&& F && F == F if expr1 is boolean true and expr2 is boolean
T && F == F true, then the entire expression is true;
F && T == F otherwise, it is false
T && T == T
|| F || F == F if either expr1 or expr2 are boolean true, then
T || F == T the entire expression is true; otherwise, it is
F || T == T false
T || T == T
! !F = = T if expr is true, then generate false; otherwise,
!T = = F generate true

43
In the table above, the letter T represents true, and the letter F represents false. This table
of Ts and Fs is normally called a truth table, since it lists all the possible outcomes of
boolean equations such as the ones we've see above.
It should be known that in C, if an expression evaluates as non-zero, the result will be
considered true, and if the expression evaluates as zero, the result will be considered
false. Thus, the following examples of if are valid:
if (x)
{
y = 10;
}

if (!y)
{
x = 10;
}
The first example above causes the variable x to be evaluated as a boolean expression. If
x is equal to zero, this will generate a false result, and the rest of the if statement is
ignored. If it is anything but zero, the expression will be evaluated as true, and the rest of
the if statement will execute, setting y to 10.
The second example above checks to see if y is boolean false. In other words, if y is zero
(false), !y will invert this false result into a true result, allowing the body of the if
statement to be executed.
if statements such as the above are not uncommon since they are the short form ways of
writing the following lines, which are equivalent to the above:
if (x != 0)
{
y = 10;
}

if (y == 0)
{
x = 10;
}
The longer form (involving the explicit comparison to 0) is definitely preferred, as it
leaves no room for misinterpretation.

Useful Tip: Boolean Comparisons of Non-Boolean Expressions

Even though you see statements like if (x) every once in a while, you should avoid it
since it can be confusing. It implies that x is a boolean variable when it usually is not.
C++ supports a native bool data type. C does not.

You can combine more than one boolean expression together, using combinations of the
logical operators, as well as using brackets to change the order of operations. For
instance:

44
 if (((x == 3) && (y != 2)) || (z < 5))
{
printf ("wow!\n");
}
Notice how we are checking for the equality of variable x with 3 and the inequality of
variable y with 2. If both expressions are true, then the logical AND expression is true.
Notice, though, that our logical AND expression is logically ORed with the last
expression, z < 5. If either the last expression is true, or the AND expression was true,
then the entire logical expression is true, causing wow! to be printed to your screen.
Tremendous sophistication can be achieved by using logical operators and brackets to
select what sections of your code will be executed.

Useful Tip: Code Clarity

Consistency of style, proper use of brackets, and use of adequate whitespace to help with
the clarity of your code is always more desirable than any decrease in source file size that
may result from not doing so.

5.5. Putting It Together

Now that we've seen how logical operators work in our software, let's have a look at an
actual program which will use the if statement to perform a useful task. Note, to make
this program useful, a number of function calls that are covered later are used to obtain
user input.
#include <stdio.h>
#include <stdlib.h>

int main (void)

{
char buffer[100] = {0}; /* used for user input */
int month = 0;
int day = 0;
int totalDays = 0;

/* ask user to type in the month and convert */

printf ("Enter month: ");

gets (buffer);
month = atoi (buffer);

/* ask user to type in the day, and convert */

printf ("Enter day: ");
gets (buffer);
day = atoi (buffer);

45
/* calculate elapsed days */
if (month > 11)
{
totalDays += 30;
}

if (month > 10)

{
totalDays += 31;
}

if (month > 9)
{
totalDays += 30;
}

if (month > 8)
{
totalDays += 31;
}

if (month > 7)
{
totalDays += 31;
}

if (month > 6)
{
totalDays += 30;
}

if (month > 5)
{
totalDays += 31;
}

if (month > 4)
{
totalDays += 30;
}

if (month > 3)
{
totalDays += 31;
}

if (month > 2)
{
totalDays += 28;
}
46
 if (month > 1)
{
totalDays += 31;
}
totalDays += day;

printf ("Month: %d Day: %d Total Days: %d\n",

month, day, totalDays);

return 0;
} /* end main */

Useful Tip: Getting user input

We're introducing some concepts this example that we'll be re-introducing later. Here's
what you just have to trust me on for now:
- Getting a string from the keyboard can be done by calling the gets() function.
- Strings can be stored in a variable like this: char buffer[100]
- Converting a string to an integer can be done by passing the string to the atoi() function.

This is the largest program that we have studied at to date but it's also a rather useful
program. If you've ever wondered how many days have elapsed since the start of the
year, this program will calculate it.
This program is interesting in its use of the if statement to help us determine how many
days to add to the running total of days from the start of the year.
First off, the computer will ask you to enter both the month and day to work with. Both
should be a simple string of numeric digits (ie: 10 for October), and the program will
convert the user's input into the appropriate data type (in this case, an integer). After
conversion, the integer is stored in the variables month and day.
Next, the program will begin checking to see if the month specified is greater than a
specific month, and if so, it will increment the running total by the number of days in the
previous month. Finally, after running through the entire series of if statements, the day
will be added to the running total, to give an accurate assessment of the number of days
elapsed since January 1 of the year.
The above logic sounds rather convoluted at first, but it is actually quite simple if you
think about why we are adding the total number of days from the previous month.
Let's use an example of March 16. March is month 3. After setting the running count to 0,
we now begin to follow the series of if statements. The first if statement is:
if (month > 11)
{
totalDays += 30;
}

47
 Here we check to see if month is greater than November (month 11). If month was 12
(December), we would have added 30 to the running count. Since we are dealing with
March, this if is false, and we leave totalDays alone.
The next if is encountered, and again, nothing changes since March is not greater than
October.
This continues until we come across:
if (month > 2)
{
totalDays += 28;
}
This statement is asking if the month is greater than February. In our case, March is
indeed greater than February (3 is greater than 2), and because the comparison is true, we
increment totalDays by 28 days (not taking into account leap years - we'll leave that for a
student exercise!).
The next statement asks if month is greater than January, and again, March is greater (3
is greater than 1), and we again increment the counter, this time with 31.
What we've actually done here is add the total number of days in January, with the total
number of days in February, since those are the only completed months for the date
March 16. We still need to add the 16 days in March to our count, which is done by:
totalDays += day;
to arrive at the total number of days since the start of the year.

Try working through on paper other examples of dates to prove to yourself that this logic
does indeed correctly compute the number of days from January 1.

5.6. Chained if Statements

The next topic concerning if statements is the ability to chain multiple if statements one
after another. This is made possible by the fact that the C programming language allows
the programmer to specify what to do if the result of the boolean expression is false. This
is done using the else statement. The syntax for a fully exploited if statement is as
follows:

48
if (boolean expression)
{
statement1;
}
else
{
statement2;
}
If the boolean expression is true, statement1 is executed, otherwise (else) statement2 is
executed. Now we can do two different things based upon the true or false outcome of the
boolean expression.
The idea of chaining multiple if statements comes in handy when doing things like
comparing user input against a series of expected options. For instance, imagine a
program which gives the user a choice of four different options to perform. Once the user
has entered the choice, a check is made against the expected input values, and if they
match, the appropriate option is taken. Otherwise, some sort of error is displayed.
The following example uses a series of chained if statements to perform such an
operation:
if (key == '1')
{
performOptionOne();
}
else if (key == '2')
{
performOptionTwo();
}
else if (key == '3')
{
performOptionThree();
}
else if (key == '4')
{
performOptionFour();
}
else
{
printf ("ERROR: Enter number from 1 to 4!\n");
}
Notice that a chained if statement is simply an if performed as part of the else portion of
the previous if. In the example above, if the key entered is not the key labeled '1', then a
check will be made against key '2', and so on. If it was key '1', then option one will be
performed, and the rest of the if statements will be ignored (since they are only executed
when the first if fails!) Therefore, when the keystroke is indeed '1', the other 3 tests will
not be done as there is no need to, and the chained series of if statements ensures that this
principle is adhered to.

49
 Useful Tip: no elseif
There is no such thing as elseif in C. You must separate the else from the if.

5.7. Compound Statements

There are often times a C programmer wishes to do more than one statement of code if
the result of an if is true or false. We've so far seen that an if statement in C will only
execute the complete statement following the closing round bracket of the if statement (or
the else portion of an if else).
In C, a programmer has the ability to create something called a compound statement,
which is simply a series of statements which fall between an opening brace ({) and a
closing brace (}). The statements in between them are considered to belong to the same
block of code, and C will treat such compound statements as a single entity for the
purposes of an if statement or the else statement. For example:
if (x < 10)
{
printf ("x is less than 10!\n");
printf ("setting y equal to 20\n");
y = 20;
}
else
{
printf ("x is not less than 10!\n");
printf ("setting y equal to 0\n");
y = 0;
} /* endif */
The above example shows two compound statements in action: the first compound
statement prints two messages to the screen and sets the variable y equal to 20. The
second compound statement also prints two (different) messages to the screen, and
assigns the value 0 to y. If the boolean expression x < 20 is true, the first compound
statement will be executed fully (the two printf() function calls and the assignment of 20
into y). If false, the second compound statement is executed fully.
Notice again the style being used for both bracing and spacing: this is a common
compound statement style used in professional C programming, as mentioned earlier.
With compound statements, a programmer has the freedom of being able to do as much
or as little as is required for a particular programming problem.
Even though the compiler doesn't force you to put braces around a single statement that is
used as a body (like was done earlier), it is a good idea to do so. That way, if you add a
second statement to the body, you don't have to remember to put braces around the
newly-compounded body.

50
 Course Requirement: Braces on Single Statement Bodies
You are required to put braces on single-statement bodies in this course.

5.8. Nesting if Statements

The final topic for if statements is the ability to nest if statements. The term "nest" is used
to denote the action of "doing an if inside of another if". An example will make this idea
immediately obvious:
if (y > 20)
{
if (x < 10)
{
printf ("x < 10 and y > 20");
}
else
{
printf ("x > 10 and y > 20");
}
}/* endif */
Here we have a compound statement, and inside of the compound statement we have
nested another if else statement. If the first if statement finds y is greater than 20, then the
second if statement within the compound statement will be executed, and based upon the
value of x, one of two messages will be printed to the screen. This is yet another example
of the freedom and flexibility of C.

5.9. Loops
So far we have looked at the abilities of C when it comes to making comparisons within
our programs. There are many circumstances, however, where simply making a decision
will not always allow a program to perform the desired result. For instance, if you wrote a
program to prompt the user to enter two numbers, and the program multiplied these
numbers and printed the answer, the programmer may want the user to try again. At this
moment, we have no method of being able to perform the action of "trying again".
The first thought some novice programmers have is: "why not just call main() again
before main() is finished?” Although it is technically permitted, this situation leads to an
interesting programming phenomenon called recursion, which is the term given to the
calling of a function from within itself. It sounds like a simple enough idea, but the
mechanism the microprocessor goes through when a function is called causes some of
your computer's memory to be used up temporarily until the function is completed, and
repeated calling of yourself begins to eat up more and more memory, until there isn't any
left. Improperly programmed recursive situations are generally fatal as far as getting
stable and robust software to be generated.

51
 Course Requirement: Recursive Calls of main()
It is typically a very bad idea to recursively call main(). You will lose marks if you do
so.

We are still left with a dilemma: how do we "try again"? The answer lies in the fact that
C, as well as most other computing languages, allows the program to specify where, and
how often, a block of code is to loop. With a loop of some sort, a programmer can
instruct the computer to execute a piece of code 10 times, 100 times, or even forever.
In C, there are three main methods of performing a loop: the while statement, the do-
while statement, and the for statement. We will now have a closer look at the while
statement now.

5.10. The while Statement

The while statement is a looping mechanism in C which appears remarkably similar to
the if statement upon first glance. The syntax of the while statement is as follows:
while (boolean expression)
statement;

If the boolean expression inside of the brackets of the while statement evaluates to true,
statement will be executed. Once statement is complete, the while statement will
perform another check of the boolean expression. As long as the boolean expression is
true, statement will be executed over and over and over. Therefore, the while statement
can be thought of as a simple if statement that has the ability to evaluate the expression
multiple times until the expression is false. Note that the statement above can be a full
blown compound statement just like in the case of an if statement. Note as well that
statement is commonly referred to as the loop body.

52
Let's have a look at a simple example using the while statement:
#include <stdio.h>

int main (void)

{
int x = 0;

while (x < 10)

{
printf ("The value of x is: %d\n", x);
x++;
}/* end while */

printf ("DONE\n");
return 0;
}/* end main */
In the above example, we initialize a variable x with the value of zero (remember initial
values can be given to variables as they are declared). We then begin the while statement,
checking to see if x is less than 10. Currently, zero is indeed less than 10, and because the
boolean expression result is true, we enter the compound statement making up the body
of the loop.
First we print a message informing the user of the current value of x. Next, we increment
x by one. The compound statement is now complete, and the while statement will again
check the current value of x against 10. x should now be 1, thus this boolean expression
is still true, and we enter the loop body a second time.
This loop body will in fact be executed 10 times, printing the following on our computer
screen:
The value of x is: 0
The value of x is: 1
The value of x is: 2
The value of x is: 3
The value of x is: 4
The value of x is: 5
The value of x is: 6
The value of x is: 7
The value of x is: 8
The value of x is: 9
DONE!
Notice that the value of 10 is not printed on the screen. This is because we are expected
to execute the body of the loop as long as x is less than 10. When x is equal to 10, (and
hence, no longer less than 10) the loop will terminate, causing the message DONE! to
appear on the screen. To force the loop to count to 10, use the <= operator in the boolean
expression.

53
 A typical usage of the while statement is to handle the “try again” type of situation. For
example, study the following code snippet:
int done = 0
int key = 0;

. . .

done = 0;
while (done == 0)
{
. . .
if (key == 'Q') /* check for quit key */
{
done = 1;
}
} /* end while */
The while statement checks to see if the variable done is boolean false, and as long as
this variable stays false, the loop will execute. This type of loop is very common in C
programming.
Inside the loop we check, using an if statement, a keystroke pressed by the user. If the
keystroke is equal to the letter 'Q', the program will set the variable done to 1, which is
true in boolean. As soon as done is true, the loop will terminate (you can think of the fact
that done is true as literally meaning that we are done this loop!).

5.11. Pitfalls of Flow of Control

Now that we've seen the various flow of control constructs C provides, it is time to look
at some pitfalls that a programmer may come across when writing software with such
statements.
Study the following statement:
if (x == 3) ;
Remember the previous Useful Tip about the semicolons? Well, the fact is that an if
statement is not complete unless it has the boolean expression as well as a statement and
semicolon. The above example appears to be missing a statement to execute if the
expression is true. Actually, as far as the compiler is concerned, there is indeed a
statement to execute, and that statement happens to be ... nothing! This is an example of
an empty statement: a statement that has no code associated with it. If you wrote this line
into your program, the compiler would generate code to perform the test, and if the
expression were true, nothing would happen! It may appear odd in this case, but it is a
common coding error that occurs when working with flow control statements such as the
if statement. Avoid the temptation of putting a semicolon everywhere in your code –
remember, terminate a statement only when it requires termination, and that should allow
you to avoid the above pitfall.

54
A common problem that novice C programmers run across when programming a loop is
the inadvertent placement of a semicolon after the closing round bracket of a while or for
statement (and the if statement as shown above). As another example, study the following
statements:
x = 0;
while (x < 10);
{
printf ("x: %d\n", x);
x++;
} /* end while */
The above loop is never ending! Can you spot why? Note the semicolon after the closing
round bracket of the while. We've just told the compiler to generate a loop that will
execute a null statement as long as the value of x is less than 10. Since just prior to the
loop we initialized x with a value of zero, (which is certainly less than 10), this loop will
never stop looping, as it never has a chance to move x towards 10! Luckily, some
compilers today detect that such a situation exists, and a warning will be issued, giving
you an indication that something is amiss with your source code.
Another common mistake is to program a loop whose keep-going condition cannot
possibly be met during the execution of the loop body. Study the following example:
y = 30;
z = 0;
x = 0;
while (x < 10)
{
z += x * y;
y++;
}
The intent of the above loop is to sum the product of x (from zero to 10) and y and store
the result in a running counter z. But the error occurs in how we've programmed the
while statement that specifies how we reach termination. Notice that we've stated to
increment y by one each time through the loop. But our keep-going condition says to
break the loop when x reaches 10. In this case, x will never reach 10, and we've got
another endless loop!
Many endless loop problems occur due to typing mistakes by the programmer. Seldom is
there a need to write a program with an endless loop. Because of that, when most
compilers detect that condition, a warning is generated to inform you of this potentially
dangerous situation.

55
5.12. Indentation Revisited
Three bracing styles with if statements, this time
Style 1: Braces level with code block
if( i < 8 )
{
i = 8;
}
Style 2: Braces level with control statement
if( i < 8 )
{
i = 8;
}
Style 3: Braces attached to control statement (K&R Style)
if( i < 8 ) {
i = 8;
}
Choose one of these styles and stick with it. If you need to experiment, keep all code in a
project consistent and change on the next project. Chances are, though, that one style
will seem natural to you and the others will seem alien :-).

5.13. Common Beginner Errors

• Avoid putting a single = in an equality comparison.
• Avoid putting that nefarious semicolon at the end of the if or while statement, unless
you really know that it should be there. If it should be there, put it on a separate line
so that it's very obviously not a mistake.
• Put braces around single-line bodies to avoid problems later.
• When joining conditions using && and ||, make sure that you have chosen the right
one (it's easy to get confused).

5.14. Summary
This chapter has greatly expanded our knowledge of the C programming language with
the introduction of control flow statements. These types of statement allow the
programmer to choose which portions of code will be executed at a given time, through
the use of comparisons and possibly loops as well.
C uses the if statement to accomplish the act of performing a comparison. This statement
will evaluate the boolean outcome of the expression between the brackets of the if
statement, and if it results in a boolean true outcome, the rest of the if statement will be
executed.
We learned that a C programmer has the ability to cause a program to loop until a certain
condition is met. This is accomplished using the while, do ... while and for statements.
56
Care must be exercised to avoid coding situations where an infinite loop may occur - a
loop that has no way of terminating itself.

57
Chapter 6. Pseudocode
At this point, you should be thinking that our programs will be getting to be more
complex. There is a temptation to just sit down in front of a computer and start writing
code without doing some preliminary design. Except in the most trivial situations, this is
a bad idea. Either you'll waste time by having to do and redo your code or you'll write
bad, inflexible code that will cause you to be cursed by your co-workers.
Luckily, there are many methods that you can use to avoid your co-workers’ wrath. One
of the simplest to understand and implement involves using pseudocode. Pseudocode is
an English-language representation of what you have to do with your program to solve
the problem that you've been given. It should be understood that the name "pseudocode"
indicates what it is and what it is not: it is something that looks like code but it is not
something that is code. You can write pseudocode to solve a problem without knowing
anything about a computer language. As a matter of fact, if you think of it as just
describing how you do something (in English), you know how to do it!
Here's an example of simple pseudocode:
Loop while there is input from the file.
Process the input line.
Print the resulting output.
End loop.
There are six basic operations in pseudocode (from Simple Program Design, pages 10 to
12):
• A computer can receive information.
• A computer can put out information.
• A computer can perform arithmetic.
• A computer can assign a value to a variable or memory location.
• A computer can compare two variables and select one of two alternative actions.
• A computer can repeat a group of actions.
This leads to three basic control structures:
• Sequence
• Selection
• Repetition

6.1. Sequence
Sequence refers simply to having executable statements in a row, one executed after
another. There are no decisions to make and no looping. This describes the types of
programs that we've worked with so far.

58
6.2. Selection
Selection refers to making a decision. Once the decision is made, some subsequent code
is executed if the decision result is True. If the result is False, the aforementioned code is
not executed.

6.3. Repetition
Repetition refers to repeating a block of code over and over again. A decision is made,
either at the top or bottom of the loop, determining whether or not the block is executed.
If not, the next line of code after the block (and decision, if at the bottom) is executed.

6.4. Example of Pseudocode

Here is an example of pseudocode for the Rock, Paper, Scissors game (from
Programming Logic and Design by Joyce Farrell, Course Technology, Boston, 2004,
ISBN 0-619-16021-7, page 53):
Start
Ask a friend to play a game of Rock, Paper, Scissors with you
While answer is yes
Extend yourHand and myHand
If yourHand is "paper" then
If myHand is "rock" then
winner = yourHand
Else
If myHand is "scissors" then
winner = myHand
Else
winner = tie
Endif
Endif
Else
If yourHand is "scissors" then
If yourHand is "rock" then
winner = myHand
Else
If myHand is "paper" then
winner = yourHand
Else
winner = tie
Endif
Endif
Else
If myHand is "rock" then
winner = tie
Else
If myHand is "paper" then
winner = myHand
Else

59
 winner = yourHand
Endif
Endif
Endif
Endif
Ask a friend to play a game of rock, paper, scissors
End while
Stop
In this example, we can see many examples of all three types of pseudocode.

Useful Tip: General program structure

Here's an example of general program structure:
1. Do stuff that happens at the end of the program, including declaring and initializing
variables.
2. Perform the main loop.
3. Do stuff that happens at the end of the program.

6.5. Why Pseudocode?

It is well established that time spent in pre-coding design will be rewarded in time
efficiency when producing software. If you design your software with emphasis on what
it does rather than the details of how the language works, your success will be much
greater!
The next chapter will give some examples of working with pseudocode with flow of
control situations. For more detail on using pseudocode, read Simple Program Design.

6.6. Common Beginner Errors

• Always design your software before coding it!!!!!!

6.7. Summary
This chapter has introduced the concept of using pseudocode for design.

60
Chapter 7. Designing for Flow Of Control
7.1. Some Help With Flow of Control
So far, the code that we've had to deal with doesn't require a lot of analysis to produce.
Flow of control issues, though, make you think a bit more. If you're experienced with
doing flow of control in other languages, skip this chapter. If not, though, take a look at
Chapters 4 and 5 in Simple Program Design (by Lesley Anne Robertson, Course
Technology, Cambridge, MA, 2000, ISBN: 0-619-01590-X).

7.2. More General Flow Of Control Design Principles

When analyzing a problem, watch for any indications of actions that result from a
decision being made. That usually indicates the need for an if or while statement.
Sometimes there isn't an explicit decision that is obvious but there is an indication that
there are multiple items being processed in some way. This usually indicates the need for
a loop (e.g. while). The next section will contain some examples of translating problems
into pseudocode.

7.3. Some General Flow Of Control Design Examples

Example 1
Problem Statement
Write a program that prints the sum of prices entered by the user.
Commentary
There are two indicators of a loop here: sum and prices. The fact that prices is plural
indicates that you are dealing with more than one item. Also, a sum of only one item is
trivial, so it should be assumed that we're summing more than one.
If we were told exactly how many items we were dealing with, we wouldn't necessarily
require a loop. Since we don't know, though, we need a loop to handle it.
Solution:
• Initialize sum to 0.
• Loop while there are prices to get.
• Get a price.
• Add the price to the sum.
• End loop.
• Print sum.

61
Example 2
Problem Statement
Write a program that prints "PST Taxable" if a restaurant meal is above $3.99 and "No
PST" otherwise.
Commentary
Here we see the word "if". This is a sure giveaway that we have to make a decision.
This decision is expressed as an if-else construct.
Solution
• Get price.
• If price is above 3.99, then
• Print "PST Taxable".
• Else
• Print "No PST".
• End If.

Example 3
Problem Statement
Write a program that gets restaurant meal prices from the user, printing "PST Taxable" if
the price is above $3.99 and printing "No PST" otherwise.
Commentary
Using the first example, we can see that "prices" indicates a loop. We also have a
decision, indicated by the use of the word "if". We've got to determine whether the if is
within the while or vice versa. The giveaway here is that the decision is on a single price,
not on all of them collectively. Thus, the decision is made on a single price as we're
processing it within the loop.
Solution
• Loop while there are prices to get.
• Get a price.
• If price is above 3.99, then
• Print "PST Taxable".
• Else
• Print "No PST".
• End If.
• End Loop.

62
Example 4
Problem Statement
Write a program that gets restaurant meal prices from the user, printing "PST
Taxable" if the sum is above $3.99 and "No PST" otherwise.
Commentary
Here, we're making the decision on the sum. That indicates that the decision could be
outside the loop, after we've determined the final sum. A slight modification to
the problem would change the solution, though, as we'll see in the next example.
Solution
• Initialize sum to 0.
• Loop while there are prices to get.
• Get a price.
• Add the price to the sum.
• End loop.
• If sum is above 3.99, then
• Print "PST Taxable".
• Else
• Print "No PST".
• End If.

Example 5
Problem Statement
Write a program that gets restaurant meal prices from the user, printing "PST
Taxable" as soon as the sum is above $3.99 and "No PST" if it never is.
Commentary
We've got to be careful with this one, since we're working with the sum, which might
indicate that we need a decision after the loop. The difference, however, is that
we're not dealing with the final sum but instead with an interim sum (or running
total). The interim sum occurs within the loop and only becomes the final sum
after exiting the loop. Thus, we've got to make that decision inside the loop.
The statement of the problem is subtle, though, in that it says "as soon as the sum is
above $3.99". That indicates that we do not want to print the message more than
once. To facilitate that, we will create another variable called over399 that is 0 if
the sum is not over $3.99 and 1 otherwise.
We can't forget the second part of the problem, though. We have to print "No PST" if
the sum is never above $3.99. The only time that we know that is after the loop is
complete. So, we've got to make another decision after the loop, as well.

63
Solution
• Initialize sum to 0.
• Initialize over399 to 0.
• Loop while there are prices to get.
• Get a price.
• Add the price to the sum.
• If sum is above 3.99 and over399 is 0, then
• Print "PST Taxable".
• Set over399 to 1.
• End If.
• End loop.
• If over399 is 0, then
• Print "No PST".
• End If.

Example 6
Problem Statement
Write a program that prints the sum of restaurant meal prices, adding 6% GST to all
meals and 8% PST to individual meals above $3.99. The sum should be printed
as the total of the meals and all taxes.
Commentary
We've got a sum so we've got a loop. We have one decision to make: whether or not
to apply PST. Since we're doing this on individual meal prices, we have to do it
inside the loop. We should note that we do not have to decide whether or not to
apply GST. It's always applied.
Solution #1
• Initialize sum to 0.
• Loop while there are prices to get.
• Get a price.
• Add the price to the sum.
• Calculate GST from the price.
• Add GST to the sum.
• If price is above 3.99, then
• Calculate PST from the price.
• Add PST to the sum.
• End If.
• End loop.
• Print sum.

64
Solution #2
• Initialize sumNoPST to 0.
• Initialize sumPSTApplied to 0.
• Initialize totalMeals to 0.
• Loop while there are prices to get.
• Get a price.
• If price is greater than 3.99, then
• Add the price to sumPSTApplied.
• Else
• Add the price to sumNoPST.
• End If.
• End loop.
• Calculate PST by multiplying sumPSTApplied by the 8% PST rate.
• Calculate total_meals by adding sumPSTApplied and sumNoPST.
• Calculate GST by multiplying totalMeals by the 6% GST rate.
• Print sum of totalMeals, PST, and GST.

Post-design Exercise
Write the C code that corresponds with the pseudocode in the examples above.

More Examples
For more examples, read Chapter 6 in Simple Program Design.

65
Chapter 8. Writing Your Own Functions
8.1. Review of Functions
Earlier, we briefly covered the essential concepts regarding functions. A function is the
way a C program is constructed: indeed, without the function called main(), a C program
could not exist.
A C program, when properly written, will employ many small functions that carry out
specific tasks. These functions break up the flow of control, allowing sub-functions to
execute, alleviating the main() function from doing all the work. In this chapter, we will
look at reasons for building our own functions, and the next chapter will look at some of
the simpler library functions available through our C compiler.

8.2. Why Write a Function?

This is a common question beginner C programmers will ask, as a C program will
faithfully execute every line of code you may have inside of your main() function
without ever a need to call an external function.
One of the most important reasons is that a good C programmer will always try to make
code as modular as possible, localizing important or repeated computations and
procedures within easily maintainable and accessible functions. By creating functions to
do such tasks, we not only reduce the chance of programmer error in typing in repeated
chunks of code, but we also make it simpler to adjust the operation of a task. By simply
changing a single function within our program, rather than perhaps five or six
occurrences of a similar piece of code, we can eliminate much of the maintenance effort
in software development.
Here's an example of a typical repeated computation:
/*
* SECONDS1.C
*/

#include <stdio.h>

int main (void)

{
long hours = 22L;
long minutes = 52L;
long seconds = 11L;
long totalSeconds1 = 0;
long totalSeconds2 = 0;

totalSeconds1 = hours * 60L * 60L + minutes * 60L

+ seconds;

66
 hours = 10L; minutes = 20L; seconds = 30L;

totalSeconds2 = hours * 60L * 60L + minutes * 60L

+ seconds;

printf ("Elapsed time: %ld\n", totalSeconds1 -

totalSeconds2);
return 0;
} /* end main */
The example above shows main() declared to return an int data type. The main()
function has also been declared to accept no parameters, due to the void parameter list.
This means that main() needs no input information for it to operate.
The main() function above is attempting to compute the number of elapsed seconds
between two sets of times. The computation that is repeated is to convert the hours,
minutes and seconds into the number of elapsed seconds from midnight using the formula
h * 60 * 60 + m * 60 + s (60 minutes in an hour, 60 seconds in a minute). This
computation is repeated twice inside of our program. Because of the dual repetition, there
is twice the chance for typing error, or other such problem that can cause our program to
have a "headache". If we wished to repeat this calculation for 30 different times, it will
become exceedingly tedious to enter this formula over and over again.
Note: in the above program, the use of the capital L after the constants is done since
when working with long integers, it is usually a good idea to specify exactly to the
compiler that the constant is a long constant as well. Recall that older, non-ANSI C
compilers would do funny things if you did not specify the L after a constant if assigning
it to a long integer.

8.3. Creating a Function

Study the following rewrite of SECONDS1.C. You should notice the creation of a
programmer defined function called calcSec() which is called with parameters, much like
printf() to calculate the elapsed time.
/*
* SECONDS2.C
*/

#include <stdio.h>

/* the definition of a new function */

long calcSec (long h, long m, long s)

{
long sec = 0;

sec = h * 60L * 60L + m * 60L + s;

return sec;
} /* end calcSec */

67
 int main (void)
{
long totalSeconds1 = 0;
long totalSeconds2 = 0;

totalSeconds1 = calcSec (22L, 52L, 11L);

totalSeconds2 = calcSec (10L, 20L, 30L);
printf ("Elapsed Time: %ld\n", totalSeconds1 -
totalSeconds2);
return 0;
} /* end main */
The function we have created is called twice to calculate the elapsed seconds from
midnight for two different sets of values. Notice that the computation was coded only
once inside of the function called calcSec(), and we called this function twice to perform
this calculation.

8.4. Function Declaration Notation

As we saw in chapter 2, a function is declared using the following notation:
return-data-type name-of-function (parameter-list)
In SECONDS2.C the function calcSec() has been declared to return a long integer to the
caller of the function, and the parameter list specifies that three pieces of information will
be given to this function, the three long integers: the hour h, the minute m and the second
s. calcSec() will take these input parameters and calculate the number of elapsed seconds
since midnight using the exact same formula as in SECONDS1.C.
Notice that inside of the body of calcSec() we are assigning the results of the
computation into a variable called sec. This variable is declared to be a long integer
within the braces defining the body of calcSec(). Recall that this is one of two locations
we can declare variables, and this form of variable is referred to as a local variable. The
term "local" is used to identify the nature of this type of variable: the variable will only
be in existence as long as the function is executing. As soon as the function is complete,
this variable will disappear from existence until the next time the function is called. This
characteristic is called the lifetime of the variable. It is sometimes beneficial to think of a
local variable as a temporary variable, needed only as long as a piece of code is
executing.
In SECONDS2.C, the local sec is created as soon as we begin execution of calcSec(),
and will be removed from memory as soon as calcSec() is complete. Again, think of local
variables as temporary variables to hold intermediate computations of your functions.

8.5. Returning a Value

calcSec() takes the parameters given to it (h, m and s), computes the number of elapsed
seconds and stores the result in sec. We then use the statement
return sec;

68
 to tell the C compiler that we are ready to return a value back to the caller of this
function.
The above concept is vital to the understanding of the usefulness of functions in C.
Rather than doing the computation directly in the main() function as in SECONDS1.C,
we have the function do the computation based upon the parameters specified, and the
function will give the result back to main() to be stored in the appropriate variables.
It may be instructional at this point to think of a function as a method of performing some
operation, and the operation it performs may generate a result. If we ever needed to
calculate the number of seconds that have elapsed since midnight, we now have a
function that can compute this result, and the result can be obtained by calling calcSec()
with the appropriate parameters.
The return value can be thought of as a "special variable" in the CPU, and whenever the
return statement is used to return a piece of information, this "special variable" will be
assigned the value you wish to return. Imagine that return sec; performing code such as:
specialVariable = sec;
The caller of the function has the option of either using or ignoring the return value. In
our example here, the caller (the main() function) is choosing to use the return value.
Notice how we are assigning the return value of calcSec() into two different long integer
variables. Using our "special variable" analogy, these long integer variables are obtaining
new values from this "special variable", and we can now use these values in further
computations.

8.6. Parameters
In the above discussion, we briefly talked about passing parameters to the function
calcSec(). Parameters can be thought of as pieces of input your function can use in order
for it to carry out its work. For instance, the printf() function uses a parameter which is
normally a series of characters in between double quotation marks. When using printf()
to print formatted variables in our programs, we add more parameters after the quoted
string of characters, all separated with commas, to match the formatting codes in the
string. The printf() function is special in the world of C programming since it is capable
of accepting a variable number of parameters when called. Most functions we use in C
require the programmer to supply the correct number of parameters for the function to
succeed in doing its job, and this goes for programmer defined functions as well.
Parameters for a function are actually local variables that just happen to be initialized
with values at the time the function is called. Parameter variables are therefore simply
copies of the information supplied when the function gets called.
For calcSec(), when it is called in the main() function, we are supplying three different
pieces of information, the hour, minute and second. Notice that both times that calcSec()
is called, we are supplying a set of 3 long integer constants as parameters. When
calcSec() begins to execute, it will copy the values that are supplied at the time of the call
in main() into the parameter variables h, m and s.

69
Input parameters for a function do not need to be constants. They can be the result of an
operation, a variable or a constant. Parameters for a function can even be the return value
of another function! Remember that a parameter is simply a local variable that happens to
have its value initialized when the function is called. Have a look at the following
example program cel2fahr.c that gives an example of other types of parameters for a
function that converts Celsius to Fahrenheit.
/*
* cel2fahr.c
*
* Convert Celsius to Fahrenheit
*/

#include <stdio.h>

double celsiusToFahrenheit (double c)

{
return (c * 9.0 / 5.0 + 32.0);
}/* end celsiusToFahrenheit() */

int main (void)

{
double c = 0.0;
double f = 0.0;
double a = 0.0;
double b = 0.0;

f = celsiusToFahrenheit (10.0);
printf ("C: %f F: %f\n", 10.0, f);

c = 15.0;
f = celsiusToFahrenheit (c);
printf ("C: %f F: %f\n", c, f);

a = 5.0;
b = 10.0;
f = celsiusToFahrenheit (a * b);
printf ("C: %f F: %f\n", a * b, f);

f = celsiusToFahrenheit (celsiusToFahrenheit (20.0));

printf ("C: %f F: %f\n", celsiusToFahrenheit (20.0), f);
return 0;
}/* end main */

The main feature of this program is the function celsiusToFahrenheit(), which accepts a
single parameter called c, of type double, and will use that parameter to compute the
Fahrenheit equivalent for that Celsius temperature. Once computed, it will return the
Fahrenheit back to the caller as a double. Notice that this time, we wrote the function

70
 without using a local variable to store the result. This is because all the function is
supposed to do is compute a result and return it. In simple situations like this, there is
really no need to allocate a local variable since all we would be doing is returning the
value stored in that local variable.
In the main() function, we are calling celsiusToFahrenheit() in many different manners,
each using a different method of specifying a parameter for celsiusToFahrenheit() to
use. The first call to celsiusToFahrenheit() uses a double precision constant as the
parameter. Note that to specify a double precision constant, a C programmer normally
will place a decimal point and then at least a zero or other decimal digits. Recall that
older, non-ANSI C compilers sometimes have trouble with floating point constants when
there are no decimal digits, as some compilers will default them to an integer, much like
the long integer problem earlier.
After receiving a copy of the specified parameter, the function will faithfully convert that
Celsius value into a fahrenheit value, return that value, and then we print out the values
using printf().
The next example uses the double precision floating point variable c as the parameter.
We are able to tell celsiusToFahrenheit() to convert the value that c holds into
fahrenheit. Please note that the parameter c in the function celsiusToFahrenheit() is in a
totally different piece of memory than the local variable c in the function main(). The
parameter in celsiusToFahrenheit() is simply a copy of the variable of c in main(), just
like it's a copy of the constant in the previous example.
The next example uses two variables, a and b, each set to their respective values, and
multiplies them as the parameter to the function celsiusToFahrenheit(). After the
multiplication is completed, the result (in this case 50.0) will be given to
celsiusToFahrenheit() as the parameter. Even the result of a computation can be used as
a parameter.
The final example is the use of the return value of celsiusToFahrenheit() itself as a
parameter celsiusToFahrenheit(). We convert 20.0 into fahrenheit, and then use that
conversion as the next value to convert! It may be a rather pointless exercise, but it does
show the power of functions and their parameters: function return values can be used as
parameters themselves, offering absolute flexibility when writing code to perform some
task.

8.7. Function Comments

A style issue has to do with the description of what a function does and what it needs to
do it. This is typically handled using a function comment. A function comment
describes the name of the function, the purpose of the function, the parameters passed and
return values for the function (if any).

71
An example for the previous function is:

/*
* Function: celsiusToFahrenheit()
* Description: Converts a Celsius temperature to a Fahrenheit
temperature.
* Parameters: double c: temperature in Celsius
* Return Values: temperature in Fahrenheit
*/

double celsiusToFahrenheit (double c)

{
return (c * 9.0 / 5.0 + 32.0);
} /* end celsiusToFahrenheit */

This may seem rather trivial but, believe me, you will write more complex functions that
will require more complex comments. Another example, without the function contents:

/*
* Function: parseTicketRequest()
* Description: Takes a string that contains a request for
* a ticket for a flight and parses it. The string should
* have the following format (separated by spaces):
* originatingAirport destinationAirport flightNumber
* month day year
* The two airports are three character long airport codes.
* The month, day, and year are all integers. If all
* fields are present and correct, 1 is returned and the
* values are filled in to the parameters. Otherwise, 0 is
* returned and the contents of the parameters are
* undefined.
* Parameters: char *input: input string
* char *orig: originating airport code
* char *dest: destination airport code
* int *flight: pointer to flight number
* int *month: pointer to month of flight
* int *day: pointer to day of flight
* int *year: pointer to year of flight
* Return Values: 1 if all fields are correct;
* 0 otherwise
*/

Now, you can't possibly tell me that it is easier to determine that information present in
the preceding function comment by looking at the code!

72
 Useful Tip: main() Function Comment
You don’t need a function comment for main() because it’s covered by the description
within the header comment.
Useful Tip: Function Comment Veracity
Of course, it is possible for function comments to become outdated. Make sure that you
change your function comments when the function changes to make the comment
incorrect. Also, take a quick look comparing the function comment with the contents
when you take over possession of an unfamiliar function for the first time.
Course Requirement: Function Comments
You must have function comments for all functions you create except main(). This
comment must be immediately before the function definition.

8.8. Prototyping Functions

With the advent of the ANSI C standard, compiler makers have been adhering to a
principle adopted by the standards committee, and this same standard is used by (and
strictly enforced by) C++ compilers. The standard is in regards to how a C compiler
should interpret a function call prior to the function having been defined.
Imagine a main() function which wants to execute a function called foo(). If the function
foo() is not declared before the main() function in your source code, as has been the case
in the earlier examples in this chapter, the compiler will have to make assumptions about
what the function will return.
All C compilers have been written to assume that a function, which has not been formally
defined prior to its first call in a C source code file, will return an int. In fact, if a function
does not have its return type specified, the compiler will default it to an int for you! This
may be a simple way of saving some typing time for a programmer, but it is a dangerous
programming practice, especially when you call a function that returns something other
than an int.
It is sometimes possible for a C programmer can structure the source code to place all
functions to be called ahead of main() so that the compiler has been able to compile these
functions, and know what the return type of the function is. In this way, there will be no
ambiguity in compilation. But what about cases such as the following (note that this
program has an error purposely embedded to demonstrate prototyping):

73
/*
* FOO.C
*/

#include <stdio.h>

/* here's where the prototypes should go! */

long foo (void)

{
return (test(50));
}/* end foo */

long bar (void)

{
return (foo());
}/* end bar */

int test (int a)

{
long x = 0;

if (a < 10)
x = foo();
return (x * bar());
}/* end test */

int main (void)

{
long a = 0L;

a = test (7);
printf ("a: %ld\n", a);
return 0;
}/* end main */
Although FOO.C is quite cooked in its operation, it nonetheless exhibits the problem
we've been discussing. Notice how we've defined functions foo(), bar() and test() ahead
of the main() function, as we've done in all our other examples. The problem lies in the
fact that the function foo() ends up calling function test(), so, following our earlier
approach, we should have placed test() in front of foo(). But looking inside of test(), we
notice that we end up calling foo() based upon the outcome of a conditional test. Because
we are calling functions prior to them being declared, the compiler will, in this example,
think that the function test() returns only an integer, which will cause problems further
down the road in our software.
This problem is not unlike the old chicken and egg problem - which comes first? The
solution that ANSI C supplies for us is the use of a function prototype. The function
prototype is a replica of the function declaration, except for the fact that the function

74
body is not declared along with the prototype. The main purpose of the prototype is to
declare to the C compiler the return values of the function. As well, it defines ahead of
time the data types of the parameters. Armed with this knowledge, the C compiler would
be able to perform a number of useful checks for us.
♦ it can warn us when we have a return data type mismatch (when we try to return a
data type that the function was not declared to return)
♦ it can check to ensure the data we pass are of the correct data type that the function
expects.
This ability makes ANSI C compilers a real boon to programmers since the compiler can
do a lot of the error checking for us.
The prototypes for the functions in FOO.C are as follows:
long foo (void);
long bar (void);
long test (int);
Note that the prototype almost exactly matches the syntax for a function declaration,
except for formally declaring the names of parameters, and actually defining the function
body. The use of the data type void as the parameter of some of the above functions is
simply an indication to the compiler that, under no circumstances, shall those functions
accept a parameter. This again allows the compiler to perform more error checking on
your code.
A function prototype should be placed in the source code before the first attempt at using
that function. If not, the entire reason for using this error checking mechanism will be
defeated!

Useful Tip #1: Placing Prototypes

Most programmers will try to place their prototypes before any actual function
declarations, and after the header files are included in your program. In the example
FOO.C, this location is where the comment /* here’s where the prototypes should go */
lies.
Useful Tip #2: Making Prototypes
You can have parameter names in the parameter list for the prototype, even though they
aren't used. This means that the easiest way to create a prototype would be to copy and
paste the actual function definition, putting a semicolon at the end of the bracket closing
the parameter list.
Course Requirement: Prototyping
You must prototype all functions that you write except main() (which has its own implicit
prototype). You must also provide void in the parameter list if there are no parameters
(this is different from C++). Failing to do so will lose you marks.

It should be noted that the header files you include, such as stdio.h, or stdlib.h, usually
include their own function prototypes. This is because the standard library functions that

75
 the C compiler manufacturer supplies need to be prototyped just like a programmer
defined function. Because of this, we can be assured that we will never obtain an
erroneous result back from a function simply because the compiler assumed a different
return type that was anticipated.

8.9. Functions That Do Not Return Any Values

There are many circumstances where a programmer may wish to write a program that
does not return a value to the caller of that function. Such functions, of course, will be
declared to return void, meaning that no information will be given back to the caller.
Situations where this may arise are numerous. For example, envision a program which
presents a menu to the user. The menu itself may be programmed to be displayed using a
function called displayMenu(). This function has the job of showing the text of the menu
on the screen, and nothing more. Once all the text is shown, the function will terminate,
and nothing will be given back to the caller of displayMenu().
As the example above shows, functions that return void simply perform some sequence
of statements. This type of function is usually termed a procedure in most other
programming languages, to differentiate it from a function that will return a value. In the
C language, both functions and procedures are one and the same; only the return type
differs.

8.10. Designing Using Functions

Now we have to determine exactly how we should use functions. A common method is
that you start with a statement of what a solution to a problem does and you break it
down in steps (typically using pseudocode). At each line of pseudocode within each step,
you ask yourself “Do I know how to do this easily?" If the answer is "yes", then you
can continue on to the next line. If the answer is "no", then you have to break that line
down into other lines that describe what that line means or does. Each of the lines that
has to be broken down can represent a function call.

Partial Example:
Design a program that allows a student to make their own timetable, given established
course times. It will be a menu-based program that has the following menus:
Main Menu:
1. Display my current timetable.
2. Display a course’s times.
3. Quit.

Current Timetable Menu:

<stuff representing your timetable>
1. Delete course.
2. Return to main menu.
76
 Course Menu:
<stuff representing your timetable>
<more stuff representing course times>
1. Show course.
2. Add shown course section.
3. Return to main menu.

Quit Menu:
1. Save changes.
2. Discard changes.
3. Return to main menu.

Assume that you have easy access to the student’s current timetable and to the course
scheduling info through provided function calls.
Step 1:
Loop
Display Main Menu.
Do Action.
Repeat until done.
Quit
Commentary:
The above is a rather simple statement of a solution to the problem. If you ask yourself
the question “Do I know how to do this easily?", it should be pretty obvious that you
would answer "yes" to the Loop and Repeat lines but "no" to the Display, Do, and Quit
lines. Thus, you would break Display Main Menu, Do Action, and Quit down further in
Step 2.
This would also likely be an indication that the Display Main Menu and Do Action items
should represent functions within our program.
Step 2: Breakdown of Display Main Menu
Display Main Menu:
Display Title
Display each menu line
Prompt user

Commentary:
We start with a label that indicates what we are breaking down (Display Menu Item).
This label could correspond to a function name. Within the body itself, we could

77
probably answer "yes" to the first two lines but "no" to the prompting of the user. Thus,
the prompting gets broken down in another step (probably as another function).
Step 2: Breakdown of Do Action
Do Action:
Choose action
Do specific action

Commentary:
This is pretty simple. We do the action by first determining what we need to do and then
doing that specific thing. Both can be functions that can be broken down further.
Step 2: Breakdown of Quit
Quit:
<fill in here>
Commentary:
We don't know the details so we'll leave this out of our example.
Step 3: Breakdown of Do Specific Action
Do Specific Action:
Did they select item #1?
If so, then do item #1
Etc.
Did they select something else
If so, print error message
Commentary:
This particular breakdown is only one of many ways that we could do it. We would also
have to break down each individual action in another step.
Step 3: Breakdown of Prompt User
Prompt User:
Print prompt
Increase cursor flash rate
Beep annoyingly if they take too long

General Commentary:
It should be obvious that this could go on for a long time, with many steps and much
breaking down of functions. It's much better to put the work in here than having to
redesign your solution later when you discover that it won't work the way it is supposed
to.

78
 Useful Tip: Always Design First
You should always design your solution "on paper" first. Of course, it doesn't have to be
physically done on paper; you could do it in a text editor using pseudocode. For any non-
trivial program, though, going straight to coding is almost always A Bad Idea.
Useful Tip: The Magic Question
At each line of pseudocode within each step, ask yourself “Do I know how to do this
easily?"

8.11. Naming Conventions, Revisited

Here's a naming convention you should use:
• Start function names with a lowercase letter.
• Start variable names with a lowercase letter.
• Start macros with an uppercase letter.
• Put constants defined using #define in UPPERCASE, using underscores to separate
words (e.g. #define SIZE 10).
• Make constants defined using const start with the letter k (e.g. const int kSize = 10;).
• Make function and variable names expressive and understandable.

Useful Tip: Naming Conventions

Always be consistent in your naming. It will save you tons of time in debugging!
Course Requirement: Naming Standards
Take a look at the SET Coding Standards to see what is required.

8.12. Common Beginner Errors

• Always have a main().
• Always have prototypes. When you change your function definition, change your
function prototype at the same time.
• Don't define a function within another function. Function definitions must be
separate from one another.
• Don't put a semicolon at the end of your function definition.
• Always design before coding!!!!!!

8.13. Summary
In this chapter we learned about the concepts needed to enable a programmer to define
their own functions. A function can be used to perform often-repeated programming
tasks, as well as making a program modular, allowing much simpler maintenance later

79
down the road. We also looked at how to declare and use the return value of a function,
as well as how to declare and use the parameters for a function.
We also looked at the concepts regarding function prototyping, which allow a C compiler
to be informed of the return value data type as well as the data types of the parameters for
a function. An ANSI C compiler can then perform a large amount of error checking for
us.
Finally, we looked at the need for designing before coding. Functions tend to fall easily
out of designing on paper, making your life easier.

80
Chapter 9. Rules about Variables
It is useful at this point in our study to look at the proper rules regarding how variables
are allocated and accessed in the C language. Because C is a compiled language, there
must be no ambiguity about how memory is accessed, or where it will be allocated.

9.1. Scope
The scope of a C variable is defined as "from where" a C variable can be used from. The
declaration of a variable or function is active over a certain region of the source code text,
which is known as its scope. The scope of a variable can be thought of as where you can
access a certain variable, and where you can't. For functions, it can be thought of as when
the C compiler knows the return type of a function, and the exact parameter lists for the
functions.
In this text, we have seen two styles of C variables: global and local. If you recall, a
global variable is a variable that is defined outside of a function declaration, and can be
accessed by all functions within the C program. Certain programming constructs become
quite easy to master because of this feature: for example, a globally defined character
buffer for string input is not only useful, but desirous in some circumstances since you do
not usually want multiple instances of a large input buffer eating up the memory space
your program has to work with. A table of values, such as the number of days in each
month, is another example of a useful global variable.
A local variable, on the other hand, is a variable that is declared internally to a function.
This class of variable includes parameters for a function, as well as the locals declared
within a function. The big advantage locals have over globals is that they will never
conflict with your global variables even if they have the same name. A local variable
called counter will never interfere with a global variable called counter. If you have one
hundred functions inside of your C program, and you declare a local variable for each of
these 100 functions called counter, none of those 100 instances of counter will affect the
globally defined variable counter, or each other for that matter.
To study the concepts behind scope in C, let us use the following program, which
declares a number of globals, and uses them:
/*
* SCOPE.C
*/

#include <stdio.h>
#include <stdlib.h> // for atoi()

int a = 0;
int b = 0;
char buffer[100] = {0};

int test (void);/* prototype */

81
int main (void)
{
long c = 0L;

printf ("Enter a number: ");

gets (buffer);
a = 5;
b = atoi (buffer);
c = (long)a * (long)b;
printf ("a: %d b: %d c: %ld\n", a, b, c);
return 0;
}/* end main */

int test (void)

{
int c = 0;

c = a * b;
return c;
}/* end test */
The above program declares a number of globals, and inside of the two functions main()
and test(), a local variable by the name of c is being declared. The c in test() will never
conflict with the c in main() because of the scope rules in effect: a local is hidden to all
functions except the one that is currently executing.

Useful Tip: Naming Repeated Variables

It is a good practice to use the same variable names for local variables across different
functions. If you think about it, it wouldn't make much sense to have a counter
represented by variables names i, j, count, counter, n, x, and y in different functions
(since they don't overlap and interfere with each other).

The scope characteristics for C variables are fairly simple:

• Globals are accessible throughout your program at any time. These variables are
created as your program starts and the memory occupied by these variables is freed
when the program terminates.
• Locals are only accessible as long as your C program is executing a function that has
that local variable defined. These variables are created as your program begins a
function (or any block of code, for that matter) in which the variable is declared. As
soon as that function (or code block) completes, the local variable ceases to exist.
• A local variable with the same name as a global variable has precedence over the
global variable.

82
9.2. Lifetime Rules
Along with the concept of scope comes the idea of the lifetime of a variable. The lifetime
of an object is defined as ”the period of time during which the object is allocated
storage”. There are a number of classes of lifetime as well.
Firstly, static lifetime refers to variables that exist as long as the program is running. All
global variables are static lifetime variables since they are allocated memory when the
program begins and retain that memory until the program terminates.
Next, local or automatic lifetime refers to variables that are created when a compound
statement begins and are destroyed when a compound statement ends. All local variables
and parameters to a function are created as local lifetime variables, as well as variables
defined at the start of a compound statement such as a for or while loop. As an example:
for (x = 0; x < 5; x++)
{
int y = 0;

y = x * x;
printf ("x: %d x squared: %d\n", x, y);
} /* end for */
The variable y is created inside of the open and close braces for the for statement, which
is its lifetime. The scope of this variable is only between those braces as well, thus y is
invisible and inaccessible outside of the for statement. y is created when the for
statement begins to loop, and is destroyed when the for statement is complete. This is
exactly how local variables for a function operate.

9.3. Static Variables

A local can be forced to be a static variable by simply pre-pending the type modifier
static in front of the type declaration. For example:
void foo (void)
{
static char x = 0;
. . .
} /* end foo */
declares a statically allocated byte of memory called x which exists throughout the entire
time your program runs, but is only accessible during the time its scope rules are in
effect.
A static local, although given memory at the time your program begins, is still only
accessible when the function that declared it is executing. The advantage of this is that
you can save the results of local computations in such statics, and only have particular
functions access them rather than all functions!
Often a programmer may want to count how often particular functions are being called.
Rather than declare a number of global variables such as countFunction1,
countFunction2, and so on, the programmer can simply declare a number of static local

83
 variables, all with the name count. Because they are local, even static locals with
identical names will never conflict because of the scope rules for local variables.
Global variables and functions too can be declared to be static, and the usefulness of this
stems from the fact that the C compiler will make the lifetime of such a function or global
valid only inside of that source code module. You can declare a static global x inside of
test.c, and that global x will only be useable inside test.c. If you linked test.c with foo.c,
and foo.c had it's own static global x, you will find that there are two separately
accessible global variables called x inside of that program!

9.4. The Processor Stack

It should be noted here that local lifetime variables are created on the stack of the CPU
executing the program. The stack is a section of memory set aside where values are
placed and taken off of in a stacking method: whatever is placed (pushed) on the stack
last will be the object that is removed (popped) first, much like making a stack of
magazines on a table: the last one placed on top is the first one removed. The typical
scenario that is followed by a C program when a function call is made is given by this
pseudocode:
push the last parameter onto the stack
push the second last parameter on the stack
. . .
push the first parameter on the stack
push the code address of the next line of C code
to be executed after the function is complete
call the function

the function will create a frame pointer to point to where

the stack variables are currently stored
(the parameters)
the function will move the stack pointer to allocate enough
space for the local variables

the function now begins

. . . (code for function)
the function now ends
the stack pointer is returned to where the frame pointer
was pointing (clearing out the locals)
the return address is unstacked, and jump there

we are now at the line following the function call:

the C program must unstack all parameters to
"clean" the stack after the function call

It is obvious that a rather large amount of work is accomplished by a simple call to a

function. This is the overhead of a function call, as alluded to earlier in this chapter. You

84
 can see how the lifetime rules are utilized by the above scenario: the local variables are
created on the processor stack each and every time the function is called, and when the
function is complete, the stack is cleared out and the variables disappear until the next
invocation of the function.

9.5. Memory Layout

As a final note, a brief discussion of a typical C program's memory layout is in order:
high memory
--------------------------------
end of heap

. . .

start of heap
--------------------------------
start of stack

. . .

end of stack
--------------------------------
end of global/static variables

. . .

start of global/static variables

--------------------------------
end of code

. . .

start of code
--------------------------------
low memory
In the diagram above, memory grows in address from low memory (address 0) to high
memory (address dependent on the computer system). Thus, the start of code is lower in
memory (in address values) than the end of code.
The term heap refers to memory reserved for dynamic allocation of data types and
structures, to be covered in a more advanced text. It should be noted that as your program
uses more and more heap space, less memory is available for dynamic allocation of data.
This is not really a concern for small C programs, but is of great importance to programs
such as database managers that must handle the heap space carefully to maximize the
amount of memory available for data inside of the database.

85
 The stack is a curious entity - it is used to house the return addresses and local variables
for the program - but if you notice in the above diagram, the stack is 'upside down'
compared to the rest of the memory segments. Stacks grow 'downwards' in memory, due
to how the internal architecture of stack handling is designed in virtually all
microprocessors.

Useful Tip: Preserving Stack Space

The stack is usually a lot smaller than other memory areas. Make local array variables
static if they are large in order to avoid using up valuable stack space.

9.6. Common Beginner Errors

• Don't give local variables the same names as global variables.
• Don't use inconsistent naming styles for variables (e.g. cursor_x and cursorX). You
might end up with two variables with similar names while thinking that you've got
only one.

9.7. Summary
We then had a discussion about the scope rules and lifetime rules of different classes of
variables in C. Scope defines where a variable will be accessible, and lifetime defines
when variables will be created and destroyed.
Finally, we took a brief look at what exactly happens when a function call is made on a
typical microprocessor. This helps explain some of the scope and lifetime rules discussed
above. A look was also made of the typical memory model of a microprocessor-based
computer, which should give a programmer a better feeling of how memory is utilized in
the computer.

86
Chapter 10. Other Control Statements
10.1. do - while Statement
The do - while statement is very similar in nature to the normal while statement, except it
always allows at least one iteration (run) through the loop body before performing a
check on the boolean expression. The syntax for a do - while statement is as follows:
do
statement
while (expression);

Note that the do and while keywords are separated by the loop body. The loop body will
continue to be executed as long as the expression evaluates as boolean true. Again, it
should be stressed that with this form of the while loop, the check of the boolean
expression is done at the end of the loop, thus allowing at least one run through the loop.
Depending on the particular programming problem, this form of the while loop may be
more efficient than the standard while loop. Most other programming languages refer to
this looping construct as a repeat until loop, since you will repeat (or do) the loop body
until (or while) some condition causes termination.

Useful Tip: Repeat-until

The difference between a do-while and a repeat-until is that the conditions are reversed.

The following is a rewrite of the while example from the previous chapter, showing how
the same thing could be accomplished using a do - while statement:
int done = 0;
int key = 0;
. . .
do

87
{
. . .
/* check if quitting */
if (key == 'Q')
{
done = 1;
}
else
{
done = 0;
}
} while (done == 0);
Notice that this time, we did not set done ahead of the loop since there is no need to
"fudge" the condition as the condition is not checked until the end of the body of the
loop. Since in both cases we wanted to execute the loop body at least once, the second
example is a somewhat better approach for the "try again" type situations.

Useful Tip: Bracing do-while

No matter what bracing style you use, it is acceptible to put the while on the same line as
the closing brace.
Useful Tip: do-while Semicolons
Make sure you don't forget the semicolon closing the do-while.

88
10.2. The for Statement
The final looping statement we will study is the for statement. This is the most complex
of the three looping constructs, but it is one of the most powerful. The syntax of the for
statement is as follows:
for (initialization; boolean-condition; how-to-get-there)
statement;

Notice that for the for statement, there are three distinct parts within the round brackets:
an initialization section, a boolean condition section, and a section which specifies how
we are to arrive at a termination situation. The three sections are all separated using
semicolons, and if any of the two semicolons are missing, the for statement will not
compile correctly. You can, however, leave out any of the sections.

Let's have a look at a typical for statement, and dissect its operation:
for (x = 0; x < 10; x++)
{
printf ("value of x: %d\n", x);
}
In the above for statement, we can see the three sections of the for in use. Firstly, we
initialize the value of the variable x to zero. The initialization instruction within the for
statement is executed only once at the start of the loop, thus x will be set to zero only
once.
Next we have the boolean condition. The C compiler will generate code that will cause
the loop body to execute as long as this condition is not false. In our example above, this
for statement will loop as long as the value of x is less than 10. As soon as it is 10 or
above, this condition is false, and we stop the loop. This condition will be checked
exactly once prior to each execution of the loop body. The first check is done as soon as
the initialization is complete, thus a for statement may not always execute the loop body.
Some programmers like to refer to this boolean condition as a termination condition,

89
 because as soon as it is false, the loop will terminate. However, it should be noted that the
actual statement of the condition doesn't indicate when the loop should terminate but
instead when it should not terminate. Thus, you might want to refer to it as the keep-
going condition instead.
The last part of the for statement is the specification of what you change to stop the loop.
In the above example, we will reach the termination of the loop by incrementing x by one
each iteration (repeat of the loop). This increment will be done exactly once after each
run through the loop body, and just prior to the next check of the boolean condition. The
above loop will execute the printf() statement exactly 10 times, printing the values of x
from 0 through 9 inclusive, since when x is equal to 10, this will break the boolean
condition.

Useful Tip: Semicolons After Other Control Statements

Remember the useful tip about the semicolon and the if statement (that is, don't put a
semicolon after an if!)? Well, it applies to while, do-while, and for as well.
NOTE: That doesn't mean that you should never put a semicolon after one of those
statements. It simply means that you should only put it in if you truly want the program
to do nothing if the condition is true.
Useful Tip: Difference between C and C++ with for statements
Unlike C++, you cannot declare a variable in the for statement itself in C.

10.3. Putting It Together

The following program, called MULT.C will ask the user to enter which times table they
want a printout for, and a for loop will be employed to generate the table for values from
0 through 12. Please note it uses some new functions, and new data objects that we
haven't yet seen, but will be discussed in the next chapters.
#include <stdio.h>
#include <stdlib.h>

int main (void)

{
static char buffer[100] = {0}; // keep it off stack
int x = 0;
int timesTable = 0;

/*
* get user to enter the times table, and
* convert it to an integer for later use
*/

printf ("Which times table do you want: ");

gets (buffer);
timesTable = atoi (buffer);

90
 /*
* output times table via for loop
*/

for (x = 0; x <= 12; x++)

{
printf ("%d x %d = %d\n",
timesTable, x, timesTable * x);
} /* end for */

printf ("\nDONE!\n");
return 0;
} /* end main */
In the above program, notice that after we've obtained user input, the for statement will
run from the values 0 through 12 inclusive, allowing a printout of the appropriate times
table to be generated. This is typically how for statements are programmed: if you need
to run through a specific number of loops, simply code the number of loops as part of the
initialization and boolean conditions, and you are guaranteed to execute that many times.
One final word about while, do ... while and for loops: just like their cousin the if
statement, these loop constructs can be nested inside of each other too. You are capable
of having a loop within a loop if the programming problem you are solving demands such
a solution.

10.4. Other Features of Loops

There are two special keywords that C provides that allow us to manipulate loops out of
the ordinary manners we seen thus far.
Often, a programmer may wish to abort a loop early, without having to complicate the
conditional expression that controls the loop. This is done using the statement break,
which instructs the C compiler to generate code that will leave the loop if this statement
is executed. For example:
for (x = 0; x < 1000; x++)
{
...
if (key == 'Q')
{
break;
}/* end if */
}/* end while */
The above loop will execute normally as long as the user does not press 'Q' on the
keyboard. Once the user has done this, the body of the if will execute the break
statement, causing the loop to terminate. In C, early or abnormal exits from a loop are
possible, and usually cases of extraordinary circumstances (such as fatal errors) can be
handled with ease to allow loop termination.

91
 The second keyword is continue. This is the almost the opposite of break in that it
causes the loop to move to the next iteration faster than would normally have occurred.
Again, an example will make this concept clear:
for (x = 0; x < 5; x++)
{
if (x == 2)
{
continue;/* skip value 2 */
}
printf ("x: %d\n", x);
}/* end for */
The output of this program appears as such:
x: 0
x: 1
x: 3
x: 4
As the comment suggests, the use of continue above will cause the loop to essentially
"ignore" the case where x is equal to 2. A programmer may write a loop to run through a
list of values (as above) and skip the unwanted values, rather than trying to program a
complicated set of loops to maneuver through a series of disjointed values.

Useful Tip: break and continue in if statements

break and continue break out of loops. They do not have any effect on if statements.

10.5. The switch Statement

We will next examine the flow of control statement switch. In this chapter, we introduced
the concept of being able to chain if statements together, which works quite nicely in
being able to detect a specific value in a variable. When a specific value is found, an
appropriate piece of code may be executed. This type of situation lends itself quite nicely
to the programming of menu type situations, where a menu is presented to the user, and
the user must enter their choice. The choice is scanned using a chained series of if
statements until all choices have been exhausted. If a choice is found, the appropriate
code can handle that choice, otherwise an error can be generated for the user.
The switch statement allows us to program such sequences without resorting to a long
line of if ... else statements. The switch statement allows us to compare specific cases
against the value of a variable the switch will operate upon. This flow of control
statement allows a programmer to list the cases in a very easy to read and understand
manner.
The switch statement has the following notation:
switch (variable)
{
case CONSTANT1:

92
 statement1;
break;

case CONSTANT2:
statement2;
break;

. . .

default:
statementDefault;
break;
} /* end switch */
Notice that the switch outlines the possible cases a variable is expected to take on, giving
a simple, bird's eye view of the possibilities that exist for the different expected values.

The constant that comes after the case keyword is a constant that represents a value
expected. If the programmer were expecting the value 5, the case would read as:
case 5:
The statements that come after the case comparison are executed if the comparison works
out true. That is if, say, CONSTANT1 matches the current value of variable, whatever
statements come after the case will be executed. One statement or 100 statements may
come after the case line.
The individual cases are terminated using a break statement, much like a loop was
broken using break. Because a case is terminated using the break, a compound
statement is not needed for multiple lines of code for a particular case.
You can have as many case statements as you like to ensure that all possible conditions
are checked for. This format is much more compact and readable that a series of chained
if else style comparisons, although the code generated is almost identical.
You will notice that we have a special case at the end of the switch, the default case.
This is the case that is executed when all others fail, and is a convenient way of saying "if
it ain't any of the ones above, then do this one!". In the example of a menu, the default
case can inform the user they've made an incorrect choice. A programmer may wish to
leave out the default case if there is nothing special to be done on unexpected values.

93
Useful Tip: Location of default
It is usual but not required to put a default case at the end of a switch. If it would make
more sense elsewhere, you can relocate it. An example of when it would make more
sense elsewhere is when your default is shared with another case and you have a desired
order for cases (e.g. alphabetical and your default has the same behaviour as pressing 'b').

It is interesting to note that if a programmer wanted to do the same code for more than
one case of a variable, the programmer simply puts more than one case one after another.
For example:
switch (x)
{
case 1:
case 2:
case 6:
/* do the code here */
statement1;
break;

default:
/* do default code here */
statement2;
break;
} /* end switch */
In the example above, if x is equal to either 1, 2 or 6, then statement1 will be executed.
This is a simple method of being able to check for multiple values that are expected to
use the same C code.
Another feature of switch is that if two or more cases do virtually the same work, except
one case has a slightly different action than the other, a programmer can allow cases to
"fall through" to the next case in the list. For example:
switch (x)
{
case 0:
y = 5;
/* fall through to case 1 */
case 1:
z = 10;
break;

case 2:
z = 20;
break;
} /* end switch */
Notice that in case 0, we set variable y to 5, but there is no break statement. Because of
the missing break, we automatically "fall through" to the next case (without any
comparisons!), where variable z is set to 10. We then hit a break statement, causing the

94
 switch to terminate. Thus, the only difference between cases 0 and 1 is that case 0
happens to set y to 5. If you find that your programming cases are identical except for a
few lines of code, the above approach may be beneficial.

Useful Tip: Switch Fall-through Situations

You should put a comment whenever you have a fall-through situation, so that anyone
looking at your code knows that you meant to leave the break out. It is a good idea to
explicitly mention the case that you are falling through to in case you (or some
nepharious wrongdoer) decides to put another case after this one instead.

That all being said, there are limitations to what you can compare using a switch
statement. Some general rules are that the items must be able to be compared using the
equality operator, that your case values must be constants, and that you cannot compare
floating point numbers.

Useful Tip: Switch Indentation

There is quite a bit of controversy over how to properly indent a switch statement. There
are arguments for having the case statements further indented from the switch statement
or not. There are also arguments for indenting the code that comes after the case
statement beyond the standard single indentation from the switch. The best thing to do is
to choose a style and stick with it.

10.6. The Evil goto Statement

There is a goto statement in C. The language is set up so that you should never have to
use one, though. If you really, really, really feel that you have no way around it, examine
the problem carefully and see if it couldn't be done using the other control flow constructs
(e.g. if, while, do, for, switch, break, continue) in conjunction with a variable to allow
you to get out of where you don't want your code to be. This is most common when
trying to break out of nested loops or loops that contain switch statements.

Course Requirement: goto

Never use a goto statement. If you do, you will lose TONS of marks.
Useful Tip: goto
The use of goto may also cause the end of the world as we know it.

10.7. Common Beginner Errors

• Don't forget the semicolon at the end of the do-while.
• Don't forget that a for statement must have exactly two semicolons within the round
brackets.

95
 • When you create a switch statement, make sure that the break statements are
everywhere where they should be. An easy way to do this is to instinctively type the
break statement immediately after you type the line that contains the case statement.

10.8. Summary
The use of the do - while, for, break, continue, and switch statements was introduced.
The use of the goto statement was introduced in order to make you not use it.

96
Chapter 11. Standard Library Functions
Previously, we learned the intricate details of how to write our own functions. Now we
will explore some of the most widely used library functions that come with virtually
every C compiler, on virtually any operating system platform.
A library function is a function that has been previously written by another programmer,
and has been added to a file called a library file, for the sole purpose of being included
with the code a programmer may be writing.
The use of library functions by a programmer frees the programmer from "reinventing the
wheel", so to speak. Since most of the functions supplied as part of the libraries for C
compilers are general input, output, file, math and string handling functions, a
programmer can concentrate on solving a programming problem from a higher level,
writing less low level code.
Library functions are compiled at the company that produces the compiler, and placed
into a library file for linking with the code a programmer will produce. The library file is
therefore a collection of object files that contain functions that allow manipulation of
input, output, files, math and strings.
As you should recall, the purpose of the linker program is to take the object file output of
a compiler and link it in with other object files, as well as whatever library files are
necessary to create a complete executable file.
Generally, compiler installation software will copy the most used libraries on your
system when setting up your compiler, allowing you to have a good selection of
predefined functions at your disposal. We have already seen one of these functions in
action: printf(). This is the standard library function supplied with all C compilers to get
information from inside of your software to the outside world through the computer
screen. There are other library function variations on printf() that allow output to go to
the printer, modem, or even a file. Similar library functions exist to get information from
the keyboard into your software. A programmer, unless absolutely necessary for speed or
efficiency purposes, will never need to write a line of character based input/output low
level code since just about everything needed has been predefined in the standard
libraries.
We will now begin a look at the various types of library functions available grouped
according to their purposes. This chapter will introduce the various functions using the
function prototype format as described in the previous chapter. The header files needed
will also be given, as well as a brief programming example that uses the named functions.
The two groups of functions we will look at here are the input/output oriented functions
as well as the math-oriented functions. Other library functions will be introduced in other
chapters as the need arises. Please note that there are some data types used in the
following discussion that have not yet been introduced, but should not affect the general
understanding of the functions being demonstrated. Input Functions

97
The first group of library functions to study is the input functions.
Name:
char *gets (char *memory);
Header:
stdio.h
Description:
The gets() function is used to obtain a string of characters from the keyboard, and place
them into memory in a character buffer (see the chapter on arrays) pointed to by
memory. Although we have not yet seen what a character buffer/pointer are, as well as
the strange use of the * (see the chapter on pointers) in the above prototype, it is a very
simple function to use if you follow the example below as a guideline on how to use this
function. If successful, it will return memory, otherwise it will return NULL (defined in
stdio.h).
#include <stdio.h>

int main (void)

{
char buffer[100] = {0}; /* declare input buffer */

printf ("Enter a string and hit <enter>: ");

gets (buffer);
printf ("The string the user has entered is: ");
printf (buffer); /* yes, you can do this */
return 0;
} /* end main */

Note: gets()'s biggest downfall is that the function has no knowledge about how big the
buffer is in terms of bytes. A much more robust function to use is fgets().

Useful Tip: Visual C++ Deprecation / Unsafe Warning

If you are compiling the above code using Visual C++ 2005 or 2008 (or newer), you may
get a warning or error saying that the gets() function is deprecated or that it might be
unsafe. That means that Microsoft has decided that you shouldn't use gets() any more
but that it will keep gets() support in until they decide to take it out. To eliminate this
warning for this and other such functions (and you will get it a lot), use the following line
after your #include statements:
#pragma warning(disable: 4996)

Name:

98
fgets(char string[ ], int n, FILE *fp)
Header:
stdio.h
Description:
fgets() gets a line from a file, similar to what gets() does with the keyboard. However, if
you use the special variable stdin as the file, you can get input from the keyboard instead.
The second parameter gives you length protection since it reads at most n-1 characters (it
knows to leave room for the null-termination (explained in the section on strings)). If,
however, it reads less than n-1 characters, it leaves the '\n' in the string as an indicator.
To get rid of it, use:
if( strlen(buf) != 0)
{
buf[strlen(buf)-1] = '\0';
}
Modifying the example from gets() is easy:
#include <stdio.h>

int main (void)

{
char buffer[100] = {0}; /* declare input buffer */

printf ("Enter a string and hit <enter>: ");

fgets (buffer, 100, stdin);
printf ("The string the user has entered is: ");
printf (buffer); /* yes, you can do this */
return 0;
} /* end main */

Name:
int getch (void);
int getche (void);
Header:
conio.h
Description:
These functions will extract a character directly from the keyboard, meaning no buffering
of characters will be done. getch() will get a character without echoing that character on
the screen, whereas getche() will do the same but place the character typed on the screen
as well. These functions will wait until a character is typed before returning with the
ASCII value of that character.

99
Two important notes: 1) These are often not portable to other operating systems. 2)
Visual C++ v6 does not handle getch() and getche() very well when combined with other
I/O functions.
#include <stdio.h>
#include <conio.h>

int main (void)

{
int key = 0;

printf ("Enter a single character: ");

key = getch();
printf ("The key entered is: %c", key);
return 0;
} /* end main */

Name:
int atoi (char *string);
long atol (char *string);
double atof (char *string);
Header:
stdlib.h
Description:
These functions will take the character pointer parameter and convert the string of
characters into an integer (by atoi()),a long (by atol()), or a double (by atof()) for use in
mathematical computations. This is necessary since math cannot be done on a string of
raw characters typed in from the keyboard. The characters must be converted into the
appropriate data type before any mathematical operation is performed.
#include <stdio.h>
#include <stdlib.h>

int main (void)

{
char buffer[100] = {0};/* buffer of 100 chars */
int x = 0;

printf ("Type in a number & hit <enter>: ");

gets (buffer);

/* convert string into an integer */

x = atoi (buffer);

printf ("The number is: %d\n", x);

printf ("The number times 5 is: %d\n", x * 5);

100
 return 0;
}/* end main */

Useful Tip: atof() Has a Misleading Name

Even though atof() implies that it produces a float, it doesn't. It produces a double. If
you don't include the appropriate include file so the prototype is available to the compiler,
your result will be wrong.

11.1. Mathematical Functions

This section will introduce some of the most widely used mathematical functions that the
standard libraries of C provide for the convenience of the programmer. Advanced
mathematical operations such as square roots have to be programmed in C because there
is no such operator in the C language.
Prototypes will be given for most of the commonly used math functions. All math
functions require the header file math.h to be included with your C source code. Also, all
math functions require double precision floating point parameters and will return double
precision floating point values.
/* trigonometric functions */
double cos (double x);
double sin (double x);
double tan (double x);
double acos (double x);
double asin (double x);
double atan (double x);
/* square root */
double sqrt (double x);
/* logarithms */
double log (double x);
double log10 (double x);
/* base raised to the power of exponent */
double pow (double base, double exponent);
An example of a few of these functions:

#include <stdio.h>
#include <math.h>

int main (void)

{
double a = 0.0;
double b = 0.0;
double c = 0.0;
double d = 0.0;

101
 a = 33.67;
b = 1024.567564;

/* take square root of b */

c = sqrt (b);

/* raise a to the power 5 */

d = pow (a, 5.0);

/* print results */

printf ("sqrt of %f is %f\n", b, c);

printf ("%f to the power of 5 is: %f\n", a, d);
return 0;
} /* end main */

Name:
int abs (int x);
long labs (long x);
double fabs (double x);
Header:
math.h (and stdlib.h for abs() and labs())
Description:
The above functions will change their respective parameters into their absolute values.
An absolute value means that if a number is negative, the number will be changed to its
positive value, and if it is already positive, nothing happens.
#include <stdio.h>
#include <stdlib.h>

int main (void)

{
int x = 0;
int y = 0;

x = -33;
y = abs (x);
printf ("x: %d y: %d\n", x, y);
return 0;
}/* end main */

11.2. String Handling

Name:

102
size_t strlen (char *s);
Header:
string.h
Description:
This function will compute the length of a null terminated string. It returns it as a size_t
item, which is typically the same as a primitive data type. Which primitive data type
size_t is equivalent to depends on the system. Under Visual C++ v6, size_t is equivalent
to an unsigned long, indicating that strings can be over 4 billion characters long.

Name:
int strcmp (char *s1, char *s2);
int strncmp (char *s2, char *s2, int n);
Header:
string.h
Description:
The first function will compare the ASCII contents of two strings, pointed to by s1 and
s2, and return a value based upon the following table:
o < 0 if s1 is less than s2
o 0 if s1 is identical to s2
o >0 if s1 is larger than s2
The second function will compare at most n characters before returning a result.
Name:
char *strcpy (char *s1, char *s2);
char *strncpy (char *s1, char *s2, int n);
Header:
string.h
Description:
The first function will copy the bytes pointed to by s2 to the memory pointed to by s1. A
programmer must ensure that the memory pointed to by s1 is dimensioned large enough
to hold the bytes pointed to by s2 and its null terminator.
The second function will copy at most n characters before returning. Both functions will
return s1.

Name:
char *strcat (char *s1, char *s2);
char *strncat (char *s1, char *s2, int n);

103
 Header:
string.h
Description:
The first function will concatenate the characters pointed to by s2 to the end of the string
of characters pointed to by s1. See the warning about s1 from the previous function.
The second function will concatenate at most n characters before returning. Both
functions will return s1.

11.3. Character Classification Functions

Name:
int isalnum (int c);
Header:
ctype.h
Description:
This is a macro that will return TRUE if the ASCII character c is between A to Z or a to z
or 0 to 9 inclusive, otherwise FALSE will be returned.

Name:
int isalpha (int c);
Header:
ctype.h
Description:
This is a macro that will return TRUE if the ASCII character c is between A to Z or a to z
inclusive, otherwise FALSE will be returned.

Name:
int iscntrl (int c);
Header:
ctype.h
Description:
This is a macro that will return TRUE if the ASCII character c is between ASCII 0x00 to
ASCII 0x1f inclusive or ASCII 0x7f, otherwise FALSE will be returned.

Name:
int isdigit (int c);

104
Header:
ctype.h
Description:
This is a macro that will return TRUE if the ASCII character c is between 0 to 9
inclusive, otherwise FALSE will be returned.

Name:
int islower (int c);
Header:
ctype.h
Description:
This is a macro that will return TRUE if the ASCII character c is between a to z
inclusive, otherwise FALSE will be returned.

Name:
int isprint (int c);
Header:
ctype.h
Description:
This is a macro that will return TRUE if the ASCII character c is printable character (ie:
not a control character), otherwise FALSE will be returned.

Name:
int ispunct (int c);
Header:
ctype.h
Description:
This is a macro that will return TRUE if the ASCII character c is any printable character
that is not alphanumeric or a space, otherwise FALSE will be returned.

Name:
int isspace (int c);
Header:
ctype.h

105
Description:
This is a macro that will return TRUE if the ASCII character c is a space character (' '),
otherwise FALSE will be returned.

Name:
int isupper (int c);
Header:
ctype.h
Description:
This is a macro that will return TRUE if the ASCII character c is between A to Z
inclusive, otherwise FALSE will be returned.

Name:
int isxdigit (int c);
Header:
ctype.h
Description:
This is a macro that will return TRUE if the ASCII character c is hexadecimal digit (0 to
9 inclusive, a to f inclusive or A to F inclusive), otherwise FALSE will be returned.
Name:
int tolower (int c);
Header:
ctype.h
Description:
This function will convert c into a lower case character, if and only if it was an upper
case character.

Name:
int toupper (int c);
Header:
ctype.h
Description:
This function will convert c into an upper case character, if and only if it was a lower
case character.

106
11.4. Miscellaneous Functions
Name:
void exit (int value);
Header:
stdlib.h
Description:
This function allows you to terminate your program at any point in the program. The
caller of the program (the operating system, for example) will receive the value.
The use of the exit() function should be carefully weighed as its use causes significant
problems with the destruction of objects in C++.

11.5. Common Beginner Errors

• Make sure that the appropriate include files are included for the prototypes.
• Don't forget that fgets() will often have an extraneous '\n' at the end of the string.
• Make sure to use the #pragma mentioned in the previous Useful Tip to eliminate
extraneous warnings.

11.6. Summary
In this chapter we had an introductory look at some of the various predefined functions
most C compiler manufacturers supply for us. These functions are all part of the standard
library of functions, and enable a programmer from (hopefully) reinventing the wheel in
many cases. The use of these functions is unlimited, and the main requirement is that the
header files necessary for the individual library functions be included so that prototype
definitions can be compiled along with your code.

107
Chapter 12. Simple Pointers
Pointers are one of the most difficult topics in C. We will introduce the concept of a
pointer variable in a (hopefully) gradual way in order to lead us to more advanced topics.
The following is adapted from C by Dissection by Al Kelley and Ira Pohl (Addison
Wesley, Toronto, 2001, ISBN 0-201-71374-8, pages 264 to 265, 267, and 271 to 272)

12.1. Pointer Syntax

If v is a variable (such as which we've been used to since early in the course), then &v is
the location in memory of its stored value. This location is also referred to as its address.
Pointer variables can be declared in programs and then used to take addresses as values.
The declaration
int i = 0;
followed by
int *p;
declares i to be of type int and p to be of type "pointer to int". The legal range of values
for any pointer always includes the special address 0 and a set of positive integers that are
interpreted as machine addresses on a particular C system. Typically, the symbolic
constant NULL is defined as zero in stdio.h. Some examples of assignment to the
pointer p are:
p = &i;
p = 0;
p = NULL;
In the first example, we think of p as "referring to i", or "pointing to i", or "containing the
address of i". The second and third examples show assignments of the special value zero
to the pointer p. It is considered a special value because nothing can be validly stored
there.

12.2. Addressing and Dereferencing

We want to examine some elementary code to show what is happening in memory. Let's
start with some declarations:
int a = 7;
int *p = NULL;
The second declaration is read as "p is a pointer to an int variable with an initial value
of NULL." Anytime one declares a pointer variable, it has an underlying type such as int
that it can reference. The declaration of a causes the compiler to allocate memory space
for an int. The declaration of p causes the compiler to allocate memory space for a
pointer to int. After the assignment statement:
p = &a;
we can use the pointer p to access the value stored in a. This is done through the
dereference (or indirection) operator *. This operator is unary (as opposed to the
108
 multiplication use of the asterisk). Because p is a pointer, the expression *p has the value
of the variable to which p points. The pointer p is said to have a value which is the
address of the int variable a. The term indirection is taken from assembly programming.
The direct value of p is a memory location, whereas *p is the indirect value of p, namely
the value at the memory location stored in p. Consider the statement
printf("p points to %d\n", *p);
Because p points to a and a has value 7, the dereferenced value of p is 7 and that is what
is printed. Now consider
*p = 3;
printf("a is %d\n", a);
The first statement is read "the item pointed to by p is assigned the value 3." Because p
points to a, the stored value of a in memory is overwritten with the value 3. Thus, when
we print out the value of a, 3 is printed. This is an indirect way of changing the value of
a variable through the use of a pointer variable.
In simple terms, the dereference operator * is the inverse of the address operator &.
Consider the following code:
double x = 1;
double y = 2;
double *p = NULL;
p = &x;
y = *p;
First, p is assigned the address of x. Then y is assigned the value of the item pointed to
by p. The last statement is equivalent to
y = *(&x);
which is equivalent to
y = x;
A pointer variable can be initialized in a declaration but the notation is confusing to the
beginning programmer. Here is an example:
int a = 0;
int *p = &a;
This declaration tells the compiler that p is a pointer to int and that the initial value to p is
&a. Do not read this as "the item pointed to by p is initialized to the address of a."

Useful Tip: Initializing a Pointer in a Declaration

When declaring a pointer variable, you should always realize that the asterisk indicates
the data type and that dereferencing is not happening.

12.3. Pass-by-reference
Whenever variables are passed as parameters to a function, their values are copied and
any changes within the function are not reflected in the calling environment. This pass-

109
 by-value mechanism is strictly adhered to in C. An alternative is to use the addresses of
the variables to achieve the effect of pass-by-reference and, thus, be able to change the
contents of the variables in the calling environment.
For a function to achieve pass-by-reference, we must tell the function where the variables
that we wish to modify are located. To do so, pointers must be used in the parameter list
in the function definition. Then, when the function is called, addresses of variables must
be passed as parameters.

12.4. Pass-by-reference Example

The function swap() in the following program illustrates these ideas.
#include <stdio.h>

void swap(int *p, int *q);

int main (void)

{
int a = 3;
int b = 7;

printf("%d %d\n", a, b);

swap(&a, &b);
printf("%d %d\n", a, b);
return 0;
} /* end main */

Because the addresses of a and b are passed as parameters to swap(), the function is able
to exchange the values of a and b in the calling environment.
void swap(int *p, int *q)
{
int tmp = 0;

tmp = *p; // First line of interest

*p = *q; // Second line of interest
*q = tmp; // Third line of interest
} /* end swap */
We read the three noted lines above as follows:
• The variable tmp is assigned the value of the variable pointed to by p.
• The variable pointed to by p is assigned the value of the variable pointed to by q.
• The variable pointed to by q is assigned the value of tmp.
After these three statements have been executed, a has the value 7 and b has the value 3.

110
12.5. Another Example Of Pass-By-Reference
Here's another example of using pass-by-reference to return multiple values from a
function:
#include <stdio.h>
#include <math.h>

void convertHms (double a, double *b, double *c, double *d);

void convertHms (double a, double *b, double *c, double *d)

{
double x = 0.0;

*b = (int)a;/* cast to an int to chop decimals */

x = (a - *b) * 60.0;
*c = (int)x;
*d = (int)((x - *c) * 60.0);
} /* end convertHms */

int main (void)

{
double hour = 0.0;
double h = 0.0;
double m = 0.0;
double s = 0.0;
char buffer[100] = {0};

printf ("Enter hour (ie: 10.25): ");

gets (buffer);
hour = atof (buffer);

convertHms (hour, &h, &m, &s);

printf ("hour: %f = h: %f m: %f s: %f\n",
hour, h, m, s);
return 0;
} /* end main */

The above program asks the user to enter the time in a fractional format (meaning that
you enter the time using fractions of an hour) and will convert that fractional hour into
the appropriate hour, minute and second. For example, 10.25 hours will be converted to
10 hours, 15 minutes and 0 seconds, and 22.341666667 will be converted into 22 hours,
20 minutes and 30 seconds.
Notice that there are three values that must be returned by a function which will convert
the fractional hours. We could have made the doubles h, m and s global, and had the
function convert_hms() set the globals with the appropriate values, but that is messy and
also bad programming style. A much more elegant solution is to pass to convert_hms()
pointers to the three variables it must fill: a pointer to h, a pointer to m and a pointer to s.
111
 The ampersand tells the C compiler to generate code that calculates the address in
memory where the variable exists. The address-of operation always generates a pointer to
the variable it operates on. To give convert_hms() pointers to h, m, and s, we use the
address-of operator: &h, &m, and &s. Note that this generates three double precision
pointers (double *).
The actual function convertHms() has been declared to accept a double parameter a, and
3 double pointer parameters. Parameters b, c, and d are all pointer variables to double
precision floating point numbers in memory. The function will proceed to convert the
number passed in a into the hours, the minutes and the seconds. Notice that the converted
values are supplied to the caller of this function using the * (contents-of) operator, which
is the opposite of the address-of operator. The contents of, say, parameter variable c in
this function are filled with the converted minutes, therefore, in this manner, the local in
main() called m will be able to obtain the minutes from convertHms().
This technique is extremely valuable in terms of creating functions that are meant to
supply back more than one piece of information. Because the return value mechanism is
limited to one piece of information at one time, passing in addresses of variables to fill in
is a wonderful solution, and ensures that functions can be expanded infinitely.
NOTE: A better name for convertHms() would be convertHourMinuteSecond() to
avoid the issue of "what's an Hms?" :-).

12.6. Casting
You will notice a curious expression used in the convertHms() function:
*b = (int)a;
The keyword int inside the round brackets is used to convert, or cast (alter) the data type
of variable a temporarily from a double precision floating point number to an integer.
Using an int cast against a double precision floating point value will truncate, or chop
off, the fractional digits, leaving only the whole number. This is how we can obtain the
integer portion of our fractional hour. 10.25 will be converted into the integer value 10,
and because it is to be stored in the contents of an double precision memory location
pointed to by b, it is automatically casted or reconverted back into a double before being
stored inside the memory pointed to by b. Casting is used throughout C programs to
easily convert from one data type to another.
As a further example of casting, suppose you wanted to take a look at the individual bytes
of a short integer (remember that an short is two chars side by side), try this:
short x = 0;
char *a = NULL;
. . .
x = 1234;

/*
* taking the address of x would
* normally get an short pointer,
* so cast to a char pointer!

112
 */

a = (char *)&x;
printf ("1st byte (in hex): %x\n", *a);
a++;
printf ("second byte (in hex): %x\n", *a);
As the comment above states, using &x in your program, (if x was an short), would
generate an integer pointer (short *). The casting expression (char *) is used to change
the pointer from being an integer pointer to that of a character pointer. Pointers can be
created and manipulated for any data type you want, at your free will! By not using
casting, however, your program may not compile cleanly as most C compilers will
generate errors or warning about pointer type mismatches.

12.7. Common Beginner Errors

• Always initialize pointer variables when declared.
• Distinguish between the * and & operators when used in front of variables.

12.8. Summary
This chapter briefly introduces pointer variables through the concept of assigning the
addresses of variables to pointer variables in order to indirectly change or access values.
Pass-by-reference is introduced as a way of returning more than one item from a
function. Casting is used to convert between data types.

113
Chapter 13. Arrays
13.1. Introduction
Over the last few chapters we have introduced enough programming concepts for the C
language to enable a programmer to write some rather powerful and interesting software.
In this chapter, we will expand this knowledge base by introducing our first complex data
object: the array. This data object is sometimes referred to as a data structure because
there is a structure or organization to how the data for this object lies in memory. You
will also sometimes see an array referred to as an aggregate data type.
In real life, there are many examples of arrays. An accounting worksheet is grouped as a
two-dimensional array or matrix of cells (some number of rows by some number of
columns) where an accountant may place financial figures, and computations are done on
these figures to arrive at subtotals for each column and row. A street of houses is another
example. Each house has an offset (the house number) from the start of the street, and
houses are generally a linear progression from one end of a street to another. Hotel rooms
are another real life example.
An array in C can be thought of as a linear sequence of memory locations reserved for the
storage of objects of a particular data type. The programmer has the ability to manipulate
these linearly arranged objects with the notations introduced in this chapter.
An array in C programming has many uses. An array can be used to hold data obtained
from a user, or it can hold characters typed in on the keyboard (for example, the
keystroke buffer that is used in conjunction with gets()) or even hold memory locations
of functions that can be called without using the function name. The possibilities in C are
almost endless!
The following is a pictorial representation of an array of short integers, of size 5
elements, starting at address 536 (arbitrary value):

Table 13-1: Array Layout

Memory 536 538 540 542 544

location

Array offset 0 1 2 3 4

Value -5 123 0 3192 -32

114
13.2. Base Address
The above table shows that there are five short integers placed in memory, starting at
memory location 536. This location is termed the base address of the array (or
sometimes the base location). Notice that the memory locations increment by a value of
2 for each integer: recall that an short is composed of two bytes of memory.

13.3. Array Offset

The table above also shows the array offset to each element in the array. The first
element, which is the value -5, is at array offset 0 from the start of the array. The next
integer is at an array offset of 1.
An array offset is simply the element within the array that we wish to work with, and it
corresponds to a location in memory starting at the base memory location, plus the size of
the data object multiplied by the offset. This general rule of thumb applies to arrays of all
data types: chars, ints, longs, floats and of course doubles. Note, however, that the
compiler is responsible for determining where an array object lies in memory, not the
programmer. The programmer only needs to use the array offset to reference a particular
element in the array.
In the sample array above, the first element is at memory location 536 plus an array offset
of 0, which is of course memory location 536. The second element is at array offset of 1.
Since a short integer takes two bytes of memory, the actual byte offset is 1 * 2, as a short
takes two bytes, giving us a memory location of 538. Offset 2 is at memory location 540
(location 536 + 2 * 2). If the sample array were an array of double, we would be
multiplying the array offset by the size of a double to compute where in memory an
element resides. A programmer need not worry about where in memory these objects lie.
This discussion simply aids in our understanding of C and a program's memory layout to
know how a compiler treats objects in an array.

13.4. Array Offset Operator

The generalized notation for accessing an element within an array is simply the name of
the array, with the offset inside of the square brackets. Using this notation, a programmer
can extract information at that array offset, or place new information into that array
offset. In fact, the square brackets [ ] are nothing more than a special operator that
informs the compiler to add an offset, which is scaled by the size of the data type, to a
base address.

13.5. Notation
In C, when we declare an array variable to the compiler, we must always specify the
dimension of an array (in some way or another). This is necessary so that the compiler
can reserve the correct amount of bytes for the purposes of our program.
The following is an example of a C array declaration:
short x[5];

115
This declaration will reserve memory for 5 consecutive short integers in an array called x
for this example. It should be plain that an array in C is an easy method of being able to
reserve large blocks of memory for data storage, and array notation allows simple access
to each element in this block.
The generalized notation for declaring a variable to be an array is:
data-type variable-name[# of elements];
Once we have our array declared, in our code, we can access an object within an array by
using subscript notation. This is done by placing the object's array offset inside of
square brackets following the array name in our code. Although this is similar to the way
we declare an array, we are performing a totally different operation when this is done in
code. As an example, using our integer array x from above, we can access elements in the
array using notation such as:
x[2] = 5; /* assign 5 into object number 2 */
C always references array elements starting at the "zeroth" element. In other words, the
first element in our array is always x[0], the second is x[1], the third is x[2], and so on.
The idea of a zero based array is explained using our array offset discussion: the first
element of an array is at array offset zero from the start of the array. The second element
is array offset by 1 in the array, and so on.
To help rationalize our understanding of arrays thus far, consider the following scenario.
Imagine a program that must keep track of the mileage driven by a sales person for a
week, and then report the total mileage. A first crack at this program would probably set
up 5 integers, similar to the following:
int day1 = 0;
int day2 = 0;
int day3 = 0;
int day4 = 0;
int day5 = 0;
int total = 0; /* used to hold the total mileage */
The 5 integer variables would thus be able to hold all 5 days worth of mileage data.
To generate a report that outputs the total mileage driven by the sales person, we may
compute the total as thus:
total = day1 + day2 + day3 + day4 + day5;
So far, so good, and in fact, this is a trivial programming scenario for anyone thus far.
Now, to complicate matters, the requirement for the program will switch to maintaining
the mileage for an entire month of time. Now we have some problems. Using the method
shown thus far, we would need to create up to 31 independent integer variables to hold
the mileage data for all 31 days of a given month. This is now becoming cumbersome, as
to compute the total mileage for the month, we would have to write code to sum up 31
independent integer variables. It gets worse – imagine keeping track of one full year of
data!

116
This is where arrays come into play. Let’s return to our original scenario of one business
week of data. The following snippet will show how to declare an array to hold the
mileage data, and how to sum up the array elements.

int mileage[5];
int x = 0;
int total = 0;

. . .

for (x = 0; x < 5; x++)

{
total = total + mileage[x];
} /* end for */

This is far easier to deal with than the 5 independent variables. Note how we declare the
array of integers – we have told the compiler to reserve 5 consecutive int data objects in
memory. We have our summation variable called total set to zero as before. We have one
extra variable, x, which will be used to run a loop.
Recall that an array starts at offset zero, and if the array has 5 elements, the last element
of the array will be element number 4. Hence, as long as x is starting at zero, and remains
less than 5, we are accessing valid elements of the array.
To sum up the array, we add each individual element of the array to the summation total.
This is exactly the same as what we had previously, where we added day1 to day2 to
day3, etc.
By running the loop, we now have an algorithmic method to generate the total. An
algorithm is a software formula to compute a result. Algorithms are important because
they will simplify work, in the long term. Here, it appears we have done more work to
compute the result of 5 days of mileage. But now consider the task of doing one month.
We have a very easy way to handle this – change the 5 in the code snippet above to 31,
and we’re home free!
And to do a year, change the 5 to 365, and instantly, we have the ability to compute one
year’s worth of data.
Hopefully this illustrates clearly the benefits of a common grouping of data elements
under a single name. Arrays are so important to C that this chapter will drill home a
number of techniques you can use to implement arrays as part of your programming
efforts.

117
 Useful Tip: Use constants for array sizes
It is best to use constants (through #define) to define array sizes.

13.6. More About Array Offsets

The C language is somewhat strict in how you write software, as you must always define
variables, functions, etc. before you get a program to compile correctly. Unfortunately,
the compiler is only capable of enforcing syntax rules, and does no sanity checks on
whether the code we have written will perform a task adequately.
One of the main dangers of working with arrays is to try to access an element outside of
the dimension we have specified. Take for example:
long val[10];
The above declaration reserves ten long integers in memory one after another, a total of
40 bytes of memory being reserved. In our code, we have access to ten longs through the
array offsets val[0] to val[9] inclusive.
If you inadvertently use an array offset of val[10], you should be aware that you are now
trying to access the 11th long integer in the array. However, our array has been
dimensioned to hold only 10 long integers. What does the compiler do?
In C, nothing special! The compiler will happily generate code to access this 11th long.
If our code contained something similar to:
val[10] = 55L;
we have now overwritten a piece of memory past the end of the array with four bytes
organized as the value 55. If we had a crucial variable stored after the end of the array,
we have now destroyed the contents of that variable, which will lead to all kinds of havoc
in our program. For this reason, C programmers must be extra vigilant in keeping array
offsets within their bounds.
It is up to the programmer to ensure that what an array offset is doing is legal for the
particular application. Array offsets such as:
x[-2] = 5;
are legal. It is only valid, however, as long as what is in memory two integers before the
start of the array is something that can be overwritten without worry. (The example above
is something that is done more often with pointers than with arrays.)

13.7. Initialization of Arrays

Often in C, we would like to set up an array so that the array values we need are ready to
go when we execute our code. There are three ways to accomplish this. Study the first
method below:

int main (void)

{

118
int anArray[7]; /* declaration */

anArray[0] = 100;
anArray[1] = 32700;
anArray[2] = -31231;
anArray[3] = 0;
anArray[4] = 256;
anArray[5] = 255;
anArray[6] = 4321;
. . .
} /* end main */
The above code declares an array of 7 integers, and later in our code, we give individual
elements of the array their values. However, to initialize a number of elements in this
fashion is awkward in terms the amount of repetitive code is required.
Recall how to initialize a simple variable:
int x = 5;
double y = 10.33;
For array initialization we use the following second type of initialization:
int anArray[7] =
{
100, 32700, -31231, 0, 256, 255, 4321
};
Here we have an initialization of the array anArray, and is useful to supply your C
program with values at compile time, rather than at run time.
The third method will seem to contradict an earlier statement:
int anArray[] =
{
100, 32700, -31231, 0, 256, 255, 4321
};
The array appears to be dimensionless, due to the use of [] (the empty square brackets).
The compiler will not complain about this declaration, however, as the programmer is
supplying 7 integers as initial values for this array. The programmer has effectively told
the compiler that the array has an implicit dimension of 7. This notation is useful in
many circumstances where the contents of an array are being modified in the source code
often, so rather than constantly change the dimension, leave it out, and let the compiler
figure out the size of the array from the number of initialization values.
In the latter two examples, what the C compiler will do is allocate a block of memory for
the array, and fill in the values for you, saving the trouble of writing the code to do the
assignments one element at a time.
What follows is a more concrete example of where using and initializing an array can
simplify a program tremendously:
#include <stdio.h>

119
int calcNumDays (int m, int d);

int daysPerMonth[12] = {31,28,31,30,31,30,31,31,30,31,30,31};

int calcNumDays (int m, int d)

{
int numDays = 0;
int x = 0;

/* add from jan to last fully completed month */

for (x = 0; x < (m - 1); x++)
numDays += daysPerMonth[x];
numDays += d;
return (numDays);
} /* end calcNumDays */

int main (void)

{
char buffer[100] = {0};
int month = 0;
int day = 0;
int total = 0;

while (1)
{
printf ("Enter month (999 to end): ");
gets (buffer);
month = atoi (buffer);
if (month == 999)
{
break;
}

printf ("Enter day: ");

gets (buffer);
day = atoi (buffer);

total = calcNumDays (month, day);

printf ("month: %d day: %d total: %d\n",
month, day, total);
}/* end while */

printf ("done!\n");
return 0;
} /* end main */

120
 As you can see from the math inside the function calcNumDays(), the array version of
this calculation is much simpler to code and understand than the version we had a few
chapters ago using a chained series of ifs. We have replaced the series of if statements
with a simple run through an array of values, from the start of the year to the month just
completed, to get the total days in the complete months elapsed so far. We then add the
number of days into the current month, and bingo, we have the total. Of course, error
checking should still be done to ensure all results of the program are valid.

Useful Tip: Initializing Arrays

If you have any elements of an array initialized, any element that you did not explicitly
specify are implicitly initialized to 0.
Useful Tip: When a Table is Better Than Chained if Statements
It is always much more efficient to code tabular forms of information into an array to be
referenced through some form of offset for the particular piece of information that is
needed, rather than perform a complicated series of if statements to isolate the values
desired. NEWDATE.C is a prime example of an array based table of data.
Useful Tip: while(1)
while(1) means "loop forever". Of course, you can always break out, like we did here.
Course Requirement: Always Initialize Arrays
Always initialize your array variables, even if it’s just with a single 0 in the very first
element.

13.8. Multi-Dimensional Arrays

C allows you to have multidimensional arrays. This is a further extension of the idea of
an array, which allows you to efficiently organize and access groups of information. One
of the most common multidimensional arrangements is a row-column array, otherwise
known as a two dimensional array. Many useful programs use this format - for example,
the spreadsheet programs you may use on a day to day basis use a two dimensional array
to organize the rows and columns of cells!
With more than two dimensions, an array may get hard to visualize, but to simplify your
thought patterns, think of a multidimensional array as nothing more than an array of
arrays!

13.9. Common Beginner Errors

• Make sure that you don't access past the end of the array.
• Don't confuse the dimensions of multi-dimensional arrays.

13.10. Summary
This chapter has introduced the concept of the array, which was defined as a linear
progression of data objects in memory. The advantage of such a structure is that we can

121
easily access individual members of such an array using the name of the array and the
offset to the data object. It was noted that C arrays are zero based, meaning that the first
element is actually at offset zero, the second element at offset 1, and so on. We also saw
that in our generated code, offsets are multiplied by the size of the data type in bytes to
obtain the byte offset from the base address of the array.

122
Chapter 14. Strings
So far, we've glossed over the concept of character strings. Unfortunately, dealing with
strings is one of the more awkward parts of the language. Let's address it with an
example to lead off.

14.1. C and Character Strings

Let us have a look at a very simple C program which will calculate the length of a string
of characters:
#include <stdio.h>

/* prototype */
int length (char string[]);

/*
* note that we are passing an array reference to
* the function length()
*/

int length (char string[])

{
int len = 0;

len = 0;
while (string[len] != '\0') /* '\0' ends string */
len++;/* increment len */
return (len); /* and return it */
} /* end length */

int main (void)

{
char buffer[100] = {0};

printf ("Enter a string: ");

gets (buffer);
printf ("Length of %s is: %d\n", buffer,
length (buffer));
printf ("Length of %s is: %d\n", "hello",
length ("hello"));
return 0;
}/* end main */

In the main() function, we declared a local variable called buffer as an array of

characters, 100 in total, which will contain the string of characters we type in on the
keyboard when we call the gets() function. We have just instructed the compiler to
reserve, in memory, space for 100 consecutive characters of information.
123
 Near the start of program, the first function we declare is called length(). This function is
used to calculate the length of any arbitrary string of characters. The term "string of
characters" is a fancy term for an arbitrary concept in C programming. In C, a string is
defined an array of ASCII (char) values one after another, ending in a null terminator,
which is the ASCII code with value zero. This is known as null-termination.
The null terminator is represented by the escape sequence '\0'. This is not a two-character
sequence. That is, this is not a backslash followed by a character ‘0’. This is a way of
denoting a single character. The presence of a backslash followed by a number or series
of numbers within single quotes allow you to specify unprintable characters. The series
of numbers indicate the octal (see the Appendix on Counting Systems) value of the
ASCII character.
A string in C always has a null terminator, since this is the programming convention used
to inform the various string handling functions in the standard library that the string of
characters is terminated (ended) inside memory. As an example, the string "Ab2%" is
represented as the following array:

Array offset 0 1 2 3 4

char 'A' 'b' '2' '%' null-

terminator

ASCII 65 98 50 37 0

In our example program here, a string is passed to the length() function by means of a
rather odd looking parameter declaration:
char string[]
This parameter declaration, with no dimension value inside of the square brackets,
informs the compiler that this function is going to be passed a reference (or base address)
to some arbitrary sized array of characters (a string). However, at the time we develop
this function, we are not sure of the exact length. C allows array parameters to be defined
in this manner, which offers flexibility in working with arrays, as array-handling
functions need not know the exact size of the array it will work with. Often, a second
parameter will be supplied that would indicate the maximum number of elements to work
with.
The function sets a counter called len to zero, and will use this counter to step from
element to element inside of the array of characters. As long as the current string array
element string[len] is not equal to the null terminator, we know we haven't yet reached
the end of a C string. (Remember that C strings are zero terminated, meaning that there is
always a '\0' character at the end of any valid C string!)

124
 As soon as string[len] is '\0', we terminate the while loop, whose only job was to
increment len, and then return the current value of len. This is the length of the string in
characters.
In our main() function, we get some user input, and then call the length() function to
calculate the length of that user input, and finally, display how long the string was (in
characters typed). Note how we can pass to the length() function a reference to the base
address of the keystroke buffer by simply using the name of the array buffer. This is how
we can pass array references to any function that requires knowledge of the start of an
array. The array name all by itself is always the base address of that array.

Useful Tip: Array Names

The name of an array is the same as the address of its very first (or zeroth) element.
Memorize this.

In main() we also print how long the string "hello" is. Remember that a string of
characters is nothing more than an array of characters with a null-termination, hence we
can give any function that needs an array reference a string such as the above. "hello"
will of course result in a string of length 5.

14.2. The Innards of Strings

We have discussed strings without really describing what the compiler does with them.
As we know, a string is normally defined as a series of characters in between double
quotation marks such as:
"ABCdefGHI"
Strings such as the above are used all over in C. For instance, all the formatting
information for printf() is supplied through the use of formatting strings. Text we want
placed on the screen through the use of printf() is supplied using strings.
When the compiler comes across a string as shown above, it does two things. Firstly, it
will take the characters in the string, and place them aside as it generates the executable
file for the source code. As it places them aside, it will also append a null termination
(the character '\0') at the end of the string in order for the string to be properly defined.
The null termination is a marker that is used to tell functions such as printf() where the
string ends in memory so that these functions don't attempt to print out the character
representation of all of memory.
What the compiler puts in the place of the string inside the code is the base address of
the first character in this string of characters. What printf() obtains is the address of
where the first byte of the string exists in memory. All that these functions (and any
function that expects a character-based string) do is simply run through the characters in
the null terminated array one at a time until the null terminator is reached. As far as the C
compiler is concerned, a string is an address of where the first character of the string
resides in memory, and passing a string as a parameter to a function is simply passing the
base address of this array.

125
 If you recall the function gets(), it requires a character buffer to be passed as its
parameter. As we've seen in this chapter, a character buffer is just an array of characters
of some dimensioned size. The character buffer name is the base address of where the
first character of the buffer resides. All that gets() does is obtain a keystroke from the
keyboard, and place that keystroke in the next available location in the character array.
When <enter> is hit, gets() will place a null termination after the last character received,
terminating the string properly.
As you should hopefully see, both quoted strings and character input buffers are one and
the same as far as the C compiler is concerned: they are nothing more than base addresses
of where their respective first characters lie in memory! This is a crucial concept to
understand for the next chapter, which deals with pointers - a data type which allows a
programmer to point to any piece of data in memory, including arrays of information.

Useful Tip: Overwriting Null-termination in Strings

Whenever manipulating the characters within a character array, especially if you are
attempting to change the length of a string of characters, it is a good idea to ensure that
the null termination is always updated to the correct location inside the array. Accidental
overwrites of the null termination with a non-null value is a good (bad?!?) way to crash
an application!

14.3. String Library Functions

At the start of the chapter, it was mentioned that dealing with strings in C was awkward.
That is because you can't deal with strings with many of the usual operators like =, = =,
+, <, >, etc. Instead of using the operators, there exist string library functions that are
used.
In the next few sections, assume that you have the following two string variables:
char str1[30] = "fred";
char str2[30] = "barney";

Assignment using strcpy instead of =

You can't do this:
str1 = str2;
Instead do this:
strcpy(str1, str2);
strcpy() will copy the contents of the second parameter into the first parameter. Thus,
the contents of str1 will no longer contain "fred". It will contain "barney". str2 will be
unchanged.
There are other variations on strcpy() such as strncpy() that do assignment in slightly
different ways as well.

126
Comparison using strcmp instead of ==, <, and >
You can't do this:
if( str1 == str2 )
Instead do this:
if( strcmp(str1, str2) == 0 )
strcmp() will compare the contents of the second parameter with the contents of the first
parameter. If they are equal, 0 is returned. If str1 is less than str2, an integer less than 0
is returned. If str1 is greater than str2, an integer greater than 0 is returned.
It is important to explicitly compare the result returned from strcmp() to 0. Since 0 is
equivalent to FALSE, doing
if( strcmp(str1, str2) )
will give you exactly the wrong result if you think it will be true if the two strings are
equal.
Comparison of strings is done according to the ASCII values of the characters in the
strings. Character-by-character comparison is done until an inequality is found between
characters at the same offset in both strings.
There are other variations on strcmp() such as stricmp(), strcmpi(), strncmp(), etc. that do
comparison in slightly different ways (such as case-insensitivity) as well.

Concatenation using strcat instead of +

You can't do this:
str1 = str1 + str2;
Instead do this:
strcat(str1, str2);
strcat() will copy the contents of the second parameter to the end of the contents of the
first parameter. The null-termination will be overwritten and put at the end of the new,
longer string.
You must make sure that there is enough room in str1 to hold both str1, str2, and a null-
termination. The compiler will not do it for you!

strlen to find the length of a string

It is very valuable to know how long a string is. strlen() is a library function that will
give you this information. Pass the string as an argument. strlen() returns the length as a
size_t item, which is typically the same as a primitive data type. Which primitive data
type size_t is equivalent to depends on the system. Under Visual C++ v6, size_t is
equivalent to an unsigned long, indicating that strings can be over 4 billion characters
long.

127
14.4. Common Beginner Errors
• Don't use normal operators like =, = =, +, etc. to work with strings. Use the string
library functions.
• When using strcpy() and strcat(), make sure that you don't access past the end of the
array.
• When using strcmp(), explicitly compare the result to 0.

14.5. Summary
This chapter introduces the way that C handles working with strings. Selected string
library functions were used to replace operators that are not supported by the language.

128
Chapter 15. More Pointers
15.1. Review of Array Concepts
In the last chapter, we introduced the concept of arrays. The array, as we discovered, is a
structured data type that allows access to a linear set of data objects in memory. The array
allowed efficient access to objects in memory since the C compiler could generate code
that took advantage of this linear nature.
It was stressed that, when working with arrays, the C compiler is mainly concerned with
offsets from the base address of the array. The base address was defined as the memory
location where the array began. When the compiler generates code, the offsets that are
added to the base address are scaled by the data object size. An offset for an array of
longs will be multiplied first by 4 (4 bytes in a long), before adding it to the array base
address to locate the first byte of that data object.
To refresh our memories about the arrangement in memory for an array, study again the
following chart showing a fictitious array of short integers:

Table 15-1: Array Contents

Memory 536 538 540 542 544

location

Array offset 0 1 2 3 4

Value -5 123 0 3192 -32

15.2. Addresses
Notice that the starting location of our array is 536. This location is an address in
memory, which is nothing more than a place in memory where data can be stored. It turns
out that we can access every single byte of memory in our computer by using the
appropriate address.
Addresses for microcomputers typically begin at location 0, and end up at one less than
the number of bytes the computer is designed to hold. Although it is useful to understand
exactly how your PC accesses memory (both read only ROM and writable RAM), all a
program really needs to know is that all bytes of memory (both RAM and ROM) can be
accessed using the appropriate addresses for your PC if permitted by your operating
system.

129
 The idea behind being able to address a particular location in memory is akin to being
able to locate an individual living in a city. We look up the address of that individual, go
to that address, and by being let in through the front door, we can see what is inside of
that address (for example, the number of occupants). To help you visualize how memory
addresses work, think of your computer's memory as being a large city of millions of
inhabitants. Each citizen (byte) must have their own address of where they live (there is
no such thing as a homeless byte in computerland!). To see who's at home at a particular
address, use the memory address of that byte, and have a look at that location (use a read
operation on that memory location). To move a new person into that address (to change
the value), use a write operation to change the contents of that address.
As should be obvious from the above analogy, an address in a computer's memory space
allows access to the contents of a particular memory location so that software can extract
those contents, or place new information into that location. This idea of looking at or
placing new information into a location in memory is usually referred to as peeking and
poking information into memory.
As another example of an address, let us have a look at a character string in memory:
“HELLO”

Table 15-2: Character String Representation In Memory

Memory 1016 1017 1018 1019 1020 1021
Address

Character 'H' 'E' 'L' 'L' 'O' null-

terminator

ASCII value 72 69 76 76 79 0

The table above should re-iterate the concept that an address has no particular or specific
value: it is simply a location inside the computer's memory where a value can be read or
written to. When you write a program in C that uses character strings, arrays, etc., these
data objects reside in memory locations that vary from computer to computer, program to
program, somewhere inside those memory locations. Addresses allow you to obtain and
manipulate those bytes of memory, giving you almost complete control over where you
place objects in memory, or what you do with them.
Of course, this power can be (and often is) abused, or even worse, accidentally misused,
resulting in mysterious side effects like system crashes, weird characters on the screen
and the like.

15.3. Initializing a Pointer Variable

We can declare a pointer variable to access characters of memory:
char *stringPointer = "HELLO12";

130
 The first line is the declaration of a pointer variable. As you should recall, the declaration
begins with a data type, in this case char, since a character pointer is simply a pointer to
objects of type char. Also recall that the asterisk in front of the name of the variable tells
the C compiler that we are declaring a pointer to that data type.
To make the pointer useful, we must make it point to something in memory. A
convenient object to point to is the first byte of a string of characters. We can assign into
stringPointer the address of where the capital H is in the string HELLO12. It appears
that we are setting our pointer variable equal to the string, but, if you remember from our
discussion about what the compiler does with string constants, all the double quotes of
the string indicate is that an address should be generated. This address will indicate where
the first character of the string lies in memory. The assignment of the string
"HELLO12" to the pointer variable is really the assigning of the address of where in
memory the first byte of those 7 bytes (plus null terminator) exist.
If the string "HELLO12" happened to reside in memory locations 1016 through to 1023,
stringPointer will get the value 1016. The reason for this is because of the fact
stringPointer is a pointer variable: the string begins at that address, therefore the pointer
variable will get that address. It cannot be stressed enough how important this concept is
for a good understanding of how pointers work.
We will look at other initialization techniques later in this chapter, for other pointer types.
At this point, we should introduce a constant that is heavily used with pointers. NULL
represents an address (typically address 0) that cannot be accessed. It is customary to
initialize pointer variables to NULL if the variable isn’t pointing to something valid.
NULL is usually defined in stdio.h or stdlib.h.

Course Requirement: Pointer Initialization

Always initialize pointer variables. If you don’t know what a valid value would be for
the pointer when you declare it, initialize it to NULL.

15.4. Accessing Memory Using Pointers

When using a pointer in C, the programmer has the ability of looking at memory (peeking
at the contents of an address), or storing new information in a memory location (poking
information into that address).
Now that we have an initialized character pointer we can do a number of things. In the
chapter on arrays, we studied the function called length() which could be used to
compute the length of a character string. (This function, by the way, is always included in
the standard libraries of every C compiler as a function called strlen()). With our string
pointer variable, we can pass to that function the address of our string in memory. Recall
that an array simply references the first memory location where our consecutive set of
data objects resides. Our variable stringPointer is doing this exact same operation! In
fact, you can view an array as a fixed address pointer: the address cannot change lest we
lose our reference to our reserved memory.

131
 The way we work with arrays is the same way we can work with pointers. Let's have a
look at two examples:
char x = 0;
. . .
x = stringPointer[0]; /* array notation */
x = *stringPointer; /* pointer notation */
The first line above appears to be an array reference: indeed, since a pointer variable can
be used just like an array, you can think of stringPointer as being an array of chars. It
has an added benefit: stringPointer can be changed to point to any string of characters.
Using an array offset of 0 from where stringPointer is pointing to, we see that x will get
the value 72 (the letter 'H').

15.5. 'Contents of' Operator

The second example above has some new syntax. The statement x = *stringPointer is
read as "x is assigned the contents of the address referenced by stringPointer". Since
stringPointer is currently set to address 1016, the variable x gets what's in memory at
that address, namely 72. This uses a principle called dereferencing the pointer (using an
asterisk in front of the pointer variable).
Now you may be wondering: "if they're both the same, why bother using the weirder
looking second example"? The answer lies in the fact that using pointers is almost always
more efficient in terms of the amount of executable code produced by the C compiler.
Most microprocessors are wonderfully proficient when working with addresses, but are
not as effective at working with offsets from the start of an array. The first line above
would cause the following code to be generated:
a) obtain the address contained in stringPointer
b) obtain the offset, and scale by the data object size
c) add scaled offset to that address
d) get the contents of that address
whereas the second example generates:
a) obtain the address contained in stringPointer
b) get the contents of that address
The middle steps of scaling and adding the offset are not done using pointers since the
pointer is already where we want it.
Pointer variables are not just limited at looking at a single memory location. We can
apply any mathematical operation to a pointer variable to move the pointer (the address
that the pointer variable holds) anywhere we want, although for all intents and purposes,
only addition and subtraction are really useful. Again, an example:
stringPointer++; /* example 1 */
x = *stringPointer; /* example 2 */
stringPointer += 3; /* example 3 */
x = *stringPointer; /* example 4 */
x = *--stringPointer; /* example 5 */

132
 x = *stringPointer++; /* example 6 */
Example 1 above increments the value of the pointer. The address contained in
stringPointer changes from 1016 to 1017. Example 2 will use this new address and x
will contain 69. Example 3 will increment the pointer by 3. stringPointer will now be
1020, and example 4 sets x to 79 (the contents of location 1020).
The fifth example shows some common pointer manipulations: using the contents-of
operator along with pre/post increments and decrements. Here, the operation is a pre-
decrement, which causes the value of the character pointer to be first reduced by one (to
1019), and secondly, the contents of that address are read. The last line shows a post-
increment, which causes first the current contents of where stringPointer is pointing to to
be read, and secondly, the address in stringPointer is incremented by 1.
These last two examples show that by doing pre/post increments or decrements, the
compiler will choose to manipulate the pointer differently. The pre-decrement moved the
pointer backward one byte in memory before accessing the contents of that address. The
post-increment accessed the byte first, then incremented the address. This shows that the
order of operations for pointer variables will follow the same rule that normal variables
follow: that pre-increments and pre-decrements will be done first, then the operation (the
contents-of operator in this case), and finally post-increments and post-decrements.

Useful Tip: Pointer Dereferencing

A good way to think of a dereferenced pointer (or a pointer variable with the asterisk in
front of it) is as “go to the location the pointer is pointing to and do something with what
is there.”
Useful Tip: Brackets with dereferenced incremented pointers
Usually, programmers write the increments and decrements in the following manner:
x = *(stringPointer++);
by putting the increment or decrement inside of brackets when combining them with the
contents-of operator. This makes the statement clearer.

15.6. Putting It Together

Let's have a look at a C program using pointers:
#include <stdio.h>
#include <string.h>

char * capitalizeString (char *string);

char * concatenateString (char *str1, char *str2);

char * capitalizeString (char *string)

{
while (*string != '\0')
{
if ((*string >= 'a') && (*string <= 'z'))

133
 {
/* current character is lower case! */
*string += ('A' - 'a');
}/* endif */
string++;
} /* end while */
return string;
}

char * concatenateString (char *str1, char *str2)

{
char *stringStart = NULL;

stringStart = str1;

/* get to end of string */

str1 += strlen (str1);

/* copy str2 to end of str1 */

while (*str2 != '\0')
{
*str1++ = *str2++;
}

/* add new null termination to end of str1 */

*str1 = '\0';

return stringStart;
}/* end concatenateString */

int main (void)

{
char s1[80] = "hi!";
char *s2 = NULL;

s2 = "goodbye!";

capitalizeString (s1);
printf ("s1: %s\n", s1);
concatenateString (s1, s2);

/* print string using pointer */

printf ("s1: %s\n", s1);

return 0;
}/* end main */

134
 The first function above, capitalizeString() has a single parameter, which is a pointer to
a character. Notice how we run through the entire string, character by character. The
while loop will loop as long as the contents of the pointer to the string does not equal '\0' -
the null termination. As we go, we check to see if the currently pointed to character in the
string is a lower case character (between 'a' and 'z' inclusive). If so, we subtract from that
character the letter 'a' and add to it 'A', which generates for us an upper case character
(remember that subtracting 'a' is subtracting the ASCII equivalent for that character -
same goes for adding 'A'). The function is complete as soon as we reach the null
termination of the string (ASCII zero), and the current value of the pointer is returned,
which is a pointer to the end of the character string.
The while loop will stop as soon as the content of memory is a value of zero, which is the
null terminator. Note as well that this is not quite as good a trick as it used to be as
modern software must operate in non-ASCII environments (for example Japanese,
Chinese, etc.). Internationalization of software requires explicit checking for things like
terminators, rather than assuming that a terminator is zero. For the purposes of this
course, we will assume we are in a North American locale, hence, a termination will be
zero, just to keep examples as short and sweet as possible.
The second function is concatenateString(), and is used to add one string to the end of
the second string. Notice that this function has two character pointers as parameters. The
first string pointer will have the second string added to the end of it. We first add to the
string pointer the length of the string itself, by calling the standard library function
strlen(). This gives us a pointer to the end of the string. Next, we simply copy characters
one at a time from the second string to the first string. Notice how we increment both
pointers as we copy characters. The statement:
*str1++ = *str2++;
is a fairly familiar one in C whenever data is being copied quickly from one place to
another. The C compiler will generate code to extract the current contents of str2 and
place it into memory where str1 points, then increment both str1 and str2.
This function will place a new null termination into the end of the first character string
once the copy is complete. This is needed since we must inform the rest of our software
where the new end point of the first string is.
We must always ensure, whenever we manipulate character strings such as the above
example, that we always update the null termination accordingly - if we don't we run the
chance of having strings 'fly off' into never-never land as functions like printf() will
never stop until they reach a null terminator character!

15.7. Scaling Of Pointer Math

There is one important feature to note about using pointers: depending on the data type
the pointer is pointing to, whenever arithmetic operations are performed upon these data
types, the arithmetic operation will reflect the size of the data type. For example, when
incrementing a char pointer, the pointer address is incremented by one to move to the
next byte. When decrementing a short pointer, the pointer address is decremented by two

135
 since an short is two bytes wide. Incrementing a long pointer increments the pointer
address by 4 since a long contains 4 bytes. The same rules apply for int, float and
double.

15.8. Pointers and Arrays

Pointers and arrays are somewhat interchangeable in terms of how we can access
information that can be referenced by them. The notation for arrays (the usual square
brackets with an offset inside) is applicable to pointers, in addition to the contents-of
operator, because all the array notation does is use the pointer address as the base address
of the array. Also, we can use the name of an array as a pointer, because all an array
name is is simply a pointer to the first location in the array!
For example, our now familiar character buffer for getting user input is:
char buffer[100];
We use this array by calling the gets() function, as in the following example:
printf ("Enter a string of characters ...");
gets (buffer);
The gets() function is generally prototyped in stdio.h as:
char *gets (char *s);
Notice that the parameter for gets() is a pointer to a character. Why can we pass buffer as
a parameter to gets()? Simple! The array name is the address (a pointer) to the first
element in the array. buffer is a pointer to the first char inside of the 100 char array.
The interesting actions we can do with this array stem from the fact that arrays and
pointers are somewhat interchangeable. Study the following program:
/*
* PTRDEMO.C
*/

#include <stdio.h>
#include <string.h>

int main (void)

{
char buffer[100] = {0};
char buf2[100] = {0};/* reserve some memory */

printf ("Enter a string of characters ...");

gets (buffer);

printf ("From the 3rd character onwards ...\n");

printf ("%s\n", buffer + 2);

printf ("copy what's in buffer to buf2\n");

strcpy (buf2, buffer);

136
 printf ("buf2: %s\n", buf2);

printf ("copy buffer to buf2 at offset 2\n");

strcpy (buf2 + 2, buffer);
printf ("buf2: %s\n", buf2);

printf ("copy buffer to buf2 at offset 4\n");

strcpy (&buf2[4], buffer);
printf ("buf2: %s\n", buf2);
return 0;
}/* end main */

This program uses a standard library function strcpy() which is prototyped as follows:
char *strcpy (char *s1, char *s2);
It will copy whatever is pointed to by s2 into memory pointed to by s1, until a null
terminator is located in memory pointed to by s2. Only characters will be copied since
both parameters are declared as being char pointers.
In the printf() which says we will display everything from the 3rd character onwards,
notice how we specify this fact. Recall that the name of the array is a pointer to the first
character. The name plus one then must be the address of the second character, and the
name plus two is then the address of the 3rd character. To obtain a pointer to the 3rd
character onwards, simply use the address of the array, and add 2 to it! name + 2 is a
pointer to the 3rd character of the array (or the element of the array at offset 2)!
This same idea is used in the next few examples which copy strings of information
around using strcpy(). The second string copy will copy what's in buffer into memory at
offset 2 of buf2. We generate this pointer location by using the name buf2 and adding
our offset of 2 to it.
But remember that array references can be used as well. buf2[4] is an array reference to a
character in the array at offset 4. This expression yields the value of a single character at
offset 4. To get a pointer to this single value, we could have used buf2 + 4 as the earlier
examples, but we also can use the address-of operator (&), and write:
&buf2[4]
which means "take the address of the character at offset 4 of the array buf2". A C
programmer has two methods of arriving at the exact same location when working with
arrays!
The next example reiterates our earlier discussion about the increment and decrement
operators when working with pointers. Improper use of these operators will cause
incorrect values to be obtained or stored in memory, leading to corruption or possible
lockups of your software.
int xx[5] = { 33, 44, -1, 0, -33 };
In the above line of C code, we are initializing an array of integers (5 in total) to a certain
set of values. The table above shows where in memory this table happens to be stored,
137
 and the values at the expected locations and offsets. Note that this is an integer array,
therefore, all addresses are two bytes apart.
int *c = NULL;
int y = 0;
int z = 0;
int y1 = 0;
int z1 = 0;

. . .

c = xx;
y = *++c;
z = *--c;
y1 = *c++;
z1 = *c--;
In the above example, we have set the pointer c to where xx begins in memory. Next, we
are using the increment and decrement operators to move the pointer around. The
increment and decrement operator affect the pointer in different manners depending on
how the operator is placed. (Remember, they will affect any variable in these same
ways!)
When ++ and -- are in front of a variable, it tells the compiler to increment/decrement the
variable first, before proceeding to do anything else in the expression. When it comes
after, it means increment/decrement the variable last, after everything else in the
expression is complete.
y will be set to the value 44, because we move c up by one integer in memory, then look
in memory at that location. z will be set to 33 because we move c back by one integer
before looking in memory.
y1 will be set to 33 because we look in memory first, then move the pointer ahead by one
integer, and z1 will be set to 44 because we look in memory first, then decrement the
pointer.

15.9. The scanf() Family

The way you use the scanf() family of functions is somewhat simple: you supply the
function with a formatting string, much like the string given to printf(), except that rather
than printing values, we are telling the function to extract values of a certain type. We can
tell this function to look for integers, characters, strings, longs, doubles; actually,
anything that you can print with printf(), you can obtain with a scanf() family function.
As an example, let's look at the following program: /* SCANF.C */

#include <stdio.h>

int main (void)

138
{
int x = 0;
long a = 0;
long b = 0;
char buffer[100] = {0};

printf ("enter a string: ");

scanf ("%s", buffer);

printf ("\nenter an integer: ");

scanf (" %d", &x);

printf ("\nenter 2 long ints: ");

scanf (" %ld %ld", &a, &b);

printf ("\nresults: %d %ld %ld %s\n", x, a, b, buffer);

return 0;
}/* end main */

The above program asks us to enter four different pieces of information: a string, an
integer, and two long integers. Notice that after the printf() calls to print out the prompts
to the user, we end up making three calls to scanf(), except the formatting string passed
as the first parameter is different in each case.
The first scanf() is looking for a string, which is signified by the %s in the formatting
string. %s in scanf() has the same function as in printf(), namely, that a string is about to
be manipulated. The function will look for a space-delimited string (i.e. until a space is
reached), and place the characters into the array called buffer.
The second scanf() is supposed to look for an integer. As you can see, the formatting
string is " %d", which says to expect an integer. The second parameter is our now
familiar address-of operator, which will pass the address of where in memory the
variable x is located. It should be apparent that scanf(), when extracting the values asked
for, will place them into the appropriate data object, which is supplied by the addresses
passed as the second, third, etc parameters. The space before the %d is used to tell
scanf() to skip all white space before scanning for a number: this ensures that scanf()
will ignore spaces or carriage returns already entered from the previous scanf(), and get
us the value we want.
The final scanf() function call is looking for two longs, thus the formatting string is
“ %ld %ld”
which tells scanf() to skip white space, look for a long, skip some more white space and
get another long. The two longs will be returned through the addresses of variables a and
b (&a and &b).

139
 If you're wondering about why you don't pass an address for the string in the first scanf(),
remember that the name of an array is already a pointer to the memory it occupies! You
generate the pointer to the array by simply passing the name of the character array.
The reason we waited until now before using this function is the fact that explaining the
necessity of passing addresses of variables to the function scanf() would needlessly
confuse most neophyte C programmers. Everything you can do with scanf() can be done
explicitly with gets() and the appropriate conversion function (atoi(), atof(), etc), with
much less hassle.
Importantly, scanf(), just like getchar(), uses an internal keystroke buffer to record
keyboard activity. This has implications on how scanf() will extract information from
this buffer. For instance, if you look for an integer using %d, and the user of your
software typed in a bunch of alphabetical characters, scanf() will fail to extract an
integer. Unfortunately, because no processing of the input characters took place due to
the data mismatch, the next time you try to extract information, scanf() will look at the
remaining characters in its keystroke buffer, until it processes the carriage return. This is
one of the side effects of the buffering that scanf() performs for us.
Unfortunately, that means that scanf() will be rather ill-behaved if you type in anything
that is different from what it is expecting. If you have a loop that gets input, it could end
up looping forever, getting the same invalid input over and over! This is, obviously, A
Bad Thing. That's why it's best not to use scanf(). Instead, you can use a combination of
fgets() and sscanf().

15.10. Using fgets() and sscanf() for Getting Integer Input

Here's a function that you can use to get integer values from the user in a safer manner. It
has an inherent limitation, though, such that it uses a value of -1 to indicate invalid input
so you can't tell if the user really entered -1 or if they entered something bad.
int getNum(void)
{
char record[121] = {0};
int number = 0;

/* use fgets() to get a string from the keyboard */

/* record contains the characters entered on the
* keyboard */
/* stdin indicates that we want to get the
* characters from the keyboard */
fgets(record, sizeof(record), stdin);

/* extract the number from the string; put the

* number just entered into the number variable */
/* sscanf() returns a number corresponding with the
* number of items it found in the string; since
* we're looking for one number, it should

140
 * return 1 */
if( sscanf(record, "%d", &number) != 1 )
{
/* if the user did not enter a number
* recognizable by the system, set number to
* -1 */
number = -1;
}

/* return the value that we've got */

return number;
}

To avoid the "loop-forever-on-invalid-input" problem inherent to scanf(), we can get the

entire line from the keyboard all at once and then examine it for valid input.
The act of getting the entire line from the keyboard is done by fgets(). You've seen gets()
so far. fgets() is the file version of gets(). We can use it in a non-file situation, though,
by using a pre-defined variable that represents the keyboard as a file: stdio. fgets() takes
three parameters. The first parameter is the same as you would send to gets(): the
location of a buffer to put the string in. The second parameter is the reason that we are
using fgets() instead of gets(). It is a number that indicates the upper limit on how many
characters (including null-termination) we are willing to get from the keyboard. If we set
this second parameter to the size of our buffer, we are guaranteed to never overflow that
buffer from this call. Lastly, the third parameter is given by the stdio variable that we
mentioned before. That tells the function that we want to get the input from the
keyboard.
Once we get the string into the buffer, we can scan it using sscanf(). sscanf() is a variant
of scanf() that works on strings, not keyboards. That eliminates the problem of invalid
input messing you up.

Course Requirement: scanf()

Do not use scanf(). Use sscanf() (or one of the conversion functions like atoi(), atol(), or
atof()) instead, in conjunction with gets() or fgets(). Failing to do so will lose you marks!

15.11. Common Beginner Errors

• Always make sure that pointer variables have valid values or are NULL.
• Be very aware of the difference between, for example, *myVar, myVar, and
&myVar.
• When dealing with a string in a loop, be conscious of the need to check for the end of
the string (the null-termination).

141
15.12. Summary
This chapter has introduced the concept of the pointer, which was defined as a method of
accessing individual data objects anywhere in memory through the use of an address. The
advantage of such a data type is that we can move around memory very efficiently using
simple math on the pointer variable itself. The pointer variables hold an address, which
points to a particular object in memory. Using the contents-of operator, we are able to
extract or place information out of or into the memory occupied by the data type the
pointer points to.
Pointers can be initialized to point to things like characters strings, since, as we saw with
arrays, all a character string is is simply an address which points to the first character of
the character array. Pointers can also be generated using the address-of operator, which
will instruct the C compiler to obtain the address of a data object and place it into the
pointer variable. Thus, using pointers, we can instruct a function to be able to pass more
than one piece of information back from the function by simply passing the addresses of
the data objects to update.
C has a wealth of standard library functions that manipulate strings. In fact, most library
functions that require strings as parameters will use pointers internally to get each
character one at a time out of the passed string addresses.
Also discussed was the fact that an array and a pointer share similar characteristics. An
array name is simply the base address of the array, which is exactly what a pointer is. In
this manner, a C programmer can access data within an array using either pointers or
array notation.

142
Chapter 16. The C Preprocessor
16.1. #define
In this chapter, we attempt to cover a few specific loose ends in our introductory look at
the C programming language. The topics covered here do not affect our overall
understanding of C, but introduce a few concepts that can make our lives easier when
programming in C.
So far, we have used constants by specifying the value of the constant each and every
time we used a constant. For instance, the following lines of code illustrate the use of
constants as we currently understand them:
int a[5] = {0};
int z = 0;
. . .
for (z = 0; z < 5; z++)
{
a[z] = 0;
}
In the above examples, the number 5 is used as a constant in sizing the array and as a
loop control condition. If this value of 5 had a special meaning for our software, and later
in life, it turned out that each time we had to use the value 5, we should have used the
value 6, the programmer will have to re-edit the source code and replace all occurrences
of 5 with 6.
This is of course time consuming, tedious and error prone. A better solution is to direct
the compiler to use an alphanumeric name to represent a constant, which is accomplished
using the #define compiler directive.
The #define directive allows a programmer to give a name to a constant, and whenever
that name is used in the program, a program called the preprocessor, which executes
prior to the compiler and whose job it is to scan header files and convert all directives
into their appropriate values, will replace all occurrences of the name with the constant
value it represents. For our above examples using the constant 5, see how a #define
simplifies coding:
#define IMPORTANT_VALUE 5
int a[IMPORTANT_VALUE] = {0};
int z = 0;
. . .
for (z = 0; z < IMPORTANT_VALUE; z++)
{
a[z] = 0;
}
Everywhere we had used the value 5, we now use the name IMPORTANT_VALUE,
which was declared using the #define directive as being equivalent to the constant 5.

143
 In the first statement, the array a is being sized to IMPORTANT_VALUE, which the
preprocessor will replace with the value 5. Using IMPORTANT_VALUE does exactly
the same thing as using the value 5 directly. Later, to initialize the array to zero, we use a
loop control value of IMPORTANT_VALUE. The advantage of using the #define
directive is that if an important constant will change during the course of a program, we
can easily change all occurrences by changing the specified value after naming the
constant. If we change IMPORTANT_VALUE to 10, the array is automatically resized,
and the loop will still initialize all values in the array to zero with one small change in our
source code.
The general syntax for #define is:
#define name-of-constant value
An important note: there is no semicolon after specifying the value! This is similar to
how there is no semicolon after a #include line. Also, it is common coding style for
professional programmers to try to name constants using totally uppercase letters. In this
way, it is easy to spot constants being used in a program.

16.2. Macros
In many software applications, such as spreadsheets and word processing, the software
package provides a method of performing commonly repeated keystrokes and mouse-
strokes in the form of easy to use macros. In C, a macro is a similar concept, and uses the
power of the preprocessor software combined with the #define directive to allow the
compiler to perform some interesting work.
As an example, study the following directive:
#define cls() clrscr()
The above directive informs the preprocessor that whenever it sees the text cls() in our
code, it should replace it with the text clrscr(). This is a portable method of defining a
clear screen function since not all compilers have a clrscr() function. When working with
a different compiler, you simply have to change the directive to use the appropriate clear
screen function call equivalent (if it exists)!

Useful Tip: Clearing the screen in Visual C++

There is no built-in way of clearing the console (or clearing the screen) in Visual C++.
An alternative is to use system("cls");

Here is another example:

#define AVG(x,y) (((x) + (y)) / 2)
This is another example of the power of the macro. Here, we've defined a macro called
AVG which will take two macro parameters, x and y. The data types for these parameters
are irrelevant since all the preprocessor will do is substitute the expression that appears
later on the line (a search and replace operation).
To use AVG, you may code something similar to:

144
 int a = 0;
int b = 0;
int c = 0;
int d = 0;
. . .
a = 5;
b = 9;
c = AVG (a, b);
d = AVG (c, 5 * b);
Notice how we simply place values into the parameters of the AVG macro, similar to
how we pass parameters to a function. The preprocessor will translate the last two
statements into:
c = (((a) + (b)) / 2);
d = (((c) + (5 * b)) / 2);
which is exactly what an average is supposed to do.
The reason programmers use macros sometimes as opposed to function calls is that a
function call has associated with it. This overhead is the parameter passing and local
variable creation, along with the storage of the return location in memory where the
program can continue when the function completes. If a programmer desires to wrap a
small computation with a meaningful name such as AVG, it may be more beneficial
speed-wise to create a macro rather than a function call.
The extra brackets in the AVG macro definition ensure that there will be no ambiguity
when substitutions are made. This is an essential macro programming concept.

Useful Tip: Macro brackets

Always put brackets around the macro "parameters" in the expression.

Commonly found macros include min() and max(), available in most compiler include
files.

16.3. Conditional Compilation

The last topic to cover in this chapter is conditional compilation. This is a term given to
compiling source code that contains code that should, or should not be compiled
depending on the circumstances. For example, during the debugging phase of a project,
programmers will place extra code into a program to help debug the danger spots of a
program. When the program is deemed to be bug free, normally the debug statements
should be removed. If the program is then released, and a new bug is discovered, the
programmer may need to add all the debugging statements back into a file. This ends up
being a very frustrating process!
The concept of conditional compilation in C is not a new one for C programmers. We
have already seen a powerful method of keeping code from being compiled: simply place
unwanted code inside of a pair of comment delimiters, namely /* and */. Any code
within the pair of comment delimiters will not be compiled just like any comment text is

145
not compiled. The C programmer can then simply add comments around debugging
statements to keep the compiler from compiling such debugging statements when a
program goes to market.
It still becomes a bit of a pain to search through your source code to locate debug
statements in order to comment them out. A better solution is to use the conditional
compilation directives in C.
To begin a section of code which is to be conditionally compiled, use the following
directive:
#ifdef CONDITIONAL_FLAG
and terminate the section of conditionally compiled code using:
#endif
The constant CONDITIONAL_FLAG can be any name the programmer chooses. If the
constant has been defined anywhere prior to the #ifdef statement, the compiler will
compile the conditionally compiled section of code, because the constant has been
defined. If not, the compiler will ignore all statements until the #endif is located.
The advantage of this is that a programmer can use a specific constant, such as the word
DEBUGGING_IN_PROGRESS and simply define or remove the definition of this
constant whenever switching from a debugging to production mode. The conditionally
compiled statements will then either appear or disappear whenever needed, without any
effort on the part of the programmer!
Project maintenance software, such as an IDE or MAKE also allow the compiler to be
informed that a constant exists even before the file is compiled. At the project
maintenance level, flags such as DEBUGGING_IN_PROGRESS can be turned on and
off whenever a new phase of the project is entered!
As an example, study the following code snippet:
#define DEBUGGING_IN_PROGRESS

. . .

#ifdef DEBUGGING_IN_PROGRESS
printf ("we are debugging\n");
printf ("variable x: %d\n", x);
#endif
In the above snippet, since the constant DEBUGGING_IN_PROGRESS has been
defined, when the compiler sees the #ifdef directive, it knows that the two printf()
statements between #ifdef and #endif are to be compiled along with the rest of the
software. By removing or commenting out just the line #define
DEBUGGING_IN_PROGRESS, and recompiling this source code, the two printf()
statements will not be compiled. This feature should tremendously simplify debugging
cycles!

146
 There is a third directive that goes along with #ifdef and #endif, called #else, which
allows either one section or another section to be compiled if a conditional compilation
flag exists or not. When experimenting with code, a programmer can place new, untested
code inside such a conditional compilation statement, and keep the original code in the
#else portion of the conditional compilation sequence. Whenever the flag is present, the
new untested code is compiled, if the flag is not present, the original code will be
compiled. This again may help a programmer through debugging sessions.
Related to the #ifdef is a conditional compilation feature called #if(flag==value). The
value inside the brackets is a conditional compilation flag, however, the #if will test the
flag for a specific value – if the match occurs, then the conditionally flagged code will be
compiled, else it will be skipped. This is often how upgrades to software are handled. If
you need to recreate an older version of an application, or a lesser featured version, use
the #if approach to conditionally compile based on a version number.

Useful Tip: Fixing compiler errors using #ifdef

Use #ifdef to help narrow down what is causing compiler errors. Do this by using #ifdef
to not compile large blocks of code until the error no longer appears. Then reintroduce
small parts of the code block until it reappears again.

16.4. Common Beginner Errors

• Don't reflexively put a semicolon at the end of preprocessor directives.

16.5. Summary
This chapter has introduced number concepts which enhance our understanding of C.
First, we looked at how to reference constants using a name rather than the actual
constant value through the compiler directive #define. Next, we looked at how the
#define directive can aid in creating macros - substitutions for simple repetitive code. A
macro can aid in making a program more portable between compilers and operating
systems. Finally, the topic of conditional compilation was discussed to give the
programmer a taste of high powered debugging capabilities in the C language.

147
Chapter 17. Structures
17.1. What is a Structure?
A structure in C is a data type that the programmer may create to help solve a
programming problem. This ability makes languages which allow programmer defined
data types very popular, since the code you create can be data driven, a term given to
programs which revolve around the data they will work with, rather than code driven,
meaning code which revolves around limitations in the language used.
The term structure itself is usually defined in a dictionary as:
composition, an arrangement of objects, an organization.
In programming, the idea of a structure relates to the language's ability to be able to
organize the primary data types available (char, int, long, etc.) into something which
more closely suits the program's needs. For instance, a hypothetical program must keep
track of manager names, the number of people they supervise, the amount of their budget
and the amount spent by their department. This is an ideal candidate for a structure, as we
have four pieces of information that are related to one another.

17.2. Structure On Paper

The following is a pictorial representation of the structure discussed above. It is often
useful for programmers to organize structures of data in a similar, visual manner before
defining a structure in their software.
Manager Name
Number of Subordinates:
Budget Allotted: $
Expenditures: $
If you had 50 of these managers to keep track of, you may have 50 sheets of paper with
the above information kept inside of a file folder in your file cabinet.
The above organization is something that can be represented in a C database style
program in a number of ways. Currently, we know that if we have to keep track of up to
50 manager names, each no longer than, say, 50 characters, we can create a data object as
such:
char managerName[50][51];
To keep track of the subordinates:
unsigned int subordinates[50];
and finally, the two financial figures:
double budget[50];
double expenditures[50];

148
 With the above array declarations, we now have 4 distinctly separate arrays to manage in
our C program, even though the 4 arrays are intimately intertwined in their significance.
Even though we have four different arrays to manage, it isn't a totally onerous job to
administer such a database. For instance, the following code snippet allows us to enter
information into entry 33 of our database:
char buffer[100] = {0};
int x = 0;

. . .

x = 33;

. . .

printf ("Enter manager name: ");

gets (managerName[x]);
printf ("Enter number of subordinates: ");
gets (buffer);
subordinates[x] = atoi (buffer);
printf ("Enter budget amount: ");
gets (buffer);
budget[x] = atof (buffer);
printf ("Enter expenditures: ");
gets (buffer);
expenditures[x] = atof (buffer);
With just a few lines of C code, we can get information into our database elements.
But what do we do about shuffling information around inside of our database? Currently,
in an array implementation as above, we would have to do the following:
/* copy entry x into entry y */

strncpy (managerName[y], managerName[x], 50);

subordinates[y] = subordinates[x];
budget[y] = budget[x];
expenditures[y] = expenditures[x];
Already you should be able to see that once information is into our array based C
program, it begins to become tedious to always be duplicating lines of code to do similar
things. It is especially infuriating knowing that the four arrays we are manipulating are
actually part of a larger organization, at least on paper. Imagine if we had to keep track of
over 100 pieces of information for each database entry!

17.3. C Structures
This is where C structures come into play. Thus far we have only been able to create
aggregate, or derived data objects based upon the primary data types. With the following
new concepts, we will learn how to create a data type which does not currently exist in
149
 the C language, and learn how to actually declare derived data objects based upon our
own programmer defined data types.
We inform the C compiler we wish to create our own data type through the use of a
structure declaration. The structure declaration informs C that the declarations between
the braces following are not variable declarations, but in reality, definitions of the
component fields that lie within the structure itself.

17.4. Structure Notation

The notation for a structure declaration is as follows:
struct tag
{
data-type field1;
data-type field2;
. . .
};
Once the above declaration is made, a new data type is added to the programmer's
repertoire: struct tag. Think of the tag as the name of the structure. The declaration
above does not allocate any memory for the structure; it simply defines a template for
future variables based upon this template. This template can then be used to declare
simple variables, arrays or even pointers.
If we take the example manager database for instance, we see that we are required to
define four different fields for our structure. We have a manager name field, maximum
50 characters, plus one for the null terminator. We have a field to keep track of the
subordinates of the manager. We have two fields that keep track of the financial
performance of a manager. In our earlier approach to this scenario, we simply created
four arrays for each of the fields to keep this information for the 50 possible managers.
To demonstrate how to create data types in C, we will define a new data type that can be
referenced through the tag managerStructure, which will contain the appropriate fields
to allow us to manipulate the manager data with ease. We will define this data type using
the structure declaration notation described above:
struct managerStructure
{
char name[51];
unsigned int subordinate;
double budget, expenditures;
};
Notice that we begin with the keyword struct, followed by the tag, the name the
programmer gives the data structure itself. In this case, we have tagged the structure
managerStructure. This is simply a name to be used in conjunction with the keyword
struct to allow a C compiler to know which structure you are referring to.
Within the braces of the declaration, you should notice that we have our now four
familiar pieces of information for the manager structure. We have declared an array of 51
characters, and given it the name name. We then declare an integer with the name
150
 subordinate, followed by a declaration of two double precision values, budget and
expenditures.
Although these declarations look remarkably similar to those of a regular variable
declaration, they definitely are not since they lie between the braces of our structure
declaration. Because of the braces, the C compiler can discern between a normal variable
declaration and a structure declaration. To complete the structure declaration, we use a
closing brace, and a semicolon.

17.5. Variable Declarations

So far, the above is only a data type. We are now in the position to actually declare our
first variable with the data type we just created. The syntax is as follows:
struct managerStructure headManager;
The above variable syntax declares a variable called headManager, whose data type is
struct managerStructure. In memory, we have just asked the C compiler to create:
[51 bytes for the name][2 bytes for subordinate]
[8 bytes for budget][8 bytes for expenditures]

Useful Tip: Saving Stack Space Again

Store large local struct variables as static.

17.6. Structure Copy

Although it may appear that we could have done the exact same thing by declaring a
single array of 50 characters, then a single unsigned integer, and then two double
precision variables, we have one very big advantage: you are able to apply the standard
assignment operation against structure variables. For instance, we can do the following:
struct managerStructure headManager;
struct managerStructure otherManager;

. . .

headManager = otherManager;
The last line above will do exactly what is expected: it will update all four fields of the
head manager variable with the four fields found inside the other manager variable. It
performs an automatic string copy, an integer copy, and two double precision copies all
at once. This is usually referred to as a structure copy or block copy.

17.7. Field Notation

To actually get information into our structure variable, or to access information already
stored in the structure variable, we must have the ability to access the individual fields
within the structure. This is accomplished using the structure member notation, which
uses the dot operator:

151
 structure-variable-name.field
The structure variable name serves much like the name of an array: it allows the C
compiler to generate an address to the start of the structure, and internally in your code,
the compiler will generate offsets to each of the fields within the structure in order to
access the information stored there. C structures perform all of the addressing work
automatically by using the above notation. As an example:
printf ("enter manager name: ");
gets (otherManager.name);
headManager.subordinates = 10;
The lines above demonstrate effectively the use of the name.field approach to accessing
the information stored within the structure variable itself.

17.8. Aggregate Structures

Returning to our earlier example of a database that must keep track of 50 managers, we
again can use our structure approach:
struct managerStructure managers[50];

. . .

gets (managers[42].name);
managers[49].subordinates = 10;
managers[3].budget = 100000.00;
managers[0].expenditures = 1234.56;
With the lines above, we have created an array of 50 consecutive manager structures in
memory, each structure still having the same arrangement of bytes as the individual
memory elements. We then set a particular element's fields to the appropriate value.
With this notiation, by using array offset notation, we are referencing a specific manager
entry (i.e.: one of the 50 managers in the array), and once we've isolated one particular
entry, we can use dot notation to access a field within this one entry. It couldn't be more
intuitive than that!

17.9. Pointers and Field Notation

We can declare variables such as:
struct managerStructure *mgrPtr1 = NULL;
struct managerStructure *mgrPtr2 = NULL;
The above declares two pointers to manager structures called mgrPtr1 and mgrPtr2.
Array and pointer manipulations on structures will perform an operation on the entire
block of memory occupied by the structure, unless field notation is used.
When a programmer has a pointer to a structure, field notation will use the arrow
operator:
printf (mgrPtr1->name); /* yes, this will work */

152
 Note that an arrow (->) is used instead of a dot. The arrow is simply a minus sign
followed by the greater than sign.
As an exercise, prove to yourself that the following two statements are identical:
mgrPtr1->budget = 100.22;
(*mgrPtr1).budget = 100.22;
Above, the arrow operator references a field within the structure that the pointer variable
is addressing. In the second example, we are taking the contents of the memory that the
pointer variable is pointing to, which in effect, gives us access to one complete manager
structure. Since we have isolated one manager, we can then use the dot operator to access
a field within this structure, hence we end up doing the exact same operation as with the
arrow operator, albeit in a much more cumbersome manner. Most programmers will
agree that arrow operators are much more convenient than contents-of, followed by the
dot operator!

17.10. Pointer Arithmetic

When moving about in memory using a pointer to a structure, the pointer address math
will reflect the size of the structure. To make this point clear, study the following
statement:
mgrPtr1++;
This statement will increment mgrPtr, but by how much? If the manager structure takes
69 bytes, then we have just incremented mgrPtr by 69! Remember, moving a pointer
always manipulates the address stored in the variable by the size of the data type the
pointer is declared to point to. Since mgrPtr is declared to point to an object organized as
a struct managerStructure, pointer math will manipulate the address stored in this
variable by the size of the structure. This is no different than pointers to any of the
primary data types. A simple rule of thumb is:
Pointer manipulations always alter an address by the number of objects, not bytes.
The C compiler will determine how many bytes to adjust the address by.

17.11. Putting It Together

Here is a small test program to demonstrate the above concepts:

/* STRUCT.C */

#include <stdio.h>
#include <stdlib.h>

struct test
{
int x;
char name[20];
double value;

153
 };

int main (void)

{
char buffer[100] = {0};
struct test a;

/* fill a with values */

printf ("Enter x: ");

gets (buffer);
a.x = atoi (buffer);
printf ("Enter name:");
gets (a.name);
printf ("Enter value:");
gets (buffer);
a.value = atof (buffer);

/* output contents of a */

printf ("x: %d\n", a.x);

printf ("name: %s\n", a.name);
printf ("balance: %f\n", a.value);
return 0;
} /* end main */

17.12. Initializing Structures

Just as we are able to initialize a simple variable, or an array, we can initialize a structure
with information at the time we declare the structure. The following example (which
initializes the headManager variable with the name Bob, 5 subordinates, a budget of
$10000 and an expense account of $2332.99) will demonstrate the required syntax:
struct managerStructure headManager =
{
"Bob", 5, 10000.00, 2332.99
};
Notice how each of the four pieces of information in the initialization corresponds to the
four fields within the struct managerStructure organization. When initializing a
structure variable, you must match initial data with the correct field, otherwise, the C
compiler will abort the compilation with a data type mismatch error.

154
 Useful Tip: Formatting struct initializations
It's considered OK to have the entire declaration, including initializers and braces, on the
same line if you wish.

17.13. Using Structures from the Standard Library

The ANSI C standard library has a number of useful structures already declared for you,
and available for use by including the appropriate header files. Here we will study the
ANSI C library support for determining the time and date of your computer.
In programming, there is often a need to display the current time that the computer
maintains internally. An ANSI standard function called time() can be used to ask the
computer for the current time, using a special data type called time_t. This is a typedefed
data type, the significance of which will be discussed in the next chapter. At this moment,
it suffices to know that the time_t type is identical to an unsigned long data type.
The following is the function prototype, declared in time.h, for time():
time_t time (time_t *t);
This declaration shows that time() requires a pointer to a time_t object, and will return a
time_t object.
The value returned by time() is not of immediate usefulness to a programmer. It is a
special, internal, computer representation of the current time. A programmer must take
this value, and give it to a number of other functions to extract meaningful information.
The following prototype, declared in time.h, is for a function called localtime(), which
will extract the local time information for a computer. Local time is generally referred to
as the time in your time zone. For example, the local time in Toronto will be 3 hours later
than the local time in Vancouver at any given moment.
struct tm *localtime (time_t *t);
localtime() will accept a pointer to a time_t value, which should be filled in using the
time() function call discussed above, and return a pointer to a structure called struct tm.
This structure, declared in time.h has the following organization:
struct tm
{
int tm_sec; /* seconds, 0 - 59 */
int tm_min; /* minutes, 0 - 59 */
int tm_hour; /* hour, 0 - 23 */
int tm_mday; /* day of the month, 1 - 31 */
int tm_mon; /* month, 0 - 11 */
int tm_year; /* years from 1900 */
int tm_wday; /* days since Sunday, 0 - 6 */
int tm_yday; /* days since Jan 1, 0 - 365 */
int tm_isdst; /* daylight savings if non-zero */
};

155
 This is an elaborate structure, which gives you everything you need to know about the
computer's view of the time and date.
Another useful ANSI C function is asctime(), which will generate a text version of the
time, based upon information inside of a struct tm variable. The prototype is:
char *asctime (struct tm *t);
This function will also require the time.h header file. The return value is a character
pointer, which will point to a buffer of text, organized as following example output of
asctime():
"Wed Jun 15 12:37:22 1994\n"
This function gives your software a simple method of obtaining a printable representation
of the time your computer has. The following program will tie these functions together to
print the time and date your computer currently is maintaining:
/* ansitime.c */

#include <stdio.h>
#include <time.h>

int main (void)

{
time_t theTime;
struct tm *t;
char *ptr = NULL;

time (&theTime);
t = localtime (&theTime);
ptr = asctime (t);

printf ("The time and date is: %s\n", ptr);

printf ("The day of the year is: %d\n",
t->tm_yday);
return 0;
} /* end main */
The program will first obtain the system time, using a call to time(). The parameter is the
address of the variable theTime, which is of type time_t. time() will fill in the current
time, allowing us to extract the local time information, by calling localtime(). The
returned pointer, of type struct tm is then given to asctime(), which will convert the
structure into a human readable string. We print this string of characters in the first
printf(), and the second will output which day of the year it is.

17.14. sizeof
An interesting and useful feature of the C language is the operator called sizeof. It is
called a operator because, whenever it is used, the C compiler will not generate code to
call a function, rather, it fills in the actual size of the data object following it instead.

156
 sizeof can be thought of as a constant which happens to equal the number of bytes of the
data object following it. For example:
double things[10] = {0.0};

. . .

printf ("size of array: %d\n", sizeof (things));

printf ("size of manager structure: %d\n", sizeof (struct
managerStructure));
The first line will display the number of bytes inside of the array called things. Since
things consists of 10 consecutive double data objects, this line will display a result of 80
(10 objects, each object 8 bytes long). The second will display the number of bytes in our
manager structure, which will cause 69 to be printed. sizeof will become very useful
when we discuss dynamic memory, and file handling.

Useful Tip: sizeof

You can use sizeof with both variable names and data types.
If you are doing sizeof in conjunction with a data type that is more than one word long
(e.g. a struct), put the data type in brackets.

17.15. Bitfield Structures

Another variation on structures involves using just parts of bytes store information more
compactly. A bitfield structure is a special type of structure which allows a programmer
to define and give field names to bits within an integer. For example, a bitfield structure
for date and time bitfields include:
typedef struct time_bf
{
unsigned int sec: 5;
unsigned int min: 6;
unsigned int hour: 5;
} TIME;

typedef struct date_bf

{
unsigned int day: 5;
unsigned int month: 4;
unsigned int year: 7;
} DATE;
Notice that in a bitfield structure, you can specify the size of a portion of a bitfield
structure, as well as the data type. In the example above, we declare that the day field will
take up 5 bits, the month field will take up 4 bits, and the year field will take up 7 bits.
Using the bitfield fields is very easy. Simply treat them as normal structure fields.

157
17.16. Summary
In this chapter, we introduced the ability in C to create programmer defined data types.
This ability allows a C programmer to create data objects that will enable a simple
solution to a programming problem, rather than trying to code around the limitations of a
language.
There are two parts to declaring a variable to be of a programmer-defined type. The
programmer must first define to the C compiler the fact that a data structure is to be used
in a program. This is done through the use of the structure declaration. The structure
declaration is nothing more than a declaration of the template, or organization of the data
structure. After declaring the structure, the programmer has the freedom to create
variables of that newly defined data type, as if the data type had been in existence
forever. Simple variables, pointers and arrays can be created based upon programmer-
defined structures.
Structures simplify work for a C programmer because they declare variables that contain
the appropriate fields of information as a single unit. A simple assignment statement
using structures has the ability to copy many bytes of data from one location to another
location in memory with almost no effort on the part of a programmer.
The pieces of data that are stored in a C structure are referred to as fields, and using field
notation, a programmer can obtain access to the individual pieces of information a
structure variable contains. Thus, a structure variable can be manipulated as a single unit,
or can be accessed right to the level of the constituent fields, offering tremendous
flexibility and potential to a C programmer.
We looked at the operator sizeof, which can be used to compute the size of any data
object, be it a variable or data type. We also looked at the use of bitfields.

158
Chapter 18. Unions, Enumerations and Type Definitions
18.1. Unions
We've learned how to tell the C compiler that we wish to write code to manipulate data
types that are not normally present in the language itself. This is done using the structure
declaration syntax, and it allows the C programmer to be able to define a data structure to
suit the needs of the data they are manipulating, rather than rework the data to suit the
programming language.
Another approach in C that allows a programmer to define new data types is the union.
The union is similar to a structure in terms of declaration syntax, but they way C
generates code to access the members of a union is vastly different.
A union is defined as a block of memory that has been designated to be able to hold a
number of different data types all at the same time. Technically, a block of memory
cannot hold, say, both a float and a char at the exact same instant, but there are times that
a programmer needs the flexibility to reassign storage within a block of memory to a
different data type to suit the immediate needs of a program. This may sound like a very
far-fetched idea, but with a couple of examples, the confusion should diminish.

18.2. Union Notation

The declaration syntax, as stated earlier, is almost identical to that of a structure
declaration:
union union-name
{
data-type field-name1;
data-type field-name2;
. . .
};
The only difference is that we use the union keyword rather than the struct keyword.
Let's see an actual real-life example of a union:
union stats
{
int rbi;
double era;
};

struct playerInfo
{
char playerName[51];
int position;
union stats vitalStatistics;
};

struct playerInfo player1;

159
Notice that, in the above examples, we have declared a union union stats and a structure
struct playerInfo. The union we have declared is meant to hold, in the same memory
block, either a runs batted in (RBI) total (useful for all players except the pitcher), or the
earned run average (ERA) figure (useful only for a pitcher). If you are familiar with
baseball, you will know that, in the American League, everyone but the pitcher will have
an RBI total, and only the pitcher will have an ERA value. This is an example of two
mutually exclusive pieces of information: never shall the two pieces of information
interfere with one another, resulting in an "either one or the other" type of situation.
The question may arise at this time: "why would you want to store either one type of
information or the other?" Simple! Rather than declaring one structure to hold
information for a pitcher, and separate structure to hold the information for a batter, take
advantage of the fact that neither the twain shall meet, and have the C compiler generate
code to handle the mutually exclusive information.
The structure declared above is fairly simple. An array of 51 characters will be allocated,
followed by an integer that will hold the position number for a particular player, followed
by the union. An example of a position number scheme may be:
#define CATCHER 1
#define FIRST_BASE 2

. . .

#define PITCHER 9
Within our structure, we store the information needed to differentiate amongst the
different positions a player may play on a team, which is the purpose of these definitions.
We could have designed our structure to have a double precision floating point value as
well as an integer to record the RBI or ERA totals. But is this approach feasible when you
have to record all the statistics for all the players in the major leagues? You will be
wasting a double for each non-pitcher, and wasting an int for every pitcher in the
database: memory that will quickly add up with large databases.
The union, therefore, gives us the ability to access a block of memory in various fashions.
With our union union stats, we see that in the same memory, we can hold either the
double or the int. In this manner, we know that if our variable player1 has the position
field set to 9 (pitcher), we can write our software to extract the double precision floating
point value for the ERA. If the position is anything else except pitcher (9), we can write
our software to instead extract the RBI total out of that same memory block - hence, the
same memory block does double duty here! In our statistics database example, this has
the ability to save memory since we no longer require to store two separate pieces of
information: only save the one that makes sense.
The memory usage for the union declared above is:

160
 Table 18-1: Union Offsets
offset 0 byte 0 of int OR byte 0 of double

offset 1 byte 1 of int OR byte 1 of double

offset 2 byte 2 of double

offset 3 byte 3 of double

offset 4 byte 4 of double

offset 5 byte 5 of double

offset 6 byte 6 of double

offset 7 byte 7 of double

The union places the 2 bytes for the int over top of the first two bytes of the double,
using the space effectively since we don't need both pieces of information at the exact
same time in our statistics example.
As should be plain from the above example, unions can be used to save memory when
working with multiple structures that have many common fields. In situations where only
one or the other field is needed, such as our statistics example above, a union is ideal. By
overlaying these fields one on top of another in a union, we have the ability to
dynamically change the type of a portion of memory at will, without resorting to odd
casting syntax or performing unorthodox operations with pointers.

18.3. Union Field Access

The syntax for accessing fields in union is identical to how you access fields in a
structure:
if (player1.position == 9)
player1.vitalStatistics.era = 0.325;
else
player1.vitalStatistics.rbi = 121;
In the above two statements, notice that to access either of the two fields within our
union, we have used the dot operator. The first dot, after the variable name, is used to
access the field within the structure (this field is called vitalStatistics, which is defined as
a union in this case). The second dot is used to access the field within the union itself.
This is a good example of a nested structure/union, as one structure/union is embedded
within another.

161
 Pointer notation is exactly the same for unions as is for structures. A pointer to a union is
simply a pointer to a memory block that can contain more than one representation of data
at one time.

Useful Tip: Union Portability

Please note that unions are often inherently non-portable between different brands of
microprocessors. For instance, a union between an array of 2 characters and a short
integer will compile cleanly on both a Macintosh and a PC compatible computer. On the
PC, when storing a short integer into the union, the low byte is stored first, then the high
byte. On the Macintosh, the high byte is stored first, followed by the low byte.
Code that is written to extract the high and low bytes of an integer will not access the
same portions of the integer on the Macintosh as it did on the PC. This is not a fault of C,
rather, it is a problem that comes up in the internal hardware of different microprocessors
on the market. The idea of the union is fully valid, just its portability is questionable.
In case you are wondering, this issue is known as the “big endian” versus “little endian”
problem in memory organizations. There is no right or wrong way, just different ways!

18.4. Enumerations
Another programmer defined data type that is available in the C programming language
is the enumeration. This is a interesting data type, as its basic purpose is to create
automatically numbered constants for the programmer.
Often a programmer must work with a set of numbers which are to represent different
objects. For instance, if a programmer was to work with the days of the week, a
programmer may define the days of the week as such:
#define MONDAY 1
#define TUESDAY 2
#define WEDNESDAY 3
#define THURSDAY 4
#define FRIDAY 5
#define SATURDAY 6
#define SUNDAY 7
The programmer created the above list of names, and it was the responsibility of the
programmer to define values to accompany the names above. There are many such
examples of lists that need a numeric constant in order to represent them, such as colours,
names of months, and so on.
Sometimes a programmer is faced with the prospect of adding in a new item to such as
list. For instance, what if a new day (called for the lack of a better name, NEWDAY) was
added in between MONDAY and TUESDAY? C code must be altered to reflect that
MONDAY is earlier than NEWDAY, and that TUESDAY is after NEWDAY. The new
list would become:
#define MONDAY 1
162
 #define NEWDAY 2
#define TUESDAY 3
#define WEDNESDAY 4
#define THURSDAY 5
#define FRIDAY 6
#define SATURDAY 7
#define SUNDAY 8
Notice that we had to change the values assigned to TUESDAY through SUNDAY all
because of the addition of NEWDAY. Although editors found on most C compilers these
days would make the corrections reasonably simple, renumbering a list of 1000 different
names would become very tedious.
This is where the enumeration comes into play. An enumeration instructs the C compiler
to automatically generate a list of numbers for a set of data objects the programmer will
create. This numbering action is done at compilation time, and the programmer need not
worry about what numbers are assigned: this is the C compiler's job. The compiler will
take the objects within an enumeration and increment the value given an object by one as
it compiles each object in the list. The first object is given a value of 0, the second has a
value of 1, and so on, unless the programmer overrides this feature. Hence, the value of
the lastest entry in an enumeration listing is the previous value, plus 1.

18.5. Enumeration Notation

The syntax of an enumeration is quite simple:
enum enumeration-name
{
enumerated-object,
enumerated-object = override-value,
enumerated-object,
. . .
};
The second object in the list above is set equal to a special value. This is the override
feature of enumerations: you can override, or specify values for certain objects. For the
next object, the compiler will simply increment the previously specified value, which in
this case was the override value, to generate the next value.
An example of a valid enumeration based upon our example above:
enum daysOfTheWeek
{
MONDAY,
TUESDAY,
WEDNESDAY,
THURSDAY,
FRIDAY,
SATURDAY,
SUNDAY
};

163
 With the above enumeration declaration, the C compiler would automatically assign the
numbers 0 through 6 to MONDAY through SUNDAY. If we now had to add NEWDAY
in between MONDAY and TUESDAY, we simply add it to our enumerated list, and
voila, the C compiler would renumber the list for us with no effort on the C programmers
part whatsoever! The resulting enumeration would be:
enum daysOfTheWeek
{
MONDAY,
NEWDAY,
TUESDAY,
WEDNESDAY,
THURSDAY,
FRIDAY,
SATURDAY,
SUNDAY
};
An added advantage of the enumeration is the ability to declare variables of type enum
name-of-enum. This makes it plain to the reader of a C program that a variable is only
intended to hold a certain subset of values. We can declare the following:
enum daysOfTheWeek weekday;
and perform the following code:
weekday = WEDNESDAY;
if (weekday > TUESDAY)
printf ("past the halfway mark of the week\n");
Hopefully the above examples show the power of the C compiler in doing much of the
work for the programmer as possible. Small features such as this make C the important
and exciting language that it is.

Useful Tip: enum values

You can specify the actual value for individual elements of an enum if you want.

18.6. typedef
When naming structures, unions and enumerations, a shortcut a C programmer may take
is to use the typedef keyword. The typedef is an instruction to the compiler that
whenever a typedefed name is seen, substitute the (normally large and hard to type!)
proper definition in its place. It is similar to a #define statement, but has much more
powerful substitution capabilities and allows proper type checking to be done by the
compiler when declaring variables, parameters, etc.
The generic syntax for a type definition is as follows:
typedef existing-data-type new-type-name;
To explain this concept, it is best to just use an example:
typedef struct testStructure

164
 {
int a;
long b;
double c;
} TEST_STRUCT;

TEST_STRUCT test_variable;
What we've done above is to declare a structure called struct test_structure, but because
we have placed the typedef keyword in front of it, we are actually telling the C compiler
that we are about to give struct testStructure a much simpler name. The simple name
we have given this structure is TEST_STRUCT, and we can now use TEST_STRUCT
as a type declarator just like int or float, etc.
Programmers often create typedef'ed names in upper case, to help visually distinguish
them from other data types in your application - a visual reminder that you've made a
redefinition. You will see other such examples as you proceed through the remainder of
this course!
The typedef can be used to redefine a type name for unions, enums, or just about
anything in the language. In fact, typedef is one method that is used to make software
portable across different computing platforms by taking the machine specific type
declarations (such as int) and giving them new names. A typedef can also make
experimentation with data types simple because the defined data type can be changed
once inside the typedef rather than throughout all occurrences of the original data type.
In either case, a simple change to the typedef name will allow variables declared through
the typedef to assume a different data type. For example, the following are examples of
typedefs commonly found in C code to allow simple software maintenance across
different target machines:
typedef int WORD;
typedef unsigned int UWORD;
A programmer may declare a number of variables to be of type WORD or UWORD, and
the programmer now has the ability to easily change these variables to another data type
without resorting to searching and replacing all occurrences of int to another data type.
Again, let the compiler do the work for you!
As another example, we saw in the previous chapter, the ANSI C library defines a data
type called time_t. This is a simple typedef for a long integer:
typedef unsigned long time_t;
Writing software for operating systems such as Microsoft Windows will uncover the use
of hundreds of different typedefed data types, all designed to easily port current
Windows application code to more advanced versions of this operating system.

18.7. Common Beginner Errors

• Make sure that your use of the union actually makes sense. There might be a better
way to do it (especially if you're using C++).

165
18.8. Summary
In this chapter we explored two other programmer defined data types: the union and the
enumeration. The union looks much like a structure in terms of its declaration, but the
code produced by a compiler to work with unions is vastly different. The union is a
method of being able to extract more than one type of data object from a common
memory block. The union is most often used to make a flexible data structure since a
field can be of one type for one application, and another type for a different application.
The enumeration is a declaration in C that instructs the compiler to enumerate, or
number, a list of constant declarations. The compiler is responsible to generate unique
values for a list of objects, saving much time over the traditional method of using the
#define constant declarations.
Finally, we had a look at the typedef, which is a directive to replace one declaration with
another declaration. It is useful in shortening names of structures and unions, as well as
allowing portability and experimentation in software.

166
Chapter 19. Multi-file Projects
19.1. Introductory Notes
In this chapter, we will look at a few items that may make C programming simpler,
especially when it concerns a large programming project.
Currently, we have been writing software whose source code is contained completely
inside of a single source file, except for a number of standard header files which we
include. There are times that a programmer may want to split up a source file into a
number of smaller, more manageable pieces. This chapter will explore the concepts
behind multi-file compilation, which most C compilers refer to as projects or
workspaces.
As well, we will look into creating your own header files. Finally, we will look again at a
topic which has been bantered about: conditional compilation.

19.2. Projects and Workspaces

The term 'project' or ‘workspace’ is given data which is composed of a number of related
C source code files. The IDE will take the project and create an executable file after
linking the object files from these multiple source files together. The compiler must be
aware of the files in a project, and must be able to create object files for each of these
files before the linker even has a chance of attempting to link together an executable file.
The main reason a programmer will choose to make a project, as opposed to a single C
file program, is that as C programs grow in size, it takes longer and longer to compile the
long program. Imagine a program that is 100000 lines long, taking 10 minutes to
compile. Imagine then that a programmer inadvertently has a syntax error on the 99999th
line of this file. To fix this, the entire file must be loaded into a text editor, and then the
syntax error must be fixed, and then the entire file must be recompiled. The main
drawback behind this is that 99.9% of this file did not need recompilation, only that last
bit of code, hence, an enormous amount of time was wasted in this exercise.
To speed up production of executable code, professional programmers will split large C
files into smaller, more manageable pieces, and use a project management system to get
the compiler to recompile only those C files that have changed since the last build of the
executable. By doing so, the amount of time needed to recompile the executable drops
considerably, as only the changed C files are actually processed by the compiler - the
untouched C files already have valid object modules on disk for the linker to use!

19.3. Variable Declarations

There are a number of issues a C programmer must be aware of before attempting to
work with multiple files. Firstly, when using global variables, it is highly advisable to
create a separate file to contain all globals for a multi-file project. This separate file will
then be compiled by the compiler in order to create the static lifetime storage the project
executable requires. The other files that may reference these globals, however, must

167
 know about the different globals that have been declared. There is a special keyword in C
which allows a C compiler to treat a variable declaration as an indication that the variable
has been properly defined in another file. The keyword is extern, and it is prepended in
front of the type declaration for a variable to inform the C compiler that the global being
declared has been formally defined in another file.
For instance, let us assume the file FOO.C has the following globals defined:
unsigned int variable1;
long anArray[10];
double balance;
If a second file in the project, called BAR.C, is required to access the global variable
balance, BAR.C would make a declaration such as the following:
extern double balance;
The declaration above tells the C compiler that the variable balance is a double precision
floating point value, but the storage for this variable has been declared in another file.
Thus, no storage is allocated when using the extern keyword, as its only purpose is to
allow the compiler to know the data type of an object that already exists in another
component of the project.
Again, this is an extremely important concept to grasp. The C compiler will gleefully
reserve global variable space for your program every time it comes across a global
variable declaration. However, if two or more global variable declarations with the same
name exist when the linker links your object modules together, the linker will exit with an
error, informing you of duplicate global variable declarations. The extern keyword rids
this error from your project!
To continue with this example, if a third file in the project, TEST.C, requires the variable
variable1 as well as the array anArray, TEST.C would make these declarations:
extern unsigned int variable1;
extern long anArray[10];
Notice that the extern declarations are virtually identical to the standard declarations, so
with a good editor, simply copy the globals needed from one file to another, and insert
the keyword extern at the start of each declaration.
Note as well that if each file in a five file project requires a global variable, then each file
in the project must include an external declaration as we've seen above. The more files in
a project, the more tedious this process becomes, as you should hopefully be noticing. A
remedy to this situation will be described shortly.

19.4. Prototyping
In the case of a multi-file project, there are circumstances which require a programmer to
prototype a function which exists in one file and is to be used in another. The use of the
prototype allows the C compiler to know the parameter types, as well as the return type
of the function, enabling the compiler to generate the correct code to call the function.

168
 The format for a function prototype is almost identical for a function declaration. Recall,
for example, this function from earlier:
char *
concatenateString (char *first, char *second)
{
/* code for this function */

. . .

return first;
} /* end concatenateString */
This function will have a function prototype which looks like this:
char *concatenateString (char *, char *);
If the above function is required in all five files of a project, the prototype must be
repeated in each of the five files, much like how global variables must be externally
defined in each file that requires use of the global.

19.5. Common Constants

Again, like the situations regarding common functions and common global variables,
constants which are declared in one file and used in many other files of a project must be
redeclared for each and every file that uses it. If FOO.C uses:
#define TEST_VALUE 100
and so does BAR.C, TEST.C and FILE.C, then each of the last three files above must
have the same declaration so that the compiler will understand what to do with the name
TEST_VALUE. This becomes a tedious and error prone situation to repeat constant
declarations over and over again for a multiple file project.

19.6. Programmer Defined Header Files

With the above problems described, the question that is on all programmers minds at this
moment is "why have a multiple line project at all"? The answer is: program
maintainability. If everything is squashed into a huge, 100000 line file, a programmer
who must maintain that software will cringe each and every time a change must be made,
since it will take hours just to locate the appropriate lines that the change must be made
on, not to mention the fact that it takes a fair bit of time to compile a monster source code
file such as the one described. But again, if we have so many problems with multi-file
projects, isn't one evil better than the other?
There is a solution to the above problem which makes the maintenance of multiple file
projects a breeze. Programmers have the ability to create their own header files, in the
same spirit as headers like stdio.h and string.h.
A header file is simply a chunk of C code which the compiler will compile when the
#include directive is located by the compiler's preprocessor. A programmer needs to add
common definitions such as external globals, prototypes, structure declarations and

169
constants to a header file, and include that header into each file of a multi-file project, and
the problems described above are solved forever!
For example, study the following listing:
The file TEST.H is a programmer defined header file since it does not exist as part of the
standard collection of header files for C compilers. By including TEST.H with each file
in a multi-file project, each source code file will know that three global variables exist,
that a constant called TEST_VALUE has been defined, as well as knowing the exact
parameters and return value for a function called concatenateString. Note that we may
even include headers from within a header, which leads to the common practice of
placing the #include for common header files inside of a programmer defined header file.
You may be wondering what the strange statement using the name __TEST_H__ is all
about. This is a professional programming feature, which prevents the accidental
recompilation of this header file more than once. Compiling a header file multiple times
in the same compile sometimes causes problems. By checking to see if this flag is
defined, the compiler can ensure that the file is compiled the first time it is included, and
after that point, due to the empty definition of __TEST_H__ at the end of this file, the file
will not be compiled again. Again, this is not a crucial feature, but a welcome one in
large, hard to manage projects. See below for more details on conditional compilation.
/* TEST.H */

#ifdef __TEST_H__
#define __TEST_H__

/* common standard headers */

#include <stdio.h>
#include <stdlib.h>
#include <math.h>

/* globals */
extern unsigned int variable1;
extern long anArray[10];
extern double balance;

/* prototypes */
char *concatenateString (char *, char *);

/* constants */
#define TEST_VALUE 100

#endif
For your compiler to access your header file, a slightly modified version of #include is
used:
#include "your_header.h"

170
 Notice the use of double quotes. The quotes tell the compiler to look in the current
working folder for the header file, and if not located there, it should look in the normal
location for system header files. The angle brackets that are normally used inform the
compiler to only look at the system header file directory.
Large programming projects use this feature of the C language religiously. There is
almost no reason to redefine large numbers of previously defined values, variables and
prototypes since a programmer can do it once in header. The header files a programmer
creates can be thought of as a great labour saving feature of C!

19.7. When Will A Compiler Compile A File?

A compiler will always try to keep the executable file as fresh as possible. This means
that when doing a compilation, the compiler will attempt to check and make sure all files
that are in need of compilation are compiled prior to linking object files together.
To be more exact, the control program for the compiler is responsible for doing the
checking. IDEs will always check to see what files need compilation before attempting
any link operation, unless the programmer overrides this checking feature. In most other
compiler packages, especially UNIX based compilers, a special program called MAKE is
invoked whenever a project requires compilation. The MAKE program will check, using
a sequence of rules, as to which files need compilation before proceeding to the link
phase.
Both the IDE and MAKE can be instructed to compile a file if a
header file for that file is changed, since making changes to
constants or variable declarations within header files is a more
than enough reason to recompile a file that depends on that
header. Header files are normally called dependencies for a source
code file.
# makefile example

foo.exe: foo.obj bar.obj

cc foo.obj bar.obj -o foo.exe
foo.obj: foo.c test.h
cc -c foo.c
bar.obj: bar.c test.h
cc -c bar.c
Consult the help pages for your compiler to discover how to make your compilation
package understand you are creating a multi-file project. Again, each different compiler
manufacturer treats this topic in a slightly different, and slightly incompatible manner!
A typical MAKE instruction file (called, naturally, a makefile) appears as follows:
To decipher this file, note that we have an executable file identified on the left side of a
colon. This is the target of the makefile. Next, we specify after the colon the files that
foo.exe depends upon. These are called dependencies. The next line informs the MAKE
program what command will create foo.exe when the dependencies show that a rebuild is
necessary. Here we are invoking a UNIX oriented cc C compiler.

171
 Above, bar.obj is dependant on bar.c and test.h. If either of those files are newer than the
currently existing bar.obj file, then bar.obj will be recompiled using the command line
shown.
foo.exe requires that foo.obj and bar.obj be up to date before it will be linked. By
properly specifying dependencies, if bar.obj has just been created, then foo.exe will need
to be relinked to update it to the latest version.
Think of a makefile being nothing more than a recipe for a compiler and linker to keep
an executable up to date. With large, multi-project files, nothing can beat a makefile for
ensuring that a project is delivered in a repeatable, consistent manner.

19.8. Guidelines for Splitting Files

Here are some general guidelines for determining how to split files in large projects:
• split .c files across general topic areas
• put all prototypes in .h files
• .h files including prototypes will most likely be included in all .c files
• put all global variables in a separate .c file
• put extern references to those global variables in a separate .h file
• put data types and constants that are used in multiple source files into separate .h files

19.9. Example of Splitting Files

Let's say that you had a program that read in formatted financial data from a file, filtered
and summarized it, and created a summary output file. Here's one way (of many!) that
you could split the files:
• input_format.h: contains the struct definition and constants dealing with the
formatting of the input file
• protos.h: contains prototypes
• this file will most likely have a line including input_format.h
• main.c: containing main() and not much else
• this file will have lines including protos.h and (most likely) input_format.h
• get_input.c: containing all functions concerned with getting the data in from the file
• this file will have lines including protos.h and input_format.h
• process_data.c: containing all functions concerned with processing the data once in
• this file will have a line including protos.h
• create_output.c: containing all functions concerned with filtering and outputting the
data
• this file will have a line including protos.h

172
19.10. Common Beginner Errors
• Avoid reusing already existing projects. You might end up with two versions of
main() (one from your old source file and one from your new one).
• Make sure that you set up a Win32 console application.
• Separate include and source files appropriately.

19.11. Summary
In this chapter we explored a number of concepts which allow some sophisticated
programming to be accomplished. Firstly, the topic of multiple file projects was
discussed. Basically, a multiple file project is no different from a single source file C
program except that some precautions must be taken. Any globals must be redeclared as
external to a file, since the compiler will create errors if more than one global by the
same name exists, even in a multi-file project.
As well, function prototypes, structures, unions and constants must be redeclared for each
file that requires them. A simpler approach to this problem is the use of a programmer
defined header file, which contains important information for each file in a project.
A brief discussion was held on the concepts behind how the control software for a
compiler determines how a project will be compiled. This included development
environments and the MAKE program.

173
Chapter 20. Dynamic Memory
20.1. Static Memory Allocation
In our programming to this point, we have worked with two types of memory allocation
for our variables: variables which are allocated when our software begins execution and
variables created when we need them and disappear when we no longer need them.
The C compiler, when generating the code for our program, will also generate the code to
create and destroy local variables as we enter and exit functions, as well as creating and
destroying global variables when our software begins and finishes. Because it is the C
compiler doing this work, many programmers do not realize that memory is being
allocated for the purposes of our software. Recall that local variables are allocated their
space once a function begins. Once the function is complete, the variables are no longer
needed, and the memory they occupied is released back to the operating system that
executes our software.
Static memory allocation is the term given to memory that is allocated for static
(automatically created) data objects. A global variable, local variable, or a table of values,
are good examples of statically allocated memory. The following shows a statically
allocated block of memory:
struct managerStructure managers[10];
If the above is a global or static variable, the memory, which amounts to sizeof (struct
managerStructure) * 10 bytes, is obtained when the software begins, and will disappear
when the software has completed execution. Again, it is the responsibility of the C
compiler to generate the code to create these variables, as well as cleaning up memory to
remove their presence. If the above variable was a local variable, the stack would be
accessed to allocate the memory upon function call, and clean up the memory upon
function exit - again, the work is done for you by the C compiler.
It is interesting to note that a programmer will always know how much memory has been
statically allocated by adding up the number of bytes for all global and static variables,
arrays, pointers, and so on.

20.2. Dynamic Memory Allocation

The term dynamic memory allocation is used by programmers to denote the action of
obtaining memory for data storage only when required, or on a dynamic basis. This is
obviously different from static memory allocation, where we know ahead of time how
much memory is required. In dynamic memory schemes, the programmer may not know
ahead of time how much memory an application will require, hence, a new scheme must
be found to allow memory requirements to grow as time passes.
We have been exposed to a very simple and limited variation of dynamic memory
allocation, in the form of local variables. Although these variables are considered
statically allocated variables, they nonetheless exhibit some of the basic properties of
dynamic memory: these variables do not occupy memory until they are needed (until a

174
 function is called). These variables will also remove themselves from memory when they
are no longer needed (after a function is complete).
Windows and UNIX both manage memory effectively, and encourage programmers to
write software so that memory can be shared properly between applications. Could you
imagine the problems in using Windows if a word processor ate up all memory, and thus
the operating system would disallow running your favourite spreadsheet at the same
time? Obviously, word processors, spreadsheets, etc., must be doing something to ensure
memory is always available for other programs.
But how do we judge how much memory one needs? This is where some data analysis
comes in handy. If you are writing a database that must keep track of, say, 10 objects, it
will be much simpler to allocate static storage for these 10 objects since the amount of
memory required to store only 10 objects is rather small compared to the amount of free
memory in your system.
But what happens if you have to store up to 1000? By statically allocating an array of
1000 objects, you will begin to run out of memory, or at least reduce the amount of free
memory to a minimum, leaving too little for other operations. This is where dynamic
memory allocation comes in handy, and where pointers to structures play an important
role.
Let us use the struct managerStructure example again. If you didn't want to allocate
storage for an array of 10 entries of this data type until you really needed it, you must at
least declare a variable that will point to a future piece of memory for this structure.
Hence, a dynamic memory scheme will always use pointers in some fashion - an array of
pointers to each data object allocated, or perhaps a single pointer to a dynamically
resizable block of memory! In this example, you may create an array something to the
effect of:
struct managerStructure *managers[10] = {NULL};
The line above will declare an array of 10 pointers to manager structures. We have
not yet obtained space for the 10 individual structures, therefore the amount of memory
used so far is only 10 * sizeof (struct managerStructure *) bytes (10 * 4 or 40 bytes).
This is considerably less than allocating 10 complete manager structures, which would
use up over 600 bytes of memory.
Of course, now the question begs ... "how do we allocate memory when we need it?" ...
read on ...

20.3. malloc(), calloc() and realloc()

When we come to a point where we actually need to store information for a manager, we
need to dynamically allocate, or tell the operating system that we now require memory.
This is done using the following function calls, prototyped in stdlib.h:
void *malloc (unsigned int size);
void *calloc (unsigned int numElements, unsigned int
sizeOfElement);

175
 void *realloc (void *old, unsigned int newSize);
Now we begin to get into some tricky stuff!
The notation void * simply denotes that these functions return a generic pointer, (a
pointer to something), and as a programmer, you are required to cast the return value
into a pointer of the appropriate type, otherwise endure warnings about data type
mismatches from the C compiler.
malloc() is the ANSI C standard memory allocation function, which will obtain from the
operating system a block of memory of size bytes, and return a generic pointer (address)
to that block of memory. calloc() is used to obtain more than one consecutive block of
memory of a specific size. You could, theoretically, dynamically allocate all 10 manager
structures using a single call to calloc(). realloc() is used to alter the size of an already
allocated block of memory. This allows a programmer to dynamically modify the size of
blocks of memory that have been allocated, giving your software complete control over
memory usage.
As an example, to allocate memory for one manager structure, use the following code
(assuming our array of pointer declared eariler):
ptrMgr[0] = (struct managerStructure *)malloc (
sizeof (struct managerStructure));
Remember that the size of the structure can be obtained using sizeof. For our manager
structure, sizeof (struct managerStructure) will be converted to 69 by the C compiler.
Note the cast that was performed - malloc returns a generic pointer, hence, we must
convert this to a pointer of type struct managerStructure * to keep the compiler happy,
and once we've done this, we can assign this newly casted pointer into one of the 10
pointer entries we have available in the ptrMgr array!

Useful Tip: sizeof

You should always use sizeof in your software if you need to compute the size of your
data objects in bytes, rather than counting the size by hand. Humans will make mistakes,
compilers (technically) do not!

20.4. NULL
malloc(), calloc() and realloc() are not just limited to allocating dynamic storage for
structures. For example, the following sample function will duplicate a string of
characters and return a pointer to the duplicated string. The function will check the
incoming pointer for validity (check against NULL), and then allocate a block of memory
the correct size (strlen of the incoming string, plus one for the new null terminator). If
that is successful, a string copy is performed (strcpy), and then the new address of the
string is returned. The function itself:
char * duplicate (char *s)
{
size_t len;
char *new;

176
 if (s == NULL)
{
return NULL;
}
len = strlen (s) + 1; /* include null term too! */
new = (char *)malloc (len);
if (new == NULL)
return NULL;
strcpy (new, s);
return new;
} /* end duplicate */
In the function duplicate(), you may have noticed the use of a special constant NULL. C
programmers have agreed upon a manner of recognizing when a memory allocation error
has occurred. A special address marker called NULL is used to denote that memory could
not be obtained. NULL is a predefined constant in stdio.h, and is generally set to the
value 0 (but not always, hence it is safer to use NULL than zero in all programming you
may do).
If there is no memory left in the system, you will find that malloc(), realloc() and
calloc() will return the constant NULL. With this information, you have the ability to
recognize when the system cannot give you any more free memory to allocate for your
data, and you can then report this error to the user using your software, free up any
unneeded memory that your program is using, or use disk techniques to offload the least
used data structures to disk to free up more space.
Always check the return value of malloc(), realloc() and calloc() to see if memory was
obtained or not. Using a NULL pointer in your software is simply inviting a crash of your
software.

Useful Tip: NULL Pointer Accessing

It is interesting to note that advanced operating systems like UNIX and Windows will
automatically terminate your software with a system error if your software ever assumes
that a NULL pointer is a valid address. This is to protect all other applications that may
be running concurrently, so that your inadvertent memory accesses do not corrupt any
other programs. Typically, your software would continue to run, but cause unpredictable
results, usually leading to a lockup of the system.
Remember, always check your addresses!
Course Requirement: Check NULL
Whenever a function can return NULL to indicate an error, check the return value.

20.5. free()
When a program no longer needs dynamically allocated memory, it is customary (in fact,
it is extremely good programming style) to return the unneeded memory blocks back to

177
 the operating system. This is accomplished using the following function, prototyped in
stdlib.h:
void free (void *ptr);
The free() function is used to tell the operating system that the memory pointed to by the
generic pointer parameter ptr is no longer needed by your software, and this memory can
be given to another program, or even your own program for a later allocation.

Course Requirement: Freeing Memory

Whenever allocating memory, always free up memory you do not require, after you have
completed your memory operations. Do not assume that the operating system will do this
for you.
A historical problem of this nature used to happen in the Windows 16 bit system –
memory allocated for applications to hold images, icons or cursor data, if not manually
deallocated by the programmer, were never freed up at all! This would cause memory to
leak, and eventually, 16 bit Windows would die a miserable death. Again, programmers
made the assumption that the operating system would do the cleanup, and unfortunately,
Microsoft forgot a few areas to clean.

ptr is declared as a pointer to void since free() does not care what the pointer is pointing
to, other than the fact that it must be a pointer to an already allocated block of memory. A
program should always free up any allocated blocks of memory when about to exit. This
leaves memory in its original state, and avoids wasting a precious operating system
resource. Please note: never attempt to free() a statically allocated data object, only
dynamic, as unpredictable results will occur!

20.6. Common Beginner Errors

• Always check the return value when you allocate memory.

20.7. Summary
In this chapter we had a look at the notion of dynamic memory allocation, which is used
to request from the operating system memory only when the program really needs it. In
this manner, memory is left free for other system needs such as being able to start up
other programs as your software executes. This is how commercial software allows you
to access the DOS Shell. To know how much memory to allocate using the functions
malloc(), realloc() and calloc(), it is very advisable to use sizeof() to instruct the
compiler to compute the number of bytes in the data object. It is wise to free() memory
when it is no longer needed.

178
Chapter 21. File Input and Output
21.1. Opening Notes
In this chapter, we will begin to explore the concepts behind creating and manipulating
file-based data. The ability to save information into files, as well as being able to retrieve
that information at a later date is of great interest to just about every programmer since
applications like databases, spreadsheets, word processors and editors are rather useless
without having a method of loading and saving the data that is manipulated.

21.2. Stream Based File I/O

We will begin our look at file operations by examining the concepts behind stream
based file input and output (I/O). The term "stream based" is a reflection on the type of
information that will be going into and out of the files on disk: streams of characters,
much like the character streams we place on the screen (character output) and character
streams we receive from the keyboard (character input).
With stream based file I/O, what we are basically doing is placing our textual information
that we would normally be sending to the screen, into a file instead. With stream based
I/O, we can think of the file as simply another location our stream of output characters
can be sent to. The reverse holds true as well: with stream based file input, the characters
we would normally obtain from the keyboard will come from a file instead.
Please note that all functions and constants introduced in this chapter are prototyped in
stdio.h.

21.3. Pre-declared Streams

Stream based file input and output has an inherent ability that every operating system
supplies to us: not only can we send or retrieve streams from a file, we can send and
receive information from other character oriented devices as well, such as the keyboard,
the screen, a modem and the printer. In fact, every operating system has a number of built
in character oriented streams that we can access, using the names:
• stdin: standard input
• stdout: standard output
• stderr: standard error output
• stdprn: standard printer input/output
• stdaux: standard auxiliary input/output
These values are useful when combined with the functions described in the next section
below.

21.4. Standard File I/O Functions

With stream input and output, we use slightly different functions that are identical in
nature to those used earlier, with the added ability of specifying where we want
input/output to flow from/to. These functions are:
179
char *fgets (char *buffer, int length, FILE *stream);
char *fputs (char *string, FILE *stream);
int fprintf (FILE *stream, char *fmt, param1, ...);
int fscanf (FILE *stream, char *fmt, addr1, ...);
Notice that the functions appear almost identical to the standard input and output
functions we have become used to, except for the additional parameter: a pointer to a
special structure called FILE. FILE is a typedef for the streamed file structure, and is
used to maintain specific information about the file that is being manipulated. The fields
within this structure rarely concern a programmer, unless total control over file
operations are desired, hence the structure members will not be discussed here.
The 5 standard streams that we saw earlier are in simply pre-declared pointers to file
structures that every operating system has been designed to handle (ie: stdout is of type
FILE *). Therefore we can print information to the screen using the function:
fprintf (stdout, "hello, how are you\n");
which is identical in execution to:
printf ("hello, how are you\n");
The use of the stream pointer allows us to specify exactly which file stream we want to
work with. If we wanted to send a line of text to the printer, we would accomplish that
feat using:
fprintf (stdprn, "this line goes to the printer\n");
By using different stream FILE pointers, we can send or accept information from
different sources.
A discussion of these functions is not complete without a note on the capabilities of fgets.
This function will accept a line of input from the specified stream. A line is a series of
characters, terminated by a CR/LF pair (the carriage return and line feed pair of
characters - generally, what is generated by hitting the Enter key on your keyboard), or
the number of characters specified by the second parameter has been reached. The
characters are read into the input buffer provided as the first parameter. This is a very
useful replacement for the standard gets function, which has no upper limit on the
number of characters read.
For example, let us assume that a programmer wished to read in a maximum of 50
characters into the array called input. The following code will perform this activity from
the standard input (the keyboard):
char input[51]; /* remember to reserve 1 char
* for null terminator */

. . .

printf ("Enter some data (50 chars maximum): ");

fgets (input, 50, stdin);

180
 Useful Tip: sizeof and fgets()
Use sizeof to get the second parameter to fgets().
Useful Tip: fgets() and '\n'
It should be noted that fgets() will put a '\n' at the end of the string (but before the null-
termination) if the maximum number of characters is not read in. To get rid of the '\n',
you can use the following:
if( strlen(buf) > 0 )
{
buf[strlen(buf)-1] = '\0';
}

A note about the stderr stream: in most operating systems today, the operating system
allows the standard output and input to be redirected at the command line level. You
may have seen commands such as:
DIR > LIST.TXT
which does a directory listing, and sends the listing into a file called LIST.TXT. No
screen output will appear. The redirection always works on stdin and stdout. If your
software wishes to display error messages on the screen even during times of redirection,
the only method available is to use the stderr file stream, since it cannot be redirected (at
least on most standard operating systems).

21.5. How to Open or Create a File

Useful as the above functions may seem, they are still limited by the fact that we haven't
seen a manner in which we can create files on the disk our computer uses. The 5 standard
streams above are connected to character based devices our computer is wired up to.
There still must be a way of creating streams into a file on a disk, which is what we will
look at now.
In C, to create a brand new file, or access an already existing file, we must call a function
to initiate file-oriented operations. This all important function is called fopen(). The
prototype for fopen() is:
FILE *fopen (char *filename, char *accessType);
Notice that calling fopen() will return a pointer to a stream based file. This returned
pointer is then recorded and used by a programmer with the other functions we've seen
above, in the same manner as the predefined character output streams. fopen() will return
a value of NULL if there is any error in attempting to open the file stream to disk, such as
trying to open a file on a non-existent disk, or non-existent subdirectory, etc.

21.6. Filename
The first parameter for fopen() is a pointer to a null terminated series of characters
(string) which contains the filename we want to access. This filename must be a valid
filename for the operating system we are working with. If we wanted to open a file called
181
 TEST.DAT on drive C in the WORK subdirectory on a DOS or Windows system, the
filename pointer would point to the string:
"C:\\WORK\\TEST.DAT"
Notice the use of the double back slash. Recall that since C compilers treat the back slash
as a special character within strings (for the escape sequences like \n or \a), we must
escape the backslash with a backslash. This is a problem on any machine that uses an
operating system like Windows, since DOS uses the backslash to delimit subdirectory
names. UNIX, on the other hand, uses the forward slash to delimit subdirectories, and
there is no need to escape the forward slash.
Please note that the use of the double backslash is only needed for our source code in C
on DOS and Windows machines - when accepting user input, the user need not type
double back slashes, as user input is not escaped like C-based double quoted strings are.

21.7. Access Types For Opening Files

The second parameter is a pointer to a string that specifies the access type of open
operation we are performing. The valid access type strings include:

Table 21-1: fopen() Access Types

"r" Open a file for reading only.
"w" Open a file for writing only. If the filename does not exist on the disk, create a
brand new file. If it does exist, erase the currently existing file so
that we may start fresh.
"a" Open a file for appending (writing) information to. If the file exists, allow new
information to be added to the end of the current information, if not, create a
brand new file.
"r+" Open a file for reading and writing. Old information in a file can be
overwritten with new information.
"w+" Open a file for reading and writing. If the file does not exist, create it. Old
information in a file can be overwritten with new information.
"a+" Open a file for reading and appending. If the file does not exist, create it.
Extra modifiers can be appended to the type string: "t" forces a text file to be worked
with, "b" forces a binary file to be worked with. Binary files are discussed in the next
chapter.
Notice that the "w" type string is lethal to files that already exist on the disk, and care
must be taken to ensure that files that are important are not killed by the errant use of a
"w" file open type. The + types are useful for situations where you need to read
information and write new information over top of old information in a file, the definition
of a random access file.
182
21.8. Closing an Opened File
Once a file has been opened, the programmer must ensure that the file is updated or
terminated properly before leaving a program. This act is called closing a file (obviously
the opposite of opening a file). When closing a file, the operating system ensures the
completed file is intact on the disk, and that the disk directory information is updated to
reflect the date and time of creation, and the proper file size.
To close a stream-based file in C, we use the following function, whose prototype is
defined as:
int fclose (FILE *stream);
This function will accept a parameter of a pointer to a file stream (FILE *), and will
inform the operating system that we are through manipulating the information stored
within the file referenced by the file stream pointer. If any errors occur in closing,
fclose() will return a non-zero value to inform you of such an error, otherwise zero
returned indicates no error.

21.9. Putting It Together

Let's have a look at a sample program called STREAM.C which highlights some of the
concepts we've seen in this chapter thus far.
/* STREAM.C */

#include <stdio.h>
#include <ctype.h>

int main (void)

{
char buffer[100] = {0};
int key = 0;
int x = 0;
FILE *filePtr = NULL;
/*
* ask user where char stream
* output will flow to ...
*/

printf ("where do you want output? ");

printf ("\tF = file\n");
printf ("\tP = printer\n");
printf ("\tS = screen\n");
while (1)
{
char input[10];

gets (input);
key = toupper (input[0]);
if ((key == 'F') || (key == 'P') || (key == 'S'))

183
 {
break;
}
printf ("try again!\n");
}/* end while */

/*
* now get the appropriate stream pointer
*/

switch (key)
{
case 'F':
/*
* get a filename from user
*/

printf ("enter a filename: ");

gets (buffer);

/*
* create new file on disk
*/

filePtr = fopen (buffer, "w");

if (filePtr == NULL)
{
printf ("error in opening file!\n");
return 1;
}/* endif */
break;

case 'P':// this case doesn't work under WinXP

/*
* use standard printer file pointer
* (does not exist under WinXP)
*/

filePtr = stdprn;
break;

case 'S':
/*
* use standard output file pointer
*/

filePtr = stdout;
break;

184
 }/* end switch */

/*
* generate some character stream output
*/

fprintf (filePtr, "this is text ");

fprintf (filePtr, "for a stream\n\r");
for (x = 0; x < 10; x++)
{
fprintf (filePtr, "%d times %d = %d\n\r",
x, x, x * x);
}/* end for */
fprintf (filePtr, "bye!\n\r");

/*
* if we created a file on disk,
* we must close it before leaving
*/

if (key == 'F')
{
if (fclose (filePtr) != 0)
{
printf("File cannot be closed\n");
}
}
return 0;
}/* end main */

185
Useful Tip: General program structure
Here's an example of general program structure that commonly goes along with file
processing:
1. Do stuff that happens at the end of the program.
1.1 Declare variables and initialize them.
1.2 Open files.
1.3 Perform other one-time-only tasks.
1.4 Sometimes, read the first record from the file.
2. Perform the main loop.
3. Do stuff that happens at the end of the program.
3.1 Handle summary data.
3.2 Finish up any output.
3.3 Close files.

In this program, we ask the user to choose a destination for the character output. The
choices are file, screen or printer. Once the user has chosen an appropriate destination,
the program will obtain a valid file stream pointer for the output destination. For both the
screen and printer, we set the variable filePtr, which is of type FILE * to stdout and
stdprn respectively. We are taking advantage of the built in streams for these two output
destinations.
For the file destination, we proceed to ask the user to enter a filename. Once entered, we
call fopen() with a pointer to the buffer where the user had entered the filename, and a
type of "w", which will create a brand new file on disk (or overwrite an old file). If
fopen() succeeded, filePtr will be set with a valid file stream pointer address, if not, an
error message is issued to the user.
Once we have a valid file stream pointer, we can use it along with the file stream output
functions. Here we've used fprintf to illustrate getting character based output into a
streamed file.
Finally, if we had opened a file on disk, we must remember to close that file, lest we have
the operating system improperly truncate the file. This is the purpose of the final if
statement, where we check if the option chosen was to print the output to a file. If so, we
will close the file, terminating the file correctly.

186
 Course Requirement: Closing files
Most operating systems will automatically close opened files when the application
terminates. This is not the most appropriate technique to use, for the same reasons as for
dynamic memory. To ensure full portability of software across all operating systems, you
must always manually close your files when no longer required, to avoid potential
operating system deficiencies.

21.10. Common Beginner Errors

• Always check the return code from file-related functions. If there is an error, do
something about it.
• Only use files that have been successfully opened.
• Watch out for the '\n' in fgets() buffers.

21.11. Summary
In this chapter, we have looked at the concepts behind character stream oriented file input
and output. The term "character stream" refers to the fact that the information to be save
in files is character, or ASCII oriented. We have seen that virtually all C compilers
support five standard file input and output streams including the screen, keyboard, error
display, printer and modem. A C programmer can use this fact to write very simple
software to print information to a printer, or to extract characters from the serial port of
the computer.
We also saw how to manipulate a file on a disk. The use of the fopen() and fclose()
functions allow the programmer to create/open and terminate files whenever needed. A
programmer should always close an opened file since (depending on the operating
system) the operating system will not always do this job for you correctly.

187
Chapter 22. Binary File Input and Output
22.1. Binary Files
In this chapter, we will build upon the concepts learned in the chapter on stream based
file I/O and learn about the methods available in the C programming language to transfer
binary information to and from a device or file.
The term "binary information" refers to the fact that the information is not in a textual, or
human readable format. An executable program file in Windows is considered a binary
information file since, if you took a brief look at the information stored within that file,
you would not be able to make heads or tails of it. The binary information within such
files is only useful for the task it was created for. An executable binary is useful only to
the operating system (which allows such a file to be executed).
This also applies to information your own software may need to store. We've seen
examples of data structures which contain a great deal of different types of information,
some if it textual (character strings and buffers), but most in the form of double precision
values, integers, long integers, and so on. These values are best stored in their natural
state directly inside of a file, and binary file information transfer allows simple exchange
to and from a disk of this type of information.
It should be noted that textual information, such as the stream file input/output we looked
at in the last chapter, can also be considered binary information, only we just happen to
be able to visually inspect the data stored within such streamed files.

22.2. Stream Based Binary File Techniques

We will begin to explore binary file I/O using variations on the functions we looked at for
stream-based file I/O. It turns out that C allows a programmer to open files or devices in a
streamed output format, and afterwards, read or write either streamed or binary
information to or from such a file. Since we are already somewhat familiar with opening
and closing of streamed files, we can look immediately at how to get binary information
to and from such files.
As an example, we will use the following data structure to illustrate a binary block of
data in the computer's memory:
typedef struct testData
{
double balance;
char name[20];
short position;
long socialInsuranceNumber;
} TESTDATA;

TESTDATA theData;
The typedef defines a structure called TESTDATA that contains four different data
types: a long integer, an integer, an array of 20 characters, and a double precision floating
188
 point value. The variable theData is a simple variable based upon the above structure.
theData occupies 34 bytes of memory.
At this juncture, we know how to place the character string into a file. Simply use the
following:
fprintf (outFile, "%s\n", theData.name);
and the name field of the variable theData will be sent into the file pointed to by outFile.
We are of course assuming that a file was correctly opened using fopen() earlier to allow
outFile to point to a valid FILE structure.
Now suppose a programmer wishes to save not only the name field in the file, but the
balance, socialInsuranceNumber and position fields as well. Currently, the only
method at our disposal for sending that information to a file is to perform a number of
fprintf() calls and format the information in a manner that a fscanf() can retrieve the
information later. Although such a scheme may work, it is exceedingly inefficient, since a
file on just about any operating system in use today has no more problem storing the
bytes which make up a double precision floating point value than the bytes storing the
name field. The operating system could care less if the information it is to save on disk is
human readable or not.
This is where binary file I/O comes into play. Firstly, the file must be opened using
fopen() in a binary format. This is done by appending a b after the letter which represents
the file-opening mode. To read a binary file, simply use a type string of "rb" or "r+b",
and to create a binary file, use "wb" or "w+b". Next, we need to somehow inform the
compiler we wish to transfer binary information. The following functions are employed
in the ANSI C standard library of functions (prototyped in stdio.h) to place or retrieve
complete blocks of information to and from a streamed file or device:

22.3. fread() and fwrite()

The following prototypes are instrumental in working with binary data:
unsigned int fread (void *ptr, unsigned int size,
unsigned int nitems, FILE *stream);

unsigned int fwrite (void *ptr, unsigned int size,

unsigned int nitems, FILE *stream);
Both fread() and fwrite() operate on blocks of information, rather than character based
strings. Notice that for both functions, as was the case for free(), the first parameter is
declared to be a pointer to void. This simply means that they will accept any type of
pointer, with the keyword void being a convenient method of saying "pointer to
anything".
The second parameter of both functions is an unsigned integer that specifies the size of a
single unit of the data block to transfer. With an unsigned integer, the maximum block
unit size, in 16 bit DOS, is 65535 bytes (64K in total), or in a modern 32 bit operating
system, over 4 billion bytes in total.

189
 The third field is another unsigned integer that specifies the total number of units to be
transferred. The maximum number of bytes to be transferred at one time by either of
these calls is size * nitems which amounts to a huge number of bytes!! Although capable
of massive amounts of information input/output, typically these functions are called to
transfer only one block at a time, or one structure array at a time.
The final field is, of course, a pointer to a FILE structure, which specifies the stream we
are writing to. Remember, this stream must be pointing to a binary file, opened by
appending 'b' to the standard type string for fopen
The return value of these functions indicates the total number of complete units that were
transferred using these functions. If a value less than the number of units specified is
returned, it is generally assumed an error occurred during transfer. This error could be
something like end of file, not enough room on disk, and so on.
When performing error checking on our file output, the return value of these functions is
a convenient value to check - if the returned number of blocks accessed does not match
the number of blocks requested, then report an error. Otherwise, assume all is well! What
could be simpler than that!
With our structure we defined earlier, we could save away the contents of the variable
theData by using the call:
fwrite (&theData, sizeof (TESTDATA), 1, outFile);
We can read the information back into main memory by using:
fread (&theData, sizeof (TESTDATA), 1, outFile);
To transfer blocks of information to and from disk it is simply a matter of calling fread()
and fwrite() with the appropriate block size and unit numbers. These functions are as
simple to use as the character stream functions, with the added ability of being able to
read and write binary blocks.

Useful Tip: Use sizeof with fread() and fwrite()

To get the sizes for use with fread() and fwrite(), use sizeof.

22.4. Putting It Together

The following is a sample program which uses all of the file input/output functions we've
seen so far:
/* fwrite.c */
#include <stdio.h>
#include <stdlib.h>
#include <math.h>

typedef struct
{

double balance;
char name[20];
190
 int position;
long socialInsuranceNumber;
} TESTDATA;

int main (void)

{
char buffer[100] = {0};
FILE *fp = NULL;
TESTDATA theData;

/*
* obtain information from user
*/

printf ("Enter name: ");

gets (theData.name);
printf ("Enter balance: ");
gets (buffer);
theData.balance = atof (buffer);
printf ("Enter position (0 to 9): ");
gets (buffer);
theData.position = atoi (buffer);
printf ("Enter S.I.N.: ");
gets (buffer);
theData.socialInsuranceNumber = atol (buffer);

/*
* save information to disk
*/

fp = fopen ("test.dat", "wb");

if (fp == NULL)
{
printf ("Error: can't open for writing\n");
return 1;
}/* endif */
fwrite (&theData, sizeof (TESTDATA), 1, fp);
if (fclose (fp) != 0 )
{
printf("Cannot close file\n");
return 1;
}

/*
* now reset the fields, and
* read the block back in
*/

191
 theData.name[0] = 0;
theData.position = 0;
theData.balance = 0.0;
theData.socialInsuranceNumber = 0L;

fp = fopen ("test.dat", "rb");

if (fp == NULL)
{
printf ("Error: can't open for reading\n");
return 2;
}/* endif */

fread (&theData, sizeof (TESTDATA), 1, fp);

if (fclose (fp) != 0 )
{
printf("Cannot close file\n");
return 2;
}

/*
* print data to screen
*/

printf ("Name: %s\n", theData.name);

printf ("balance: %f position: %d SIN: %ld\n",
theData.balance, theData.position,
theData.socialInsuranceNumber);
return 0;
}/* end main */

22.5. Other Stream Based Functions

A few other functions to look include:
int fseek (FILE *stream, long offset, int whence);
The fseek() function allows a programmer to move about within a streamed file using the
offset parameter to specify how many bytes to move. The parameter whence is set to one
of the following:
SEEK_SET (from the file beginning)
SEEK_CUR (from the current position in file)
SEEK_END (file end)
With this function, movement relative to start, current or end of file is possible, allowing
any file traversal to be accomplished. A value of zero returned specifies that no major
errors have occurred. A note of warning: depending on your compiler, the
implementation of fseek() will not always return an error if the end of file is breached,
thus take care when using this function.
void rewind (FILE *fp);

192
 This function will reset the file position back to the start of the file on disk. This is
identical to fseek (fp, 0L, SEEK_SET);
long ftell (FILE *fp);
This function will return the current position within a file, or -1 on error. It is useful for
recording a specific location in a file, in order to be able to return to it later. In the
following example, a function called fileSize() shows how you can determine the size of
a file:
long
fileSize (char *filename)
{
long pos = 0L;
FILE *fp = NULL;

if ((fp = fopen (filename, "r")) == NULL)

return -1L;

fseek (fp, 0L, SEEK_END);/* to end of file */

pos = ftell (fp);
fclose (fp);
return pos;
}/* end fileSize */
Here we are opening a file, and we will return -1L if an error has occurred. If successful,
we reposition, using fseek() to the end of the file, read the position using ftell(), and close
the file. We finally return the size to the caller.
int feof (FILE *fp);
This function can inform you through a non-zero return value if the end of a file has been
breached. Note that you must actually pass the end of the file before feof() will inform
you.

22.6. Common Beginner Errors

• Always check the return codes from file-related functions.
• Make sure that you distinguish between text and binary files.

22.7. Summary
In this chapter we explored the concepts behind binary file input and output. Binary is a
term used to describe data in its natural form. With binary file transfer, a programmer is
able to extract the data within integers, double precision floating point values, or even
structures and unions with virtually no effort whatsoever.
We looked at the use of streamed binary file transfer methods, using the functions fread()
and fwrite(). These functions operate on files that have been opened using fopen(). These
functions allow transfer of blocks of data to and from disk, amounting to large amounts
of file transfer with very little work on the part of the programmer.

193
Chapter 23. Parameters To main()
23.1. Introductory Notes
Another interesting feature of the C language is the command-line arguments that are
passed to the main() function. In reality, DOS, UNIX (or any operating system for that
matter) will always supply three parameters for your main() to use, if it chooses.

23.2. main() Parameter Example

As an example, study the following code fragment:
int main (int argc, char *argv[], char *env[])
{
. . .
} /* end main */
The three parameters specified for main() tell the main() function everything it needs to
know about what the user typed on the command-line to start your program. This comes
in useful for something like the TYPE command in DOS or “cat” command in UNIX:
these applications must have one argument, the filename to display on the screen. Hence,
the command must know how many arguments have been supplied, and what the actual
arguments are.
argc is an integer which tells you how many arguments are on the command-line.
Normally, with no arguments, argc is 1, meaning that the only thing the user typed is the
name of the program itself. The second parameter, argv, which is defined as an array of
character pointers, will have element 0 set to the address of a string which is the name of
the program. Other parameters will be set in elements 1, 2, etc of this array. If you had a
program called FOO, and you typed on the DOS command-line:
C:\> FOO HELLO HOW ARE YOU
argc will be set to 5, and argv[] will be set to:
argv[0] = a pointer to FOO
argv[1] = a pointer to HELLO
argv[2] = a pointer to HOW
argv[3] = a pointer to ARE
argv[4] = a pointer to YOU
The spaces between arguments on the command-line delineate the individual arguments,
which in turn tells the operating system how many character pointers to give to the
main() function.
env is available from most operating systems, and is a list of environment strings that
are present in your computer's memory. These strings inform various programs about
configuration information. For example, a common environment variable is:
PATH=C:\;C:\DOS;C:\WINDOWS;C:\GAMES

194
 and tells DOS where to locate executable programs. You have access to these strings
through the character pointer array env. The last pointer in the array will have its value
set to 0 (usually called a NULL pointer).
Because both env and argv are arrays of character pointers, you can use these character
pointers in any software that works with strings. For example, to output argument number
two in the array argv as a string on the display, use:
printf (“argv at offset 2 is: %s\n”, argv[2]);
The %s formatting code can be used in a printf() call to output an argv array member or
an env array member.

Useful Tip: argv[] Validity

Never access argv elements without first checking argc to make sure that the argv array
index is valid.

23.3. Common Beginner Errors

• Always check argc to see if argv's array index is valid.

23.4. Summary
main() can get information from the system through its parameters. argc and argv are
useful for getting arguments from the command line. env can provide environment
variables as strings.

195
Chapter 24. Operating System Functionality
24.1. Introductory Notes
In this chapter, we will explore how to work with some the functionality that your
operating system gives you, in terms of disk and other low level resource access, deep
within the bowels of your operating system.
An operating system (OS) is a special program that generally needs to be loaded first,
before any application software is loaded. The OS will provide rudimentary access to
data on your system, by supplying command-line or point and click access to files, notes,
word processing documents, and so on. The OS is responsible for maintaining this data
on long-term storage.
The operating systems used today are incredibly powerful compared to those of only a
decade ago. High-level systems like UNIX have migrated down to the level of desktop
users, through Linux. Windows XP and Vista are now de-facto standards for window-
based (graphical) systems. Smaller operating systems such as QNX or Embedded Linux
are often used in "real time systems" where access to incoming data is crucial and time-
sensitive. Note that there is no one OS that is suitable for all tasks, as each have their
strength or weaknesses.
In relation to files, your OS is responsible, when working with disks, of determining what
type of media is being used, the size of the disk, the capacity of the disk, the number of
heads, cylinders, tracks, and so forth, without ever losing a single byte of information a
user has requested to be saved. Because of these rather enormous responsibilities, the
makers of the various stable versions of your OS ensure that all new types of disks are
taken into account so that no matter what type of hardware is connected to your PC, there
is a tested and proven method of getting information to and from that hardware.
As a programmer in C, you will find that occasionally, there is a need to get closer to the
disk than the file I/O functions we've seen so far. For instance, there may be times you
will want to display to the user the files in a certain directory, so the user of your software
can see at a glance what filenames have already been chosen. Of course the user can just
blindly type the name in, and hope that a file doesn't already exist with that name, but it is
always more user friendly to show as much information to the user as possible to help
that user make informed decisions when working with your software.
With this chapter, we will take a look at some of the more useful functions that operating
systems supply to a C programmer so that we can get information such as the above with
very little effort on the programmer's part.

24.2. Listing of Files

In most OSes, there is a commonly used command to list files. In DOS, it is DIR. In
Linux, it is ls. DIR or ls is used to ask the OS command interpreter (the software that
understands what you are asking for) to list all of the files in a certain location on disk.
The OS will take the options for the listing command (which are called arguments), and

196
 begin to search the required area of the disk to see what files exist there. DIR will then
report to you pertinent information about all of the files that it finds.
With the listing command, you should be aware that an OS will accept wildcard options.
For instance, on a DOS system, if you wanted to find all C files within the directory
called WORK on drive C:, you would enter the command:
DIR C:\WORK\*.C
A similar capacity exists on a UNIX system using the generalized ls command:
ls /work/*.c
The OS will search that area of the disk, and list whatever files exist that end with an
extension of .C.
To achieve this same result in our software, we must ask the OS to supply to our software
a listing of files in the desired location on a disk.

24.3. Windows 32-bit Search Functions

For 32-bit Windows programs, both graphical and command-line based applications can
use the following Windows-specific functions:
#include <windows.h>

HANDLE FindFirstFile (char *filename, WIN32_FIND_DATA *w);

BOOL FindNextFile (HANDLE h, WIN32_FIND_DATA *w);
BOOL FindClose (HANDLE h);

/* example */

WIN32_FIND_DATA fd;
HANDLE h;

. . .

h = FindFirstFile ("*.C", &fd);

if (h != NULL)
{
do
{
printf ("Name: %s, Size: %ld\n",
fd.cFileName, fd.nFileSizeLow);
} while (FindNextFile (h, &fd);
FindClose (h);
} /* endif */

197
 Useful Tip: Compiling Projects involving FindFirstFile()
In order to get FindFirstFile() to work properly in Visual Studio, you must change the
default character set for your Project to be Multi-Byte instead of Unicode. This is done
under the Project Properties.

24.4. Changing Directories

When information must be searched for on a particular location on a disk, but the position
differs from installation to installation of the software, a programmer may want first to
switch to that particular location prior to doing any directory searches. This can be
accomplished by telling the operating system to change to the desired working directory.
On Windows, this is accomplished through the use of the _chdir() function:
#include <direct.h>

int _chdir (char *path);

The desired directory path to change to is passed as a parameter. The function will return
-1 if an error has occurred.

198
24.5. Renaming a File
Often a programmer desires to change the name of a file. Such a situation occurs most
frequently when manipulating an important file. For instance, if a database is to generate
a report, the database program will usually try to rename any existing report file to a
backup file, before writing the new report. With this action, the database can output the
desired report, while keeping a backup of the old file still on disk.
In C, we can rename a file using the following function:
#include <stdio.h>

int rename (char *oldname, char *newname);

This ANSI C function will return 0 if successful, -1 if unsuccessful. The two parameters
are simply pointers to the old name of the file and the new name of the file. An error will
occur if rename() discovers that the original file does not exist, or if the new file name
already exists on the disk, or if the new name is invalid. An example of this command is
illustrated in the following program, RENAMEIT.C. (Note: we cannot give this program
the name RENAME.C since the executable will then have the same name as the DOS
internal command RENAME!)
/* RENAMEIT.C */

#include <stdio.h>

int main (int argc, char *argv[])

{
if (argv != 3)
{
printf ("Error! Need old and new names.\n");
return 1;
}/* endif */
if (rename (argv[1], argv[2]) != 0)
{
printf ("Error during rename.\n");
return 2;
}/* endif */

return 0;
} /* end main */
The sample above will use the rename() function to rename the file specified as the 2nd
argument on the command line (remember, the name of the program is always the 1st
argument on the command line), and to the file specified as the 3rd argument. Note that a
rename is also a file mover - to move the file to another directory on the same drive is no
different than just supplying a new name! Give it a try and see in action!

199
24.6. Deleting a File
Along with the above operation, a programmer will often desire to delete a file from disk.
This is done using the following ANSI C function:
#include <stdio.h>

int remove (char *filename);

It too returns 0 if no errors occurred, -1 if an error exists.

24.7. The system() Call

If you want to make use of an operating system facility that is not readily available
through a function call, you can use the system() function. system() is compatible with
all operating systems, hence, it is valuable for portability.
#include <stdlib.h>

int system (char *command);

This function will pass the string pointed to by command to a copy of the command
interpreter available on your machine and the command interpreter will in turn execute
this program.
As an example:
/* SYSTEM.C */

#include <stdio.h>
#include <stdlib.h>

int main (void)

{
system ("dir *.c");
system ("mem");
return 0;
} /* end main */
It should be noted that the use of the system() function call is generally frowned upon
because it is very resource-intensive (it starts a completely new command interpreter).

24.8. Common Beginner Errors

• Make sure that the call that you are using will work outside of your IDE and your
particular computer.

24.9. Summary
In this chapter, we explored the topic of extracting low-level information from the
operating system of a computer. We studied how to obtain information such as file
listings from the Windows operating system so that a program can inform a user that
certain files already exist on disk. Such information is very valuable to a user of a

200
program since it enables the user to make informed choices when it comes to saving files
of data.
We also explored the ability of a C programmer to change working directories and drives,
rename and delete files from disk. Again, offering options in a software package such as
these allow much flexibility to the user of a program.
Finally, we took a look at the system() function to perform actions that aren't readily
available through other means.

201
Chapter 25. Recursion
There are times in programming when it becomes necessary to call a function from inside
the function that is currently executing. Sometimes when a function is called, it calls a
second function, which in turn will call the original function. These situations are usually
referred to as recursion, and the act of recursion is simply the act of calling oneself a
second time before the original call has had a chance to complete.
It may appear that the above idea is horribly complicated, and rather stupid as well! There
are actually a number of advanced algorithms which benefit from this idea of recursion to
efficiently perform a task.

25.1. Recursion Example

The following program, STARS.C, utilizes recursion to output a certain number of
asterisks to the screen. Note, only a single if statement is used to control the sequence of
events in this example. /*
* STARS.C
*/

#include <stdio.h>

void stars (int n)

{
printf ("*");
if (n > 1)
{
stars (n - 1);/* recursion! */
}
}/* end stars */

int main (void)

{
printf ("printing 5 stars ... ");
stars (5);
return 0;
}/* end main */

In the main() function above, we are calling the function stars() with an integer
parameter of the constant 5. We are telling stars() that we desire 5 stars to be output to
the screen. stars() begins naturally enough, printing a single star.
The if statement that follows instructs the compiler to generate code to check if the
current value of the parameter n is greater than 1. If so, another call to stars() is made,
except this time we tell stars() to print one less star!
Notice that we have called stars() a second time, before we've finished executing stars()
the first time. This is where the recursion takes place. In fact, stars() will continue to be

202
 called, each time receiving a parameter one less than the previous call, until the parameter
n is equal to 1. Once this point is reached, the function terminates normally, and all
previous functions will too terminate as coded. This is how a recursive function call
completes - it "unrolls" itself back to the original caller.
Although this is a rather trivial example of a recursive algorithm, it nevertheless
demonstrates how this concept works. In real life, we are exposed to recursive situations
daily. Take, for instance, the task of locating a file on a DOS hard disk. You will perform
DIR commands, either locating a file, or locating a potential sub-directory where a file
may be stored. You will then switch into that sub-directory, and do another DIR. If you
find another sub-directory, you can switch there too. Once you've exhausted a search in a
particular directory, you must switch back to the previous level. This is exactly how the
stars() function performed its recursion.
The main disadvantage of recursion is that if it is not managed correctly, eventually,
memory resources within your program will be exhausted. The CPU has only a limited
amount of space where it can keep track of parameters to a function, and by recursively
calling a function, we slowly chip away at this memory. For this reason, most
programmers steer clear of recursion unless it is absolutely necessary, since other flow of
control statements can suffice in most cases, and are easier to maintain as well.

25.2. Recursion vs. Looping

It should be noted that the real-life uses of recursion are quite limited. When learning
about programming for the first time, it is tempting to use recursion instead of
implementing a simple loop. Resist the temptation. Loops are almost always superior,
due to their simplicity and clarity. In other courses, you will learn about situations in
which recursive algorithms are desirable (e.g. traversal of binary trees). Wait until then.

Useful Tip: Recursion Usage

Don't use recursion unless you are absolutely certain that it's the best way to do the job
(or unless you're told to do so in an assignment ).

25.3. Common Beginner Errors

• Always make sure that the recursion can end.

25.4. Summary
In this chapter, we learned about calling a function from within itself (or another function
called by itself, etc.). Recursion has a place in programming but it should not be used to
replace proper design using loops.

203
Chapter 26. Advanced Pointer Usage
26.1. Pointer Review
In this chapter, we will look at some advanced programming techniques that utilize and
expand our knowledge of pointers.
As you recall, a pointer variable is simply a piece of memory sized large enough to hold
an address for the computer system we are working with. The pointer variable is an
address variable. The data type associated with a C pointer variable declaration is just a
reference to the type of data the address points to. For instance, a pointer to a char is an
address of where in our computer's memory a character (or byte) of memory may exist.
The pointer allows us to access memory in the form of the data type the pointer variable
is declared as. A double pointer allows a program to view memory in chunks of 8 bytes,
since a double precision floating point data object is 8 bytes in size.
It should be noted that a common term used to describe the act of referencing an object in
memory via an address is indirection.

26.2. Pointers to Pointers

In C, it is possible to declare that a pointer variable will contain an address of where
another pointer exists in memory. Since a pointer variable is used to hold an address, it
should be quite apparent by now that an address can point to anything, so why not
another pointer?
The syntax for declaring a pointer to a pointer is as follows:
data-type **name-of-variable;
Notice in the syntax the use of **. The double asterisk is used to inform the compiler that
what we have is a pointer to a pointer. The single asterisk is used to declare a pointer to a
data object, thus two asterisks in the declaration declares a pointer to a pointer to a data
object.
Some examples of valid pointer-to-pointer declarations (without initializations) are:
int **x;
double **ptr;
char **argv;
The first example declares x as a pointer to a pointer to an integer. The second example
declares that ptr is a pointer to a pointer to a double in memory. It should be no surprise
then that argv as defined above is a pointer to a pointer to a character in memory.
To access the data that a pointer to a pointer points to, we must use double indirection.
We must first use the contents-of operator to view what is in memory that the variable
itself points to. Then, we must use this address to peek at what is in memory that this
second address points to. To assign the letter 'A' into the character using the pointer that
argv is pointing to, use:
**argv = 'A';

204
Notice that in our code, we again use the asterisk, except, as for normal pointer
operations, the asterisk is now the contents-of operator. Using two of these operators tells
the compiler that what we want is the "contents of the contents of argv".
Study the following scenario:
argv contains address 10124.
In memory at address 10124 is the address 1779.
In memory starting at address 1779 is the character array
"hello".
By referencing the pointer argv, we get the value 10124, and we can access the address
of the character pointer that argv points to with the contents-of operator. Hence, *argv
will then extract a "pointer to a character" data object, which in this case is the address
1779. If we want to look in memory at what this character pointer points to, we must use
another contents of operator. This will obtain the actual character 'h', since this is the
char that **argv is pointing to.
The above scenario is seemingly idiotic. It is not readily apparent why one would be
interested in using a pointer to a pointer. One example is where a function is supposed to
return three pointers to data objects in memory. Since a function can normally only return
one piece of data at one time, we can use pointers to pass the addresses of pieces of data
to a function in order for that function to fill in the appropriate information. If a function
is supposed to generate three distinct addresses at one time, then we must pass to this
function the addresses of three pointers - we must declare three pointers to pointers!
What if you wanted a function to return the addresses of where the drive letter, the start
of the subdirectory path, and the start of the filename exist inside of a full DOS filename
specification. A possible prototype for this function could be:
void locatePathInfo (char *fullPath, char **drive, char
**path, char **name);
Notice in our ficticious prototype, the first parameter is a normal pointer to a char. We
can pass to this function a string of chars such as:
"C:\\WORK\\PRINTER\\EXAMPLE.OUT"
The next three parameters are declared as pointers to pointers. A segment of C code
which calls this function can be written as:
char *d = NULL, *p = NULL, *n = NULL;
. . .
locatePathInfo ("C:\\WORK\\PRINTER\\EXAMPLE.OUT",
&d, &p, &n);
This function, when finished, will have the character pointer d set to point to the drive
letter C in the full path, the pointer p is set to point to the letter W of the path
specification, and the pointer n is set to point to the letter E of the filename EXAMPLE.
Another situation where pointers to pointers are used is the fact that an array of pointers
can be thought of as a pointer to a pointer. Study the following declaration once again:
char **argv = NULL;

205
 argv is a pointer to a pointer to a character. Because we can use array indexing on pointer
variables, then address information can be located and read out using offset notation. For
example:
argv[3]
will be the 4th char * in memory from the base address of that array. We can use a
pointer to a pointer as an array of pointers in memory. In fact, many programmers will
declare the main function in two standard methods:
int main (int argc, char *argv[], char *env[]);
or
int main (int argc, char **argv, char **env);
It should be apparent that both declarations are equivalent since a pointer to a pointer is
nothing more than an alternate declaration of an array of pointers.

26.3. Pointers and Functions

In C, we have seen that the name of an array is a pointer to the first element in the array.
It turns out that the name of a function is also a pointer: it locates the first byte of code for
that function in memory! In fact, the rounded brackets after a function name are an
operator in C that informs the compiler that we are dealing with function instead of a
variable.
Again, a question will arise: what purpose does a pointer to code serve?
Sometimes a function will need to be told of another function it will call indirectly. This
can be accomplished by passing it the address of another function in memory.
One such common example is the ANSI C standard library function qsort. This function
is prototyped as:
void qsort (void *base, unsigned int num, unsigned int
size, int (*compare)(const void *, const void*));
This is probably one of the most complicated prototypes we've encountered so far. Notice
the use of the void * data type. The qsort function has only one purpose: to sort objects
in memory. These objects must be consecutively stored (an array of objects), thus, by
declaring the pointers as pointers to void, we can allow qsort to work with any data type
in C, as void * is nothing more than a pointer to anything. The second and third
parameters inform qsort of how many objects are consecutively in memory, and the size
in bytes of each data object.
Study the last parameter in the prototype. This is a very strangely declared parameter,
which informs the compiler that the last parameter is a pointer to a function. The
declarations is:
int (*compare)(const void *, const void *)
and declares a parameter variable called compare as a pointer to a function that requires
two const void * parameters, and will return an int data object. This pointer is required
for the qsort function in order for qsort to know how to sort the data in the array! By
206
 indirectly calling this second function through the function pointer that is passed, qsort
can easily sort any data object array because the programmer specifies an address of a
function that will do the comparisons required for sorting!
To pick this declaration apart in more detail, compare is declared as a pointer, because of
the asterisk in front of the name compare. So far we have the start of a generic pointer
declaration. But notice that we enclose the declaration inside of a pair of round brackets.
This is what makes this a declaration of a variable that contains the address of a piece of
code, as distinguished from a data pointer, that will not have the brackets around the
declaration. We follow this with another pair of round brackets with prototyped
parameters. This lets the compiler know what the parameter list of the function is like, so
that type checking can be done at compile time.
Because it is a function, functions must always be prototyped to describe not only input
parameters, but also the output data type. The int declares that compare is a pointer to a
function that will return an int.

26.4. Pointer to a Function Notation

The standard notation for declaring a pointer to a function is:
return-data-type (*name-of-pointer)(parameter
declarations);
As ANSI C is becoming more and more popular (and because C++ demands the use of
prototypes), it is suggested that all functions and pointers to function be declared with the
data types of the parameters.
As a short example of declaring a pointer to a function, a pointer to a function, called foo,
will be used to contain the address of another function that will return a character pointer,
and requires two character pointers as input parameters. The declaration is:
char *(*foo)(char *a, char *b);
Ensure you can follow the above declaration, as this understanding is crucial to your
ability to work with function pointers later in your programming.

26.5. Putting It Together

Another example will quickly clear up confusion about this function, and the use of a
pointer to a pointer.
/* QSORT.C */

#include <stdio.h>
#include <stdlib.h>

int up (const void *x, const void *y);

int down (const void *x, const void *y);
void dumpArray (int *a, int num);

int up (const void *x, const void *y)

207
{
int a = 0;
int b = 0;

a = *(int *)x;
b = *(int *)y;
if (b > a)
{
return -1;
}
else if (b < a)
{
return 1;
}
else
{
return 0;
}
}/* end up */

int down (const void *x, const void *y)

{
int a = 0;
int b = 0;

a = *(int *)x;
b = *(int *)y;
if (b > a)
{
return 1;
}
else if (b < a)
{
return -1;
}
else
{
return 0;
}
}/* end up */

void dumpArray (int *a, int num)

{
int x = 0;

for (x = 0; x < num; x++)

{

208
 printf ("%3d ", a[x]);
if ((x % 10) == 9)
{
printf ("\n");
}
}/* end for */
}/* end dumpArray */

int main (void)

{
int a[20] =
{
19, 99, 5, 10, 7,
33, 22, 57, 50, 48,
3, 77, 66, 13, 18,
9, 10, 59, 38, 28
};

printf ("This program will quick\n");

printf ("sort the array a\n\n");

printf ("original contents:\n\n");

dumpArray (a, 20);

qsort (a, 20, sizeof (a[0]), up);

printf ("\n\nsorted upwards:\n\n");
dumpArray (a, 20);

qsort (a, 20, sizeof (a[0]), down);

printf ("\n\nsorted downwards:\n\n");
dumpArray (a, 20);
return 0;
}/* end main */
The program QSORT.C demonstrates how to use the qsort function and the use of
pointers to functions. Note in the main function how we call qsort. The first call is
passing to qsort the address of the data to sort (the name of the array a), the number of
elements (20), the size of each element (sizeof (a[0]), which is requesting the compiler to
compute the size of element 0 of the array, which will be the size of a single integer), and
finally, the pointer to the function that will perform the comparisons. In this case, the
function is called up, and is defined earlier in this program.
Notice that to pass the address of the function, we simply use the name of the function,
much like how we used the name of the array a to pass a pointer to the first element to
sort. At this point, qsort knows enough information about this array and its data objects
to be able to efficiently sort this data. The qsort is short for "quick sort", and is a very
efficient sorting algorithm.

209
 Check your compiler's help documentation to determine how to write a sorting
comparison function, similar to up() and down() in the QSORT.C program. You will find
that qsort() requires a specific format to the return values for the comparison functions,
which is why up() and down() are designed in such a curious manner.

26.6. Arrays of Function Pointers

There are times where a programmer may want to call a function based upon the current
state of a variable. At the moment, the most efficient manner to do such a thing would be
to use a switch statement, comparing the value of a variable against a number of constant
states, and executing a function as a result of a successful comparison.
Unfortunately, if there are a large number of states to compare against, a large number of
time consuming comparisons may be done. As a state further and further down the list of
cases is compared, the longer it takes to complete the switch statement. In such
situations, especially where program performance is crucial, a programmer may resort to
defining an array of function pointers rather than working with a large switch statement.
This is a programming idea borrowed heavily from software such as operating systems.
In order for an operating system interrupt to be dispatched as efficiently as possible, the
operating system will define a table, or array of function pointers to the operating system
functions, and uses the function number that has been requested as an index into the
table. Once the address of a function is located in the table, it can simply jump to that
address by calling the function obtained as a pointer from the table. This avoids time
consuming switch comparisons and speeds up operating system performance without
much extra coding effort.

26.7. Putting It Together

Another small example program will highlight how a table of function pointer can be
setup and used:
/* PTRFUNC.C */

#include <stdio.h>
#include <stdlib.h>

/*
* function prototypes
*/

int func1 (int);

int func2 (int);
int func3 (int);
int func4 (int);
int func5 (int);
/*
* notice the "bizarre" declaration for the array
* ptrsToFn
*
210
 * its data type is:
*
* int (* )(int)
*
* meaning: a pointer to a function that requires an
* integer parameter and will return an integer
* each element still
* only takes up 1 integer's amount of room in
* bytes of memory!
*/

int (*ptrsToFn[5])(int) =
{
func1,
func2,
func3,
func4,
func5
};

/*
* declaration of a simple pointer to a function
*/

int (*otherPtr)(int) = func3;

/*
* the actual functions themselves
*/

int func1 (int a)

{
printf ("this is function 1\n");
return (1);
}

int func2 (int a)

{
printf ("this is function 2\n");
return (a * 2);
}

int func3 (int a)

{
printf ("this is function 3\n");
return (a * 3);
}

211
int func4 (int a)
{
printf ("this is function 4\n");
return (a * 4);
}

int func5 (int a)

{
printf ("this is function 5\n");
return (a * 5);
}

/*
* let's demo this stuff ...
*/

int main (void)

{
int x = 0;
int y = 0;

printf ("A demo of pointers to functions ...\n");

y = 0;
for (x = 0; x < 5; x++)
{
y = (ptrsToFn[x]) (x + 5);
printf ("in: %d out: %d\n", x + 5, y);
}/* end for */

printf ("test of other pointer ...\n");

x = 55;
y = (*otherPtr)(x);
printf ("x: %d y: %d\n", x, y);
return 0;
}/* end main */
Notice how we declared the array:
int (*ptrsToFn[5])(int)
This declares that the variable ptrsToFn is an array of pointers to functions which will
require a single integer parameter, and return an integer. We know that ptrsToFn is an
array, as the square brackets after the variable name indicate that this variable is an array
of 5 objects, that happen to be pointers to functions that return int data objects, and takes
a single int as a parameter.
The values we initialized this array with are the addresses of five different functions.
Note that we prototyped the functions before their use in the array initialization. This is

212
 needed to satisfy the compiler that what we are initializing this array with are indeed
functions. This program also initializes a simple pointer to a function variable called
otherPtr.
To use these pointer variables and arrays, notice in the main function how we call the
five functions in a loop:
y = (ptrsToFn[x]) (x + 5);
We use the variable x as an index into the array, and set the variable y equal to the return
value of the xth entry in that array. The parameter we pass to each of the five functions is
x + 5. The use of the round brackets around ptrsToFn[x] is not required in all compilers
since most compilers can resolve the fact that this is a reference to an array of function
pointers, and that an array reference will extract a value out of the array, which in this
case is simply an address of code.
The last example shows our normal pointer to a function. Here we use the contents of
operator to extract the address of a function from a pointer, and use this address to
indirectly call the function in question. Again, note the use of round brackets to ensure
our compiler knows that we are working with the contents of a function pointer. (Not all
older compilers are able to cleanly compile without the brackets).

26.8. Other Advanced Uses

Pointers to functions are primarily used in situations like menu handlers. A menu data
object is declared, usually as a structure which contains the menu text, the number of sub
menu items, and a pointer to the function that handles the menu options. These function
pointers are useful to declare the handlers for menu items, thus, an overall menu function
need not worry about what context the menu item is being called under. It will simply use
the supplied pointer and call the function indirectly.
Microsoft Windows, and other graphical operating systems, also use pointers to functions
in interesting manners. Windows, in particular, requires the address of functions so that
Windows can use these functions in a call back scheme. Most operations are done in a
function called WinProc, which is the main Windows procedure handler for your
application. But in a typical Windows program, WinProc is never called directly. In fact,
Windows is informed of the address of a particular application's WinProc, and Windows
will call WinProc itself whenever it sees fit!

26.9. Common Beginner Errors

• Be very careful with the syntax for pointers to functions.

26.10. Summary
This chapter showed some interesting and advanced uses of pointers. Pointers are not
limited to just containing addresses of data objects, but can contains address of other
pointers, as well as address of other functions in our program. The name of a function is a

213
pointer to the first instruction in that function in the code memory of our program, and
function pointer declarations can be built up as simple variables as well as more complex
arrays. In fact, a function pointer can be used in virtually all data objects including
structures.

214
Chapter 27. Now what?
By now, you should have a very good idea of how to write good, well-designed, clear
code in the C language. Some day, you'll be able to write good, well-designed, extremely
unclear code like these winners from the International Obfuscated C Code Contest
(http://www0.us.ioccc.org) ...

2000 Winner for Most Specific Output

Code:
int m = 754974721, N, t[1 << 22], a, *p, i, e = 1 << 22, j, s, b, c, U;
f (d)
{
for (s = 1 << 23; s; s /= 2, d = d * 1LL * d % m)
if (s < N)
for (p = t; p < t + N; p += s)
for (i = s, c = 1; i; i--)
b = *p + p[s], p[s] = (m + *p - p[s]) *
1LL * c % m, *p++ = b % m, c = c * 1LL * d % m;
for (j = 0; i < N - 1;)
{
for (s = N / 2; !((j ^= s) & s); s /= 2);
if (++i < j)
a = t[i], t[i] = t[j], t[j] = a;
}
}

main ()
{
*t = 2;
U = N = 1;
while (e /= 2)
{
N *= 2;
U = U * 1LL * (m + 1) / 2 % m;
f (362);
for (p = t; p < t + N;)
*p++ = (*p * 1LL ** p % m) * U % m;
f (415027540);
for (a = 0, p = t; p < t + N;)
a += (6972593 & e ? 2 : 1) ** p, *p++ = a % 10, a /= 10;
}
while (!*--p);
t[0]--;
while (p >= t)
printf ("%d", *p--);
}

Author: Fabrice Bellard (http://www.enst.fr/~bellard)

215
Author's Comment: This program prints the biggest known prime number (2^6972593-
1) in base 10. It requires a few minutes. It uses a Modular Fast Fourier Transform to
compute this number in a reasonable amount of time (the usual method would take
ages!).

2000 Winner for Best Use Of Flags

Code:
#include <stdio.h>

char
*T="IeJKLMaYQCE]jbZRskc[SldU^V\\X\\|/_<[<:90!\"$434-./2>]s",
K[3][1000],*F,x,A,*M[2],*J,r[4],*g,N,Y,*Q,W,*k,q,D;X(){r [r
[r[3]=M[1-(x&1)][*r=W,1],2]=*Q+2,1]=x+1+Y,*g++=((((x& 7)
-1)>>1)-1)?*r:r[x>>3],(++x<*r)&&X();}E(){A||X(x=0,g =J
),x=7&(*T>>A*3),J[(x[F]-W-x)^A*7]=Q[x&3]^A*(*M)[2 +(
x&1)],g=J+((x[k]-W)^A*7)-A,g[1]=(*M)[*g=M[T+=A ,1
][x&1],x&1],(A^=1)&&(E(),J+=W);}l(){E(--q&&l ()
);}B(){*J&&B((D=*J,Q[2]<D&&D<k[1]&&(*g++=1 ),
!(D-W&&D-9&&D-10&&D-13)&&(!*r&&(*g++=0) ,*
r=1)||64<D&&D<91&&(*r=0,*g++=D-63)||D >=
97&&D<123&&(*r=0,*g++=D-95)||!(D-k[ 3]
)&&(*r=0,*g++=12)||D>k[3]&&D<=k[ 1]
-1&&(*r=0,*g++=D-47),J++));}j( ){
putchar(A);}b(){(j(A=(*K)[D* W+
r[2]*Y+x]),++x<Y)&&b();}t ()
{(j((b(D=q[g],x=0),A=W) ),
++q<(*(r+1)<Y?*(r+1): Y)
)&&t();}R(){(A=(t( q=
0),'\n'),j(),++r [2
]<N)&&R();}O() {(
j((r[2]=0,R( ))
),r[1]-=q) &&
O(g-=-q) ;}
C(){( J=
gets (K
[1]))&&C((B(g=K[2]),*r=!(!*r&&(*g++=0)),(*r)[r]=g-K[2],g=K[2
],r[
1]&&
O())
);;}
main
(){C
((l(
(J=(
A=0)
[K],
A[M]
=(F=
(k=(
M[!A
]=(Q
=T+(

216
 q=(Y
=(W=
32)-
(N=4
))))
+N)+
2)+7
)+7)
),Y=
N<<(
*r=!
-A))
);;}
Author: Glyn Anderson
Author's Comment: This program is a simple filter that translates ASCII text into
semaphore code. It reads from standard input and converts all alphanumeric characters to
their corresponding semaphore codes.
Consecutive whitespace is translated as a single 'break' code. Newlines also cause the
next code to be output on a new line. All other characters are ignored. The codes
themselves are written to standard output as a depiction of a person holding two flags,
using only ASCII characters (semaphore smileys, if you will). For example, given the
input "Hello, world!", the output will be:

<> <> <> <> <> <>

_() ()/ ()/ ()/ _\) () (/_ _\)
[] /^ |^^ /^^ /^^ [] ^^ |^^| ^^ [] [] ^^
<>[ [][ <>][ <>][ ][ [][] ][ ][

<> []
_()_ ()/ |() ()
[] ^^ [] /^^ ^^| |^^|
][ <>][ ][] [][]

1996 Winner for Best Output

Code:
float s=1944,x[5],y[5],z[5],r[5],j,h,a,b,d,e;int i=33,c,l,f=1;int g(){return f=
(f*6478+1)%65346;}m(){x[i]=g()-l;y[i]=(g()-l)/4;r[i]=g()>>4;}main(){char t[1948
]=" `MYmtw%FFlj%Jqig~%`jqig~Etsqnsj3stb",*p=t+3,*k="3tjlq9TX";l=s*20;while(i<s)
p[i++]='\n'+5;for(i=0;i<5;i++)z[i]=(i?z[i-1]:0)+l/3+!m();while(1){for(c=33;c<s;
c++){c+=!((c+1)%81);j=c/s-.5;h=c%81/40.0-1;p[c]=37;for(i=4;i+1;i--)if((b=(a=h*x
[i]+j*y[i]+z[i])*a-(d=1+j*j+h*h)*(-r[i]*r[i]+x[i]*x[i]+y[i]*y[i]+z[i]*z[i]))>0)
{for(e=b;e*e>b*1.01||e*e<b*.99;e-=.5*(e*e-b)/e);p[c]=k[(int)(8*e/d/r[i])];}}for
(i=4;i+1;z[i]-=s/2,i--)z[i]=z[i]<0?l*2+!m():z[i];while(i<s)putchar(t[i++]-5);}}
Author: Thor AAge Eldby

217
Author's Comments: This program shows flying spheres. The program eats CPU on
lesser equipped systems. The program will run until termination. Terminal must be vt100
or better with 80 columns and 24 rows or more.

Disclaimer!
These programs (and any others that you'll find on the website) will probably need a very
specific system in order to compile and run (usually gcc under UNIX).

Another Disclaimer!
Don't even think about submitting code that looks like this as a code sample when you're
interviewing ...

218
Appendix A. Setting Up A Visual Studio Project
• Start up Visual Studio 2013 Professional (the instructions are basically the same for other
versions, the windows just look a tiny bit different).
• Choose New Project (on the Start Page; alternatively, select FILE / NEW / PROJECT in the
menus).

• Under Installed / Templates in the left-side pane, choose Visual C++. In the middle pane,
click on Empty Project. Also enter a Name for the project in the textbox under the main
windows. Make sure that the Location matches with where you want to store the project's
directory. Do not enter a network location.

219
• This will bring up an empty project. We need to create a source file within it. In the Solution
Explorer pane on the left, expand the name of the project, right-click on Source Files, and
choose Add. Then choose New Item.

• In the right-hand pane, select C++ File. Also give the file a name in the Name text box.
Make sure that the name ends in .c. Yes, I'm serious about this.

220
• Type your source code in the source window. Once you're ready to try it out, click on Start
Without Debugging in the Debug menu.

• You might get a dialog box that looks like this:

• If you get the above dialog box, select Yes. Your code will then compile and run. It is also
highly recommended that you check the "Do not show this dialog again" checkbox.

221
Appendix B. Counting Systems
Decimal
The normal method of counting for a human being is based on a numbering system which
is usually referred to as decimal. This counting system, as we all know, is founded upon
the fact that there are 10 distinct digits in the numbering system. The decimal or base 10
counting system counts in the following manner:

0 1 2 3 4 5 6 7 8 9

After 9, we put a zero in the one's place, and put a 1 in the ten's place obtaining the
number 10. We proceed to count:

11 12 13 14 15 16 17 18 19 20

and so on.

Notice that each time we count out 10 distinct digits (from 0 to 9), we add one to the next
placeholder to the left of the digit we have just incremented. This is how we proceed
from 9 to 10, from 19 to 20, from 99 to 100 and so on. The time we add one to the
placeholder to the left is when the current placeholder overflows its prescribed range of
values, as what happens when 9 changes to a 10.

Although these concepts appear absolutely trivial to anyone that can grasp the
significance of pointers in C (a bit of sarcasm, for the humour impaired!), the idea of
incrementing a placeholder until it overflows its range is vitally important to proper
understanding of the next two counting systems, binary and hexadecimal.

Binary
Binary numbering systems are, as the name suggests, base 2. That is, there are only two
different values to count before we overflow the prescribed range of values. The binary
counting system works as follows:

0 1 10 11 100 101 110 111 1000 1001

and so on.

Notice that when we reach a value of 1 in the current placeholder, and we increment this
placeholder, we will overflow the placeholder, and we replace it with zero and add one to
the placeholder to the left. When we normally would have reached the value 2 in decimal,

222
we now get 10 in binary.

The equivalents of binary to decimal can be summarized in the following table:

Binary Decimal
00000000 0
00001000 8
00001001 9
00010000 16
00100000 32
00100001 33
10000000 128
11111111 255

Although by no means complete, the table above should give you an idea of how the
binary counting system works. Normally, a programmer who wants to work with binary
will try to pad the binary number with leading zeros as was done above so that the binary
number reflects the correct number of bits a data object is capable of containing.
Remember that a byte is composed of 8 bits.

The relevance of bits to binary is the fact that a bit is capable of storing two different (or
binary) values: 0 or 1, on or off, true or false, etc. The values a bit can represent are
usually values that are total opposites of each other, so concepts such as high or low,
good or bad and so on are other valid binary values that a bit can hold.

The different data types in C are composed of multiples of bytes, and as stated above, a
byte contains 8 bits. For a byte, the total number of different values that can be
represented by 8 bits are 2 to the power 8, or 256 different values, ranging from 0 to 255
(for unsigned chars) or -128 to 127 (for chars).

For an int, the number of values is 2 to the power 16, or over 65000 different values. A
long is 2 to the power 32, or well over 4 billion! As you can see, the number of different
values is rising exponentially, since we are raising 2 to some exponent based upon the
number of bits in the data object. A short form method of writing 2 to the power (value)
is to write 2 ^ (value).

223
As a side note, bits are numbered starting from bit zero which is the first bit on the right.
For a byte, bit 7 is the first bit on the left, for an integer, bit 15 is the first bit on the left.
Oftentimes, programmers will refer to the first bit on the left as the high bit.

To convert any decimal value into binary, see what powers of 2 can be added together to
make the decimal value. For instance:

57 is equivalent to: 2 ^ 5 (32)

+ 2 ^ 4 (16)
+ 2 ^ 3 (8)
+ 2 ^ 0 (1)

giving a bit representation of:

00111001

A concept that lends itself quite nicely to binary are the bitwise operators: ! | and ^, as
well as ~. Since these operators do things to the bits within a data object, understanding
how bits are arranged in a data object will make the above operators obvious in their
operation.

For the bitwise complement (~), we simply flip all 1 bits to 0 bits, and all 0 bits to 1 bits.
Thus:

char x, y;
. . .
x = 63; /* 00111111 in binary */
y = ~ x;
will cause variable y to contain 11000000 in binary, which is equivalent to 192. Notice
that if we add 63 and 192 we obtain 255. A bitwise complement can also be computed by
simply subtracting the value in question from the largest value the data object can hold.
For a byte, ~63 is simply 255 - 63, or 192. For ~65535, we get 0, and so on. Simple!

One trick programmers like to use is to define the following constants:

#define FALSE 0
#define TRUE ~FALSE
Here, FALSE is given the value of zero, and TRUE is given the value of ~0, which is
65535, (or negative 1 if you are using signed arithmetic). In C, false is always zero, and
true is anything but zero, so the definitions above are valid.

224
The & operator (when not prepended in front of a variable) is a bitwise AND. This is a
binary operator, and needs two operands to work with. The bitwise AND will take the
bits in both operands, and AND them together. Two 1's make a 1, otherwise we get a
zero. Thus:

00101010
& 11100110 makes
00100010
The binary AND operator is a wonderful method of testing for the
presence of a certain bit in a value. If we wanted to check to
see if the second bit from the right is on inside a value,
perform the following:

if (value & 2) /* 2 is 00000010 */

statement;
If the value has the second bit from the right turned on, the bitwise AND will leave that
bit set in the result, and because the result is non-zero, statement will execute!

Bitwise OR (|) is similar except that if both bits are 0's, we generate a 0, otherwise
generate a 1. Bitwise ORs are excellent for setting particular bits in a value. For instance:

value |= 4; /* 4 is 00000100 */
will turn on the 3rd bit from the right in the variable value.

The bitwise XOR operator (^) is used to eXclusive OR two values together. XOR is the
same as OR, except in the case where two bits are 1's. This case will generate a 0 using
XOR when it generated a 1 with an OR.

Another closely coupled set of operators are the bitwise shifting operators >> and <<.
These operators will move bits within a data object back and forth by the specified
amount of bits. For example:

X = 1 << 4;
will set x equal to (binary) 00000001 shifted to the left by 4 bits, which will result in
00010000, or 16. Shifting to the right is similar in nature, except the bits move to the
right. For example, 33 (00100001) >> 2 is: 00001000 (the value 8). The shift of 1 by 4
bits is like setting x to 1, and multiplying by 2 four times. (1 * 2 = 2, 2 * 2 = 4, 4 * 2 = 8,
8 * 2 = 16), and a shift left of 2 is an integer divide of the value by 2 twice (33 / 2 = 16,
16 / 2 = 8). In fact, binary shifts are very fast multiplies and divides by the value 2. For
each shift done, we are either multiplying by 2 (shift left) or integer dividing by 2 (shift
right). Most modern compilers will recognize the fact that you may be doing powers of 2

225
for divides and multiplies, and will compile these operations as shifts for you for
optimum speed from your compiled code!

Hexadecimal
The most significant counting system of interest to programmers is the hexadecimal
system, or hex for short. This system uses yet another base to count by, this time, base
16.

This is a particularly odd looking counting system since we need 16 different symbols for
each placeholder. For decimal, we only needed 10, and the ten symbols are the numbers 0
through 9. For hex, we use the same 10 number symbols, and add the following symbols:

A B C D E F

When we reach F, and we add one more, we find we have exhausted the range of the first
placeholder, thus we must increment the next placeholder to the left. The net result is we
get, none other than, 10!

The first few counts in hex are:

0 1 2 3 4 5 6 7
8 9 A B C D E F
10 11 12 13 14 15 16 17
18 19 1A 1B 1C 1D 1E 1F
20 21 22 23 24 25 26 27

and so on.

The interesting thing about hex is that we can represent all values that a byte can hold
using only 2 placeholders in hex. It turns out that the value 255 is FF in hex.

In C, to differentiate between a decimal number and a hex number, we prepend all hex
constants with the characters:

0x
which informs the compiler that a hex number is about to be defined. FF in hex will be
written as 0xFF in a C program. C compilers will accept either 0xFF or 0xff for the same
value of 255 (decimal).

226
Some notable numbers in hex are:

0x00 - 0 0x10 - 16 0x20 - 32

0x40 - 64 0x80 - 128 0xFF - 255

Note that 0xFF plus 1 is 0x100, as is hopefully expected.

To convert a base 10 number into hex, take the base 10 decimal number, and divide by
16. We keep dividing the result by 16 until we get a quotient less than 16. The quotient
that we obtain (the whole number portion of the division result) is the first hex digit.
Take the first hex digit, and multiply it with 16 by the number of divides we did, and
subtract that result from the original number. The remaining value is taken through the
above process to extract the second hex digit, and the same for the third, and so on. For
example, the value 1300 in decimal converts into:

1300/16=81.25/16 = 5
1300-5*16*16=20/16 = 1
20 - 1*16 = 4

Therefore, 1300 in decimal is 0x514 in hex. Or even easier, use your Windows calculator
application, and switch into scientific mode, which has binary, decimal and hexadecimal
conversions available.

A question must be burning in everyone's heads by now: "why do we need to know this
weird stuff?" The answer is that virtually all computers, deep down inside of their
microprocessors and memory, all think naturally in base 16 and binary. Most hardware
for a computer is communicated to using hex and binary values. Even a little knowledge
about hex and binary will enable you to appreciate what some of the strange values your
PC's hardware manuals mean!

Octal
There is one other counting system that may be of interest: the octal system, which is
base 8. By this point, it should be apparent that base 8 is counting like the following:

0 1 2 3 4 5 6 7
10 11 12 13 14 15 16 17
20 21 22 . . .

and so on.

227
Appendix C. Recap of Course Requirements Mentioned In
This Document
Header Comments
You must have a header comment in each source code file that you create. It must
contain the information mentioned above. The format should be similar to that shown in
the example above.
Header Comment Description
The description contained within the header comment should accurately describe the
intent of the contents of the file. This is most important in multi-file projects that contain
many, many .c and .h files. Thus, a description like “This is assignment #2.” is bad while
a description like “This program takes Fahrenheit temperatures from the user and
displays them as Kelvin and Celsius temperatures. The program loops until the user
enters an invalid temperature.” is good.
main() Return Value
The main() function must return an int value. If you don't care what the value is (and you
typically don't), you can feel free to use return 0 before the closing brace in main().
Don’t return void! Don't leave out the return value!
Style Consistency
You must use correct and consistent style in this course!
Distinguishing Between Words In Identifiers
Use camel case to distinguish between words in names, unless otherwise explicitly stated.
Naming
Start your names with a lowercase letter, unless otherwise explicitly stated.
Initialization of Variables
Always initialize your variables when you declare them, unless notified otherwise.
Failing to do so will cause you to lose marks and get quite upset about it.
Global Variables Banned
Do not use global variables unless explicitly permitted by the instructor. Note that this
does not restrict the use of global constants (variables whose declarations are preceded by
the keyword const; this is to be discussed later).
There is a LARGE mark penalty that is associated with using global variables when not
allowed to.
Indentation After Decisions
You must indent the body of the if statement to show that a decision has to be made and
the indented code may or may not be executed. This also applies to while, do-while, and
228
for. The principle is to indent the body after a decision. This also implies that the
decision and body cannot be on the same line.
Braces on Single Statement Bodies
You are required to put braces on single-statement bodies in this course.
Recursive Calls of main()
It is typically a very bad idea to recursively call main(). You will lose marks if you do
so.
Function Comments
You must have function comments for all functions you create except main(). This
comment must be immediately before the function definition.
Prototyping
You must prototype all functions that you write except main() (which has its own implicit
prototype). You must also provide void in the parameter list if there are no parameters.
Failing to do so will lose you marks.
Naming Standards
Take a look at the SET Coding Standards to see what is required.
goto
Never use a goto statement. If you do, you will lose TONS of marks.
Always Initialize Arrays
Always initialize your array variables, even if it’s just with a single 0 in the very first
element.
Pointer Initialization
Always initialize pointer variables. If you don’t know what a valid value would be for
the pointer when you declare it, initialize it to NULL.
scanf()
Do not use scanf(). Use sscanf() (or one of the conversion functions like atoi(), atol(), or
atof()) instead, in conjunction with gets() or fgets(). Failing to do so will lose you marks!
Check NULL
Whenever a function can return NULL to indicate an error, check the return value.
Freeing Memory
Whenever allocating memory, always free up memory you do not require, after you have
completed your memory operations. Do not assume that the operating system will do this
for you.

229
Closing files
Most operating systems will automatically close opened files when the application
terminates. This is not the most appropriate technique to use, for the same reasons as for
dynamic memory. To ensure full portability of software across all operating systems, you
must always manually close your files when no longer required, to avoid potential
operating system deficiencies.

230