C Basics

Wednesday, 11 April 2018 8:42 PM

All information for this course will come from Andrew Taylor's and Andrew Bennett's slides/notes
Variables
Variables are used to store a value. The value a variable holds may change at any time. At any point in time a
variable stores one value, (except for quantum computers).

C variables have a type. For this course the main variables we use are:
• int - for integer values
• double - for decimal numbers
• char - for characters

Integer Representation
Typically 4 bytes are used to store an int variable.
4 bytes -> 32 bits -> 232 possible values (bit patterns)
This means only 232 integers can be represented.
These integers are: -231 to 231 - 1 (-2,147,483,648 to +2,147,483,647)
These limits are asymmetric because zero needs a pattern (the pattern being all zeros).

Integer Overflow/Underflow
Storing a variable in an int outside the range that it can be represented is illegal. This can result in unexpected
behaviour from most C implementations, or it may cause programs to halt, or not terminate. This can increase
security holes.

Bits used for int can be different on other platforms. For example, C on a tiny embedded CPU in a washing machine
may use 16 bits. For now we assume int uses 32 bits.

Real Representation
Commonly 8 bytes are used to store a double variable.
8 bytes -> 64 bits -> 264 possible value (bit patterns)
64 bits give huge number of patterns but infinite number of reals.

More on Variables int number = 0;

How to use variables: printf("%d\n", number);
1. Declare
2. Initialise Note: steps 1 and 2 can usually be done in the same line.
3. Use
If you don’t initialise a variable, it will take the last value that was stored in that variable, and this can lead to bad
stuff :(

Variable names:
Variable names can be made up of letters, digits and underscores.
Here are some unwritten rules about variable names:
• Use a lowercase letter to start your variable name
• Beware! variable names are case sensitive
• Beware! certain words can't be used as variable names; e.g. if, while, return, int, double. This is because
these keywords have special meanings in C programs

Printing and reading variables:

Use %d to read and print an int value
Use %lf to read and print a double value
Note: if you want specific decimal places for double write it like this: %.nlf, where n is how many decimal places
The format string is of the form
% [flags] [field_width] [.precision] [length_modifier] conversion_character
where components in brackets [] are optional. The minimum is therefore a % and a conversion character (e.g. %i).

COMP1511 Page 1
where components in brackets [] are optional. The minimum is therefore a % and a conversion character (e.g. %i).
Flags Flag Meaning
- The output is left justified in its field, not right justified (the default).
+ Signed numbers will always be printed with a leading sign (+ or -).
space Positive numbers are preceded by a space (negative numbers by a - sign).
0 For numeric conversions, pad with leading zeros to the field width.
# An alternative output form. For o, the first digit will be '0'. For x or X, "0x" or "0X" will be
prefixed to a non-zero result. For e, E, f, F, g and G, the output will always have a decimal
point; for g and G, trailing zeros will not be removed.

Field width Converted argument will be printed in a field at least this wide, and wider if necessary.
If the converted argument has fewer characters than the field width, it will be padded on the left (or
right, if left adjustment has been requested) to make up the field width. The padding character is
normally ' ' (space), but is '0' if the zero padding flag (0) is present.
If the field width is specified as *, the value is computed from the next argument, which must be
an int.
Precision A dot '.' separates the field width from the precision.
If the precision is specified as *, the value is computed from the next argument, which must be an int.
Conversion Meaning
s The maximum number of characters to be printed from the string.
e, E, f The number of digits to be printed after the decimal point.
g, G The number of significant digits.
d, i, o, u, x,X The minimum number of digits to be printed. Leading zeros will be added to make
up the field width.

Length Character Meaning

modifier
h The value is to be displayed as a short or unsigned short.
l For d, i, o, u, x or X conversions: the argument is a long, not an int.
L For e, f, g or G conversions: the argument is a long double.

Conversion Characte Meaning

character r
d, i Display an int in signed decimal notation.
o Display an int in unsigned octal notation (without a leading 0).
u Display an int in unsigned decimal notation.
x, X Display an int in unsigned hexadecimal notation (without a leading 0x or 0X). x gives
lower case output, X upper case.
c Display a single char (after conversion to unsigned int).
e, E Display a double or float (after conversion to double) in scientific notation. e gives lower
case output, E upper case.
f Display a double or float (after conversion to double) in decimal notation.
g, G g is either e or f, chosen automatically depending on the size of the value and the
precision specified. G is similar, but is either E or f.
n Nothing is displayed. The corresponding argument must be a pointer to an int variable.
The number of characters converted so far is assigned to this variable.
s Display a string. The argument is a pointer to char. Characters are displayed until a '\0' is
encountered, or until the number of characters indicated by the precision have been
displayed. (The terminating '\0' is not output.)
p Display a pointer (to any type). The representation is implementation dependent.

COMP1511 Page 2
 p Display a pointer (to any type). The representation is implementation dependent.
% Display the % character.
Above information from <http://personal.ee.surrey.ac.uk/Personal/R.Bowden/C/printf.html>

Giving Constants Names

It can be useful to give constants names. In general any magic number that holds meaning should be given constant
names. This makes your program readable. It can also make your program easier to update especially if the
constant appears in many places.

One method of doing this is the #define statement. #define statements go at the top of your program after the
#include statements and #define names should always be in capital letters and underscores.
#define NAME_OF_CONSTANT 100

Mathematics in C
C supports the usual maths operations: + - * /. BODMAS follows.

Beware! Division in C is not what you'd expect

When dividing with double, C division is what you'd expect: 2.6/2 = 1.3
When dividing with int, C gets a bit weird: 2.6/2 = 1. The fraction part is discarded (not rounded).

C also has the % (modulo) operator, but it is for integers only.

Mathematical functions are not part of the standard library because tiny CPUs may not support them.
The library math.h contains mathematical functions such as sqrt(), sin(), cos(), tan() (These take
double as arguments and return double).

COMP1511 Page 3
If Statements
Wednesday, 11 April 2018 8:43 PM

Many problems require executing statements only in some circumstances. This is sometimes called
control flow, branching or conditional execution. We do this in C using if statements.

A general if statement appears like this:

if (expression1) {
statement1;
…
} else if (expression2) {
statement2;
…
} else {
statement3;
}

statement1, … are executed if expression1 is non-zero.

statement1, … are not executed in expression1 is zero.
Instead it will check if expression2 is non-zero, and the same thing happens as expression1.
If none of the expressions are non-zero, else will execute statement3.

Also, there is no "boolean" type in C.

0 is regarded as FALSE
Anything non-zero is regarded at TRUE

Relational Operators
C has the usual operators to compare numbers:
Operator Meaning
> greater than
>= greater than or equal to
< less than
<= less than or equal to
!= not equal to
== Equal to
Be careful when comparing doubles for equality == or !=. Recall that doubles are approximations.

Many languages have a separate type for true & false. C just uses 0 for false and other numbers for
true

Relational operators return:

The int 0 for false
The int 1 for true

Logical Operators
C also has logical operators:
Operator Meaning
&& and operator - true if both operands are true
|| or operator - true if either operand is true
! not operator - true if and only if its operand is false

COMP1511 Page 4
Functions
Wednesday, 11 April 2018 8:42 PM

Functions allow you to:

• separate out "encapsulate" a piece of code serving a single purpose
• test and verify a piece of code
• reuse the code
• shorten code resulting in easier modification and debugging
Some functions we already use are:
printf() and scanf()

Structure of a Function The return Statement

1. Return type When a return statement is executed, the function
2. Function name terminates. The return expression will be evaluated,
3. Parameters (inside brackets, comma separated) and if necessary, converted to the type expected by
4. Variables (which are only accessible within the function) the calling function.
5. Return statement
All local variables and parameters will be thrown
Example: away when the function terminates. The calling
int add_numbers(int num1, int num2) { // 1, 2, 3 function is free to use the return value, or to ignore it.
int sum; // 4
sum = num1 + num2; Functions can be declared as void, which means that
return sum; // 5 nothing is returned. A return without a value
} statement can still be used to terminate such a
function.

Calling a function
When calling a function, you type the function name and the variable you will pass in to the function.
E.g. x = cube(number);
- cube is the function name
- number is the variable we are passing into the function.

Function Properties
• Functions have a type. This is the type of value they return.
• Type void is for functions that return no value. void is also used to indicate that a function has no
parameters.
• Functions cannot return arrays
• Functions that have their own variables created when the function is called and these variables are destroyed
when the function returns.
• A function's variables are not accessible outside the function
• return statements stop the execution of a function
• return statements specify the value to return unless the function is of type void
• A run-time error occurs if the end of a non-void function is reached without a return

Function with No Return Value Function Parameters

Some functions do not compute a value. Functions take 0 or more parameters.
They are useful for "side-effects" such as output. Parameters are variables created each time the
function is called and it is destroyed when the
void print_sign(int b) { function returns. C functions are call-by-value
if (b < 0) { (but beware arrays). The parameters are
printf("negative"); initialised with the value supplied by the caller.
} else if (b == 0) { The value is essentially copied to the
printf("zero"); parameter. A parameter variable changed in
} else { the function has no effect outside of the
printf("positive"); function.
}
}

Function Prototypes

COMP1511 Page 5
Function Prototypes
Function prototypes allow function to be called before it is defined. It species key information about the function:
• function return type
• function name
• number and type of function parameters
It allows a top-down order of functions in the file, which makes it more readable. It also allows us to have function
definition in a separate file. This is important since it is crucial to share code and important for larger programs

Library Function
Over 700 functions are defined in the C standard library.

The C compiler needs to see a prototype for these functions before you use them. You do this indirectly with
#include line. For example stdio.h contains prototypes for printf and scanf.

COMP1511 Page 6
While Statements
Wednesday, 11 April 2018 8:42 PM

if statements only allow us to execute or not execute code. This means that they execute code either
once or never. Meanwhile, while statements allow us to execute code once or more times.

Like if, while statements have a controlling expression but while statements execute their body of
statements until the controlling expression is false.

The general construction of a while statement looks like this:

while (expression) {
stmt1;
stmt2;
…
stmtn;
}

Loop Counter
Often we use a loop counter variable to count loop repetitions. This allows us to have a while loop
execute n times.

An example of using the loop counter:

// read an integer n
// print n asterisks
int loop_counter, n;

printf("How many asterisks? ");

scanf("%d", &n);

loop_counter = 0;
while (loop_counter < n) {
printf("*");
loop_counter++;
// this is the same as loop_counter = loop_counter + 1;
}
printf("\n");

Termination
We can control the termination (stopping) of while loops in many ways. However, it is very easy to
write a while loop which does not terminate. Often a sentinel variable is used to stop a while loop
when a condition occurs in the body of the loop.

Here is a general outline for using the sentinel variable pattern:

stop_loop = 0;
while (stop_loop != 1) {
//
// statements the loop needs to perform
//
if (.......) {
stop_loop = 1;
}
//
// perhaps more statements
//
}

Nested While Loops

We often need to nest while loops. When we do this we need separate counter variables for each

COMP1511 Page 7
We often need to nest while loops. When we do this we need separate counter variables for each
nested loop.

// print a square of 10x10 asterisks

int i, j;
i = 0;
while (i < 10) {

j = 0;
while (j < 10) {
printf("* ");
j = j + 1;
}

printf("\n");
i = i + 1;
}

COMP1511 Page 8
Arrays
Wednesday, 11 April 2018 8:42 PM

A C array is a collection of variables called array elements. All array elements must be of the same type. Array
elements do not have names, instead they are accessed by a number called the array index. A valid array index for
an array with n elements are:
0, 1, 2, … , n - 1
Arrays must be initialised, or else weird stuff will happen, which you don’t want.

You also can assign scanf or printf whole arrays, instead you can assign scanf/printf array elements.

// Declare and initialise the array

int array[5] = {0};
Diagrammatically the array would look like this:

// You can also initialise the array like this

int array[5] = {5, 2, 1, 6, 9};
In a diagram it would look like this:

Reading and Printing Arrays

When reading arrays, you must read each element individually. This is easiest to do with a while loop.
#define ARRAY_SIZE 20
i = 0;
while (i < SIZE) {
scanf("%d", &array[i]);
i = i + 1;
}

Likewise, if you are printing arrays, you must print each element individually
i = 0;
while (i < ARRAY_SIZE) {
printf("%d\n", array[i]);
i = i + 1;
}

Arrays of Arrays/ Matrixes

C supports arrays of arrays. This is useful for multi-dimensional data.

A two-dimensional array (a matrix) would be useful for storing both rows and columns of data.

COMP1511 Page 9
A two-dimensional array (a matrix) would be useful for storing both rows and columns of data.
Declaring and initialising an array would look like this:
int matrix[3][3] = { {1, 2, 3},
{4, 5, 6},
{7, 8, 9} };

Reading and Printing Two-dimensional Arrays

When reading two-dimensional arrays, we need to The same applies for printing arrays.
use a nested while loop, which uses two counters;
One for the row and the other for the column.
#define SIZE 42
... ...
int matrix[SIZE][SIZE]; while (i < SIZE) {
int i, j; j = 0;
while (j < SIZE) {
i = 0 print("%d", &matrix[i][j]);
while (i < SIZE) { j = j + 1;
j = 0; }
while (j < SIZE) { printf("\n");
scanf("%d", &matrix[i][j]); i = i + 1;
j = j + 1; }
}
i = i + 1;
}

COMP1511 Page 10
Strings
Wednesday, 11 April 2018 8:42 PM

The char type ASCII Encoding

The C type char stores small integers. It is ASCII stands for American Standard Code for Information Exchange.
almost always 8 bits. char is guaranteed It specifies the mapping of 128 characters to the integers 0…127.
to be able to represent integers 0..+127 The characters encoded include:
and it is most likely to store ASCII • Upper and lower case English letters: A-Z and a-z
character codes. • Digits: 0-9
• Common punctuation symbols
To print chars with printf we use %c. • Special non-printing characters: e.g. newline and space
But don’t worry you don't have to memorise ASCII. Just writing
Do not use char for individual variables. single quotes give you the ASCII code for a character
Use them only for arrays and characters. printf("%d", 'a'); // prints 97
Even if a numeric variable is only used for printf("%d", 'A'); // prints 65
values 0..9, use the type int for the printf("%d", '0'); // prints 48
variable. printf("%d", ' ' + '\n'); // prints 42 (32 + 10)
So don’t put ASCII in your programs - just use single quotes.
Reading a character - getchar
C provides library functions for reading and writing characters.
getchar reads a byte from standard input and returns and int. getchar returns a special value, EOF (usually -1) if
it cannot read a byte. EOF stands for End Of File. Otherwise getchar returns and integer (0..255) inclusive. If
standard input is a terminal or a text file, this is likely to be an ASCII code. Beware, input is often buffered before
an entire line can be read.

End of Input
Input functions such as scanf or getchar can fail because no input is available. e.g. if input is coming from a file
and the end of the file is reached. On UNIX-like systems (Linux/OSX) typing ctrl + D signals to the operating
systems no more input from the terminal. Windows has no equivalent to this, although some windows programs
interpret ctrl + Z similarly.

getchar returns a special value to indicate there is no available. This non-ASCII value is #defined as EOF in
stdio.h. On most systems EOF == -1. There is no end-of-file character on modern operating systems.

The programming pattern for reading characters to the end of input would look like this:
Programming pattern for reading characters to the end of input:
int ch;

ch = getchar();
while (ch != EOF) {
printf("'%c' read, ASCII code is %d\n", ch, ch);
ch = getchar();
}

Strings
A string in computer science is a sequence of characters. In C, strings are an arrays of char containing ASCII
codes. These arrays of char have an extra element containing a 0. The extra 0 can also be written '\0' and may
be called a NULL character or NULL-terminator. This is convenient because programs don't have to track the
length of the string.

Because working with strings is so common, C provides some convenient syntax.

Instead of writing:
char hello[] = {'h', 'e', 'l', 'l', 'o', '\0'};
You can write
char hello[] = "hello";

Note: hello will have 6 elements; 5 for the individual characters, 1 for the NULL terminator '\0'.

The C library includes some useful functions which operate on characters. Here are some below:
#include <ctype.h>

int toupper(int c); // convert c to upper case

int tolower(int c); // convert c to lower case
int isalpha(int c); // test if c is a letter
int isdigit(int c); // test if c is a digit
int islower(int c); // test if c is lower case letter
int isupper(int c); // test if c is upper case letter

COMP1511 Page 11
Reading a string/line - fgets string.h functions
fgets(array, array size, stream) reads a line of text Here are some string functions that may be useful:
1. array - is a char type array. This is where the line will be #include <string.h>
stored. // string length (not including '\0')
2. array size - is the size of the array. This size is how big the line int strlen(char *s);
can be.
3. stream - is where the line will be read from e.g. stdin. // string copy
char *strcpy(char *dest, char *src);
fgets cannot not store more characters than the array size, because char *strncpy(char *dest, char *src, int n);
it will not be big enough .
fgets always stores a terminating character ,'\0', in the array. // string concatenation/append
fgets stores a newline character, `\n', in the array, if it reads an char *strcat(char *dest, char *src);
entire line. We often need to overwrite this newline character: char *strncat(char *dest, char *src, int n);
int i = strlen(lin);
// string compare
if (i > 0 && line[i - 1] == "\n) {
int strcmp(char *s1, char *s2);
line[i - 1] = '\0';
int strncmp(char *s1, char *s2, int n);
}
int strcasecmp(char *s1, char *s2);
NEVER use the similar C function gets, which can overflow the array int strncasecmp(char *s1, char *s2, int n);
and cause major source of security exploits.
// character search
The programming pattern for using fgets looks like this: char *strchr(char *s, int c);
#define MAX_LINE_LENGTH 1024 char *strrchr(char *s, int c);
...
char line[MAX_LINE_LENGTH];
printf("Enter a line: ");
// fgets returns NULL if it can't read any
characters
if (fgets(line, MAX_LINE_LENGTH, stdin) != NULL {
fputs(line, stdout);
// or
printf("%s" ,line); // same as fputs
}

Command-line Arguments Converting Command-line Arguments

Command-line arguments are 0 more strings specified when the stdlib.h defines some useful functions for converting strings.
program is run.
If you run this command in a terminal: atoi (for ASCII to integer) converts strings to int
$ dcc -o count count.c atof (for ASCII to float) converts strings to double
dcc will be given 3 command-line arguments: "count.c" "-o" "count"
Here is a demonstration of how to use atoi:
Our main function will need different parameters if we want to access #include <stdio.h>
command-line arguments #include <stdlib.h>
int main(int argc, char *argv[]) { … } int main(int argc, char *argv[]) {
int i, sum = 0;
argc stores the number of command-line arguments + 1. i = 1;
while (i < argc) {
If argc == 1 , there are no command-line arguments.
sum = sum + atoi(argv[i]);
argv stores the program name and the command-line arguments.
i = i + 1;
argv[0] always contains the program name
}
argv[1] argv[2] … will contain the command-line arguments if
printf("sum of command-line arguments=%d\n", sum);
they are supplied }
If we want to print out the command-line arguments we would do
something like this:
#include <stdio.h>
int main(int argc, char *argv[]) {
int i = 1;
printf("My name is %s\n", argv[i]);
while (i < argc) {
printf("Argument %d is: %s\n", i,
argv[i]);
i = i + 1;
}
}

Week 6 Tutorial
char is a data type that stores single characters
%c is used to read and print characters when using printf and scanf

When you initialise an array of characters use double quotes ""

COMP1511 Page 12
When you initialise an array of characters use double quotes ""
e.g. char string[8] = "";
But when assigning specific ASCII values use single quotes
e.g. char character = 'A';
char string[9] = { 'C', 'O', 'M', 'P', '1', '5', '1', '1', '\0'};
printf("%s\n", string);
Remember: a string is an array of characters with a NULL terminator at the end.

Instead of individually assigning each ASCII character into the array you can just do this:
char string[9] = "COMP1511";
printf("%s\n", string);

Command Line Arguments

If we use command line arguments in our program the parameters in our main function should change to
int main(int argc, char *argv[]) {
}
argc is an integer telling how many commands there are.
argv is an array of strings storing what the commands are.

The program name itself is an argument.

./caesar 10
argc = 2
argv[0] = "./caesar"
argv[1] = "10"

Usage: ./caesar <number>

atoi converts ASCII to integer

int num = atoi(argv[1]);

if (argc != 2) {
perror("Usage: %s <number\n", argv[0]);
return 1;
}

putchar() and getchar()

getchar() receives characters
putchar() prints out characters

int number = getchar();

while (number != EOF) {
putchar(number);
number = getchar();
}

When we type in characters in terminal it doesn’t immediately print out the characters. This is because the
characters are held in a buffer waiting to be sent to the program once you press enter.

Cracking the caesar cipher: BRUTE FORCE!!!! There are only 25 shifts
Check against English words

COMP1511 Page 13
I/O, Reading and Writing Files
Wednesday, 11 April 2018 2:04 PM

Introducing I/O (Input/Output) Standard Streams

The Unix process environment can be represented like this: A stream is a sequence of bytes.
stdio.h defines three streams:
• stdin - standard input stream; functions like
scanf and getchar read from stdin
• stdout - standard output stream; functions like
printf and putchar write to stdout
• stderr - standard error stream; this stream by
convention is used for error messages; like
stdout it goes to the terminal but it is
semantically different to stdout; it can be
redirected separately to stdout

Can access with fprintf. This is just like printf but works with files

I/O Direction Unix Pipes

If a program runs from terminal: Unix shell allows you to connect stdout of one program
stdin, stdout, stderr are connected to the terminal to stdin of another program.
Unix shell allows you to re-direct stdin, stdout and stderr
Run a.out with stdout going to stdin of wc
Run a.out with stdin coming from file data.txt ./a.out | wc -l
This will import the content from data.txt to a.out
./a.out < data.txt wc counts chars, words and lines in its stdin
Some other useful Unix programs (a.k.a filters) designed
Run a.out with stdout going to (overwriting) file output.txt to be run like wc include grep, cut, sort, uniq.
This is export the content from a.out to output.txt This is covered in more detail in COMP2041.
./a.out > output.txt
Using stderr for error messages
Run a.out with stdout appended to file output.txt fprintf allows you to specify which stream to print out to.
This will add content from a.out to output.txt For example:
./a.out >> output.txt
fprintf(stderr, "error: cannot open %s\n, f);
Run a.out with stderr going to file error.txt printf is actually just calling fprintf and specifying stdout
./a.out 2> error.txt fprintf(stdout, …);

It is best for error messages to be written to stderr, so users see them

even if stdout is directed to a file

Accessing Files in C Opening a File

FILE *f = fopen(char *filename, char *mode); FILE *output_file = fopen("filename.txt", "w");
This creates a stream to read from a file, or a stream to write to a file.
It returns NULL if it fails to open anything - FILE * is the type of the variable. It is a pointer to the file
- fopen - opens a file
- parameter 1 - the name of the file to be opened
int fclose(FILE *f);
- parameter 2 - the mode in which to open the file
This finishes operations on a file. - return value - a pointer to the file which has been
fclose is automatically called on a program exit, although some output opened. This pointer is then used to reference the opened
may be cached until fclose is called. file for operations such as reading, writing, and closing the
file.
int fgetc(FILE *f); The return value will be NULL if the file cannot be opened or
This reads a single character from a stream. It returns EOF if no if you try to open a file you don't have permission to access
character is available.
fopen mode parameters

COMP1511 Page 14
int fgetc(FILE *f); The return value will be NULL if the file cannot be opened or
This reads a single character from a stream. It returns EOF if no if you try to open a file you don't have permission to access
character is available.
fopen mode parameters
int fputc(int, FILE *f); Mode Description
This writes a single character to a stream. "r" opens an existing file for reading purposes
"w" opens a text file for writing
int fscanf(FILE *f, …); if it doesn’t exist a new file is created
This performs scanf from a stream and returns the number of values start writing from the top of the file (i.e. it will overwrite
read. the original file)
"a" opens a text file for writing in appending mode
int fprintf(FILE *f, …); if doesn’t exist a new file is created
starts writing at the end of existing file content
This prints to a specified stream.
"r+" opens a file for both reading and writing
Example: "w+" w+ truncates the file to zero if it exists
"a+" a+ starts reading from the start of the file and writing at
// opens a file called output.txt in "writing" mode
FILE* output_file = fopen("output.txt", "w"); the end of the existing file contents

// prints "Hello!" to the file

fprintf(output_file, "Hello!\n"); Reading from file
char line[MAX_LINE_LENGTH];
// print to stdout (These are the same thing);
if (fgets(line, MAX_LINE_LENGTH, fp) != NULL) {
printf("Hello!");
printf("read line\n");
fprintf(stdout, "Hello!");
} else {
printf("Could not read a line\n");
}
OR
Writing to a file
int c;
fputs("the text I want to put in the file\n", fp); c = fgetc(fp);
OR if (c != EOF) {
printf("read a '%d'\n", c);
fprintf(fp, "the text I want to put in the file\n");
} else {
OR printf("Could not read a character\n");
int c = "!"; }
fputc(c, fp);

COMP1511 Page 15
Memory and Pointers
Wednesday, 11 April 2018 3:14 PM

Decimal Representation Binary Representation

We can interpret the decimal number 4705 as: In a similar way, we can interpret the binary
4 × 103 + 7 × 102 + 0 × 101 + 5 × 100 number 1011 as:
The base or radix is 10, and the digits are 0 - 9 1 × 23 + 0 × 22 + 1 × 21 + 1 × 20
The place values: The base or radix is 2, and the digits are 0 and 1
… 1000 100 10 1 The place values:

… 103 102 101 100 … 8 4 2 0

We can write the number as 470510. The … 23 22 21 20

subscript is used to denote the base We can write the number as 1011 10 (= 1110)

Hexadecimal Representation
We can interpret the hexadecimal number 3AF1 as:
3 × 163 + 10 × 162 + 15 × 161 + 1 × 100
The base or radix is 16, and the digits are:
0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, F
The place values:
… 4096 256 16 1
… 163 162 161 160
We can write the number as 3AF116 (= 1508910)

Binary to Hexadecimal Hexadecimal to Binary

Convert 10111110001010012 to Hex: Convert each hex digit into the equivalent of 4-bit
1011 1110 0010 10012 binary representation.
E.g. Convert AD516 to Hex:
B E 2 916
A D 5
Convert 101111010111002 to Hex:
1010 1101 010116
0010 1111 0101 11002
2 F 5 C16 To print hexadecimal values in C use the place value
Note: The two zeros in bold have been added holder %x

Memory Organisation
Memory is effectively a GIANT array of bytes. When a program is
executed, program variables are stored in memory. Everything is
stored in memory somewhere. Since everything is stored in
memory, everything has an address.

Memory addresses are effectively an index to this array of bytes.

There indexes can be very large - up to 232 - 1 on a 32-bit
platform, and up to 264 - 1 on a 64-bit platform. Memory
addresses are usually printed in hexadecimal (base 16).

In order to fully understand how pointers are used to reference

data in memory, here's a few basics on memory organisation:
• Computer memory is a large array of bytes
• A variable will be stored in one or more bytes
• On CSE machines char occupies 1 byte, int occupies 4
bytes, double occupies 8 bytes
• The & operator returns address-of a variable. We
sometimes call this a reference. i.e &number is a reference
to the variable "number"
• Almost all C implementations implement pointer values
using a variable's address in memory. Hence for almost all
C implementations, & operator returns a memory address.
• It is convenient to print memory addresses in Hexadecimal
notation.
• Variables are stored on the stack (see diagram)
• Dynamic memory is stored on the heap (see diagram)

Variables in Memory Arrays in Memory

int k; The elements of an array will be stored in consecutive memory
int m; locations.

COMP1511 Page 16
Variables in Memory Arrays in Memory
int k; The elements of an array will be stored in consecutive memory
int m; locations.
printf( "address of k is %p\n", &k ); int a[5];
// prints address of k is 0xbffffb80 int i = 0;
printf( "address of m is %p\n", &m ); while (i < 5) {
// prints address of k is 0xbffffb84 printf("address of a[%d] is %p\n", i, &a[i]);
}
k occupies the 4 bytes from 0xbffffb80 to 0xbffffb83 // prints:
m occupies the four bytes from 0xbffffb84 to 0xbffffb87 // address of a[0] is 0xbffffb60
// address of a[1] is 0xbffffb64
// address of a[2] is 0xbffffb68
// address of a[3] is 0xbffffb6c
// address of a[4] is 0xbffffb70
Note: to print the address in printf, use %p.

Pointers
A pointer is a data type whose value is a reference to another variable.
int *ip; // pointer to int
char *cp; // pointer to char
double *fp; // pointer to double
In most C implementations, pointers store the memory address of the variable they refer to; i.e. they
point to the variable, whose address they store.

The & (address-of) operator returns a reference to a variable.

The * (dereference - 'de-reference' - go to the reference to get the value there ) operator accesses the
variable referred to by the pointer.

Pointer syntax:
[type] *[some_name] = &[something];
For example:
int *my_pointer = &my_variable;
Importantly: the value of the pointer is the address of the variable it points to

For example:
int i = 7;
int *ip = &i;
printf("%d\n", *ip); // prints 7
*ip = *ip * 6;
printf("%d\n", i); // prints 42
i = 24;
printf("%d\n", *ip); // prints 24

Like other variables, pointers need to be initialised before they are used. It is best if novice programmers
initialise pointers as soon as they are declared. The value NULL can be assigned to a pointer to indicate it
does not refer to anything. NULL is a #define in stdio.h and NULL and 0 are interchangeable (in modern
C), however most programmers prefer NULL for readability.

Size of Pointers
Just like any other variable of a certain type, a variable that is a pointer also occupies space in memory.
The number of bytes depends on the computer's architecture.
• 32-bit platform: pointers are likely to be 4 bytes
• 64-bit platform: pointers are likely to be 8 bytes
• Tiny embedded CPU: pointers could be 2 bytes (e.g. your microwave)

Pointer Arguments
When we pass primitive variable types as arguments to functions, they are passed by value and any
changed made to them are not reflected in the caller function. Recall that scanf is a function, that takes
a variable from the main function as an argument. How does a function like scanf manage to update the
value of a variable found in the main function? It takes pointers to those variables as arguments!!

We use pointers to pass variables by reference. By passing the address of a variable rather than its
variable, we can change the value of that variable and have these changes be reflected in the caller
function.
int main(void) {

COMP1511 Page 17
int main(void) {
int i = 1;
increment(&i);
printf("%d\n", i); The function increment is called
//prints 2 with the address of i , and the
memory of the address is passed
return 0; to the variable *n. The function
} increases the value of the location
referenced by n by 1.
void increment(int *n) {
*n = *n + 1;
}

In a sense, pointer arguments allow a function to 'return' more than one value. This increases the
versatility of function. For examples, scanf is able to read multiple values and it uses its return value as
an error status.

You have to be extremely careful when returning pointers. Returning a pointer to a local variable is
illegal - that variable is destroyed when the function returns. But, you can return a pointer that was
given as an argument.
int increment(int *n) {
*n = *n + 1;
return n;
}
Nested calling of functions is now possible: increment(increment(&i));

Array Representation
A C array has a very simple underlying representation, it is We can even use another pointer to act as the array name.
stored in an unbroken memory block and a pointer is kept at
int nums[] = {1, 2, 3, 4, 5};
the beginning of the block. int *p = nums;
char s[] = "Hi!"; printf("%d\n", nums[2]);
printf("s: %p *s: %c\n\n", s, *s); printf("%d\n", p[2]);
printf("&s[0]: %p s[0]: %c\n", &s[0], s[0]); // both print: 3
printf("&s[1]: %p s[1]: %c\n", &s[1], s[1]);
Since nums acts as a pointer we can directly assign its value to the
printf("&s[2]: %p s[2]: %c\n", &s[2], s[2]);
pointer p.
printf("&s[3]: %p s[3]: %c\n", &s[3], s[3]);
// prints
// s: 0x7fff4b741060 *s: H We can even make a pointer point to the middle of an array:
// &s[0]: 0x7fff4b741060 s[0]: H int nums[] = {1, 2, 3, 4, 5};
// &s[1]: 0x7fff4b741061 s[1]: i int *p = &nums[2];
// &s[2]: 0x7fff4b741062 s[2]: ! printf("%d %d\n", *p, p[0]);
// &s[3]: 0x7fff4b741063 s[3]: So what is the difference between an array variable and a pointer?
int i = 5;
Since array variables are pointers, it now becomes clear why p = &i; // this is OK
we can pass arrays to scanf without the need for address- nums = &i; // this is an error
of(&) and why arrays are passed to functions by reference.
Unlike a regular pointer, an array variable is defined to point to the
beginning of the array, it is constant and may not be modified
A good explanation about pointers and arrays:
https://edstem.org/courses/1950/discussion/80387

Pointer Comparison Pointer Summary

Pointers can be tested for equality or relative order. Pointers:
double ff[] = {1.1, 1.2, 1.3, 1.4, 1.5, 1.6}; • are a compound type
double *fp1 = ff; • usually implemented with memory addresses
double *fp2 = &ff[0]; • are manipulated using address-of(&) and dereference()
double *fp3 = &ff[4]; • should be initialised when declared
printf("%d %d\n", (fp1 > fp3), (fp1 == fp2)); • can be initialised to NULL
// prints: 0 1 • should not be dereferenced if invalid
Note that we are comparing the values of the pointers, i.e., • are used to pass arguments by reference
memory addresses, not the values the pointers are pointing to • are used to represent arrays
• should not be returned from functions if they point to local variables

COMP1511 Page 18
COMP1511 Page 19
Extra C Features
Tuesday, 17 April 2018 9:06 PM

C Features we don't want you to use but are telling you what they are and how to use them
global variables
Variables declared outside of any function are available to all functions. They are called external variables or global
variables.
int g = 12;

void f(void) {
printf("The value of g is %d\n", g); // prints 12
g = 42;
}

int main(void) {
f();
printf("The value of g is %d\n", g); // prints 42
return 0;
}

Avoid global variables - they are NOT needed in COMP1511.

They make concurrency (threads) problematic and create hidden dependencies between parts of a program. They also
make code harder to reuse and pollute the namespace - create a valid name everywhere you might accidentally use
them. Global variables generally reduce readability although they can be useful for "meta"-purposes. e.g. turning on-off
debug logging through your program.

static functions
Functions are shared between files by default. This is undesirable in large programs because name clashes become more
likely. Name clashes also make code difficult to reuse.

The keyword static makes functions visible only within the file. In other words, static limits the function's scope. If a
function doesn’t need to be visible declare it static e.g.
static double helper_function(int x, double y);
It allows files to be de facto modules in C. Similarly static makes global variables visible only within the file. Beware,
static has different meanings for local (function) variables.

When a function is called, its variables are created. When a function returns, its variables are destroyed. static changes
the lifetime of a function's (local) variable. The value is preserved between function calls. Static variables make
concurrency difficult and it also makes programs harder to read and understand. There is rarely a good reason to use
static variables - so do NOT use them in COMP1511. Note: there is very different meaning to using static outside
functions; poor language design much.

For example, here is a function that counts how many times it has been called.
void count(void) {
static int call_count = 0;
call_count++;
printf("I have been called %d times\n", call_count);
}

More C Operators
C provides some additional operators, which allow shorted statements. This can make your code a little more readable
or a lot less readable.

• Pre/post-increment: ++i, i++ is the same as i = i + 1

• Pre/post-decrement: --i, i-- is the same as i = i - 1
• Compound assignment operators
○ a += b same as a = a + b
○ a -= 5 same as a = a - 5
○ a *= -10 same as a = a * -10
○ a /= 2 same as a = a / 2
○ a %= b same as a = a % b
++ and -- can be used in expression, but this is NOT recommended in COMP1511.

They can be used after the variable:

k = 7;
n = k--; // assign k to n, then decrement k by 1
printf("%d %d", k, n) // k=6, n=7

COMP1511 Page 20
printf("%d %d", k, n) // k=6, n=7

They can be used before the variable:

k = 7;
n = --k; // decrement k by 1, then assign k to n
printf("%d %d", k, n) // k=6, n=6

The for loop break and continue

There is a construct called a for loop: break causes a loop to terminate; no more iterations are
for (expr1; expr2; expr3) { performed, and execution moves to whatever comes after
statements; the loop.
} continue causes the current iteration of the loop to
terminate, and execution moves to the next iteration.
• expr1 is evaluated before the loop starts
• With while and do loops, the conditional expression
• expr2 is evaluated at the beginning of each loop; if it is
is tested before moving to the next iteration
non-zero, the loop is repeated
• With for loops, expr3 is execute, the expr2 is tested
• expr3 is evaluated at the end of each loop
before moving to the next iteration
Example of for loop:
break and continue used sparingly can make code more
for (x = 1; x <= 10; x++) { readable. However, the overuse of break and continue
printf("%d\n", x * x); can make code incomprehensible.
}
We can declare a variable if it is used only within the loop: Here is a typical use of break:
for (int x = 1; x <= 10; x++) { for (int i = 0; i < LIMIT; i++) {
printf("%d\n", x * x); // lots of complex things happens here
} if (/* need to stop loop immediately */ ) {
break; // exit loop immediately
}
For loops and while loops are equivalent. Any of the 3
// lots more complex things happens here
expressions in the for loop can be omitted, but the ';' must
}
still be present. For example:
printf("Enter starting number for Countdown:
Here is a typical use of continue:
");
scanf("%d", &n); // initial value entered by for (int i = 0; i < LIMIT; i++) {
user // lots of complex things happens here
for (; n >= 0; n--) { if (/* this is not what is wanted */ ) {
printf("%d\n", n ); continue; // go to next loop iteration
} }
printf("Blast Off!\verb|\n|"); // lots more complex things happens here
}
Although NOT recommended, the comma operator ',' can be
used to squeeze multiple statements into expr1 and expr3.
For example:
for (int x=0, y=2; x < MAX; x++, y++) {
...
}

Exiting a Program
In main, return will terminate a program.
stdlib.h provides a function useful outside main:
void exit(int status);
status is passes to exit the same a return value of main.
stdlib.h defines EXIT_SUCCESS and EXIT_FAILURE. An EXIT_SUCCESS program means the program execute successfully,
while and EXIT_FAILURE program means that the program stopped due to an error. EXIT_SUCCESS == 0 on Unix like and
almost all other systems.

Implicit Type Conversions Explicit Type Conversions

Recall that C supports 'hybrid' arithmetic operations C allows us to perform our own, explicit type casts, using the
involving certain types, in a way that mirrors our syntax (type). For example:
expectations. For example 3 + 5.8 double d1 = 1 / 2;
An integer is added to a double, giving a double result. double d2 = 1 / (double) 2;
However at the machine level, floating point addition // d1 = 0.000000
requires two double arguments and is a distinct operation // d2 = 0.500000
from integer addition.
The values of d1 and d2 will be different.
Implicit Conversions
It is good programming style to identify potentially unsafe
The compiler steps in and performs and automatic
implicit conversion types and make them explicit.
conversion known as cast, from integer to double.
#include <limits.h>
double d = 3; // 3 is converted to double
#include <assert.h>
int i = 5;
...

COMP1511 Page 21
The compiler steps in and performs and automatic
implicit conversion types and make them explicit.
conversion known as cast, from integer to double.
#include <limits.h>
double d = 3; // 3 is converted to double
#include <assert.h>
int i = 5;
...
d = d + i; // i is converted to double
assert(i >= CHAR_MIN && i <= CHAR_MAX);
char c = (char) i; // for some int i
Implicit conversions are generally performed when Note: When using explicit casts the compiler will often assume
considered 'safe'. e.g. numeric types are converted to that you know what you are doing and not issue warnings
other numeric types with larger capacity. But sometimes even when a cast is very likely to be unsafe. For example:
unsafe implicit conversions are also performed. This is an
aspect of C that is often criticised. Consider: int i = 1000;
char c = (char) i;
int i = 1000; int *ip = (int *) i;
char c1 = 100; // statically checked, OK int nums[] = {0};
char c2 = 1000; // statically checked, printf("%c\n", (char) i);
warning printf("%s\n", (char *) &i);
char c3 = i; // no warning printf("%s\n", (char *) nums);
Here the casts are used to view one type as another. This is
You should be mindful of implicit conversions. Often they often dangerous!!!
make coding easier, but sometimes they mask
programming errors.

typedef
We use the keyword typedef to give a name to a type:
typedef double real;
This means variables can be declared as numeric (real), but they will be actually be of type double. Do not overuse
typedef - it can make programs harder to read. For example:
typedef int andrew;
andrew main(void) {
andrew i,j;
…

Using typedef to make programs portable:

Suppose we have a program that does floating-point calculations. If we use a typedef'ed name for all variables e.g.
typedef double real;

real matrix[1000][1000][1000];

real my_atanh(real x) {
real u = (1.0 - x)/(1.0 + x);
return -0.5 * log(u);
}
If we move to a platform with little RAM, we can save memory (and lose precision) by changing the typedef:
typedef float real;

structs Assigning structs

We have seen simple types: int, char, double. Variables of these types Unlike arrays, it is possible to copy all components of a structure in a
hold single values. single assignment:
We have seen a compound type: array. Array variables hold multiple struct student_details student1, student2;
values. They are homogenous - as in every element in the array is of ...
the same type - and they elements are selected using integer index. student2 = student1;
Array sizes can be determined at runtime.
It is not possible to compare all components within a single
comparison:
We will now introduce another compound type: structs
structs hold multiple values (fields). They are heterogenous - as in the if (student1 == student2) // NOT allowed!
fields can be of different types . struct fields are selected using their If you want to compare two structures, you would need to write a
names and their fields are fixed. function to compare them component-by-component and decide
whether they are the "same".
Example of a struct:
If we define a struct that holds COMP1511 student details: structs and functions
#define MAX_NAME 64 A structure can be passed as a parameter to a function:
#define N_LABS 10 void print_student(student_t student) {
struct student { printf("%s z%d\n", d.name, d.zid);
int zid; }
char name[MAX_NAME];
double lab_marks[N_LABS]
double assignment1_mark; Unlike arrays, a copy will be made of the entire structure, and only this
double assignment2_mark; copy will be passed to the function. Unlike a function a function can
}; // need semi-colon to indicate you are done return a struct:

COMP1511 Page 22
 char name[MAX_NAME];
double lab_marks[N_LABS]
double assignment1_mark; Unlike arrays, a copy will be made of the entire structure, and only this
double assignment2_mark; copy will be passed to the function. Unlike a function a function can
}; // need semi-colon to indicate you are done return a struct:
// declaring the struct student_t read_student_from_file(char filename[]) {
We can declare an array to hold the details of all students: ....
}
struct student comp1511_students[900];
Note: You can't just assign strings to structs. E.g
Pointers to structs
comp1511_students[0].name = "Andrew";
If a function needs to modify a structs field or if we want to avoid the
You need to use strcpy: inefficiency of copying the entire struct, we can instead pass a pointer
strcpy(comp1511_students[0].name, "Andrew"); to the struct as a parameter:
Combining structs and typedef int scan_zid(student *s) {
A common use of typedef is to give a name to a struct type. return scanf("%d", &((*s).zid));
}
struct student {
int zid; The "arrow" operator is more readable:
char name[64]; int scan_zid(student *s) {
double lab_marks[N_LABS]; return scanf("%d", &(s->zid));
double assignment1_mark; }
double assignment2_mark; If s is a pointer to a struct, s->field is equivalent to (*s).field
};

typedef struct student student_t;

student_t comp1511_students[900];
Programmers often use convention to separate type names.
e.g. _t suffix

Nested Structures
One structure can be nested inside another:
typedef struct date Date;
typedef struct time Time;
typedef struct speeding Speeding;

struct date {
int day, month, year;
};
struct time {
int hour, minute;
};
struct speeding {
Date date;
Time time;
double speed;
char plate[MAX_PLATE];
};

COMP1511 Page 23
Malloc
Wednesday, 2 May 2018 2:08 PM

malloc and free

For example, let's assume we need a block of memory to hold a string of say 100 000 000 ints
int *p;
p = malloc(100000000 * sizeof (int));
if(p == NULL) {
fprintf(stderr, "Error: could not allocate memory!\n");
exit(1); // means something went wrong
}

// we can now use the pointer

// … lots of things to do

free(p); // free up the memory that was used

malloc() allocates memory in "the heap" and returns a pointer to a block of memory (it returns the
address of the memory it is allocated). malloc() returns a (void *) pointer and can be assigned to any
pointer type. If insufficient memory is available then malloc() returns NULL.

Memory "lives" on forever until we free it. free() indicates that you have finished using a block of
memory. Continuing to use memory after memory after free() results in in very nasty bugs. Using
free() on a memory block twice can also cause bad bugs. If a program keeps calling malloc() without
corresponding free() calls, then the program's memory will grow steadily larger. This is called a
memory leak. Memory leaks are major issues for long running programs.

Newton's 3rd Law of Memory Management:

'For every malloc, there is an equal and opposite free"
Why? Because memory is a finite resource.

sizeof
sizeof is a C operator that yields bytes when needed for a type or variable.
You use it like this:
sizeof (type)
sizeof variable_name
Note: unusual syntax (badly designed) brackets indicate argument is a type
You should use sizeof for every malloc call.
Here are some more examples of using sizeof:
printf("%ld", sizeof (char)); // 1
printf("%ld", sizeof (int)); // 4 commonly
printf("%ld", sizeof (double)); // 8 commonly
printf("%ld", sizeof (int[10])); // 40 commonly
printf("%ld", sizeof (int *)); // 4 or 8 commonly
printf("%ld", sizeof "hello"); // 6

Syntax = rules of the language

Semantics = meaning behind the language

COMP1511 Page 24
Linked Lists
Wednesday, 2 May 2018 2:08 PM
Finding an item in a list: using a shorter while loop
Same function but using a more concise while loop.
Self-Referential Structures
Shorter does not always mean more readable.
We can define a structure containing a pointer to the same type of structure like this:
// return pointer to first node containing
struct node {
// specified value, return NULL if no such node
struct node *next;
int data;
struct node *find_node(struct node *head, int data) {
};
struct node *n = head;
These "self-referential" pointers can be used to build larger "dynamic" data structures out of while (n != NULL && n->data != data) {
smaller building blocks. n = n->next;
}
Linked Lists return n;
The most fundamental of these dynamic data structures it the Linked List. }

It is based on the idea of a sequence of data items or nodes. Linked lists are more flexible Finding an item in a list: recursive
than arrays: Same function but function calls itself.
• Items do not have to be located next to each other in memory // return pointer to first node containing
• Items can easily rearranged by altering pointers // specified value, return NULL if no such node
• The number of items can change dynamically
• Items can be added or removed in any order struct node *find_node(struct node *head, int data) {
if (head == NULL) {
return NULL;
}
if (head->data == data) {
return head;
A linked list is a sequence of items. }
Each item contains data and a pointer to the next item. You need to separately store a return find_node(head->next, data);
pointer to the first item or "head" of the list. The last item in the list is special - it contains }
NULL in its next field instead of a pointer to an item .
Finding an item in a list: shorter recursive
Example of List Item Same function but a more concise recursive version.
Example of a list item used to store an address Shorter does not always mean more readable.
In C code: In a diagram: // return pointer to first node containing
struct address_node { // specified value, return NULL if no such node
struct address_node *next;
char *name; struct node *find_node(struct node *head, int data) {
char *address; if (head == NULL || head->data == data) {
char *telephone; return head;
char *email; }
}; return find_node(head->next, data);
}

Printing a list: in python syntax

// print contents of list in Python syntax
void print_list(struct node *head) {
printf("[");
List Items List Operations for (struct node *n = head; n != NULL; n = n->next) {
List items may hold large amounts of data or Some basic lists operations are: printf("%d", n->data);
many fields. For simplicity, we'll assume • Creating a new item with specified data if (n->next != NULL) {
printf(", ");
each list item only needs to store a single int. • Searching for an item with particular data
}
struct node { • Inserting a new item to the list
}
struct node *next; • Removing an item from the list
printf("]");
int data; Many other operations are possible.
}
};
Finding the last item in a list
Here are various codes showing what you can do with linked lists:
Creating a list item // return pointer to last node in list
// NULL is returned if list is empty
// Create a new struct node containing the specified data,
// and next fields, return a pointer to the new struct node.
struct node *last(struct node *head) {
if (head == NULL) {
struct node *create_node(int data, struct node *next) {
return NULL;
struct node *n;
}
n = malloc(sizeof (struct node));
struct node *n = head;
if (n == NULL) {
while (n->next != NULL) {
fprintf(stderr, "out of memory\n");
n = n->next;
exit(1);
}
}
return n;
n->data = data;
}
n->next = next;
return n;
} Appending to a list
// append integer to end of list
Building a list struct node *append(int value, struct node *head) {
struct node *n;
// Building a list containing the 4 ints: 13, 17, 42, 5
n = create_node(value, NULL);
if (head == NULL) {
struct node *head = create_node(5, NULL);
// new node is now head of the list
head = create_node(42, head);
return n;
head = create_node(17, head);
} else {
head = create_node(13, head);
struct node *l = last(head);
l->next = n;
Summing a list return list;

COMP1511 Page 25
 return n;
head = create_node(17, head);
} else {
head = create_node(13, head);
struct node *l = last(head);
l->next = n;
Summing a list return list;
// return sum of list data fields }
}
int sum(struct node *head) {
int sum = 0; Deleting all items from a list
struct node *n = head;
// Delete all the items from a linked list.
// execute until end of list
while (n != NULL) {
void delete_all(struct node *head) {
sum += n->data;
struct node *n = head;
// make n point to next item
struct node *tmp;
n = n->next;
while (n != NULL) {
}
tmp = n;
return sum;
n = n->next;
}
free(tmp);
}
Summing a list: using a for loop }
// return sum of list data fields
Inserting a node into an ordered list
int sum(struct node *head) { struct node *insert(struct node *head, struct node *node) {
int sum = 0; struct node *previous;
for (struct node *n = head; n != NULL; n = n->next) { struct node *n = head;
sum += n->data; // find correct position
} while (n != NULL && node->data > n->data) {
return sum; previous = n;
} n = n->next;
}
Summing a list: recursive // link new node into list
Same function but using a recursive call. if (previous == NULL) {
// return sum of list data fields head = node;
} else {
int sum2(struct node *head) { previous->next = node;
if (head == NULL) { }
return 0; node->next = n;
} return head;
return head->data + sum2(head->next); }
}
Inserting a node into an ordered list: recursive
Finding an item in a list struct node *insert(struct node *head, struct node *node) {
// return pointer to first node containing if (head == NULL || head->data >= node->data) {
// specified value, return NULL if no such node node->next = head;
return node;
struct node *find_node(struct node *head, int data) { }
struct node *n = head; head->next = insert(head->next, node);
// search until end of list reached return head;
while (n != NULL) { }
if (n->data == data) {
// matching item found Deleting a node from a list
return n; struct node *delete(struct node *head, struct node *node) {
} if (node == head) {
// make n point to next item head = head->next; // remove first item
n = n->next; free(node);
} } else {
// item not in list struct node *previous = head;
return NULL; while (previous != NULL && previous->next != node) {
} previous = previous->next;
}
Finding an item in a list: using a for loop if (previous != NULL) { // node found in list
// return pointer to first node containing previous->next = node->next;
// specified value, return NULL if no such node free(node);
} else {
struct node *find_node(struct node *head, int data) { fprintf(stderr, "warning: node not in list\n");
for (struct node *n = head; n != NULL; n = n->next) { }
if (n->data == data) { }
return n; return head;
} }
}
return NULL; Deleting a node from a list: recursive
} struct node *delete(struct node *head, struct node *node) {
if (head == NULL) {
fprintf(stderr, "warning: node not in list\n");
} else if (node == head) {
head = head->next; // remove first item
free(node);
} else if (head == head) {
head->next = delete(head->, node)
}
return head;
}

COMP1511 Page 26
Multiple C files, Header files
Tuesday, 15 May 2018 1:17 PM

We can make programs with multiple .c files. It is no different from making/compiling a program from one .c file.
dcc -o program_name first.c second.c
Note: You cannot have more than one main function. So make sure there is only ONE main function amongst the
code in your .c files

Using functions from another file

first.c second.c first.h
void hello(void) { #include <stdio.h> #ifndef FIRST_H
printf("Hello!\n"); #include "first.h" #define FIRST_H
}
int square(int n) { int main(void) { void hello(void);
return n*n; hello(); int square(int n);
} printf("%d", square(5));
} #endif
Notes:
• In second.c , we use for #include "first.h" use quotation marks because it is not a library but something I
made
• #ifndef is a header guard. This avoids errors if the .h file is included multiple times
• Don’t include the .h file in the command line.
• When you has define a .h file, it needs to be in the same folder as the .c files using it.

COMP1511 Page 27
Stacks and Queues
Wednesday, 16 May 2018 2:24 PM

Stacks and Queues are ubiquitous data structures in computing. They are part of many algorithms and are a good
example of abstract data types.

A type is concrete if a user of that type has knowledge of how it works.

A type is abstract if a user has no knowledge of how it works.
Stack - Abstract Data Type Queue - Abstract Data Type
A stack is a collection of items such that the last item to A queue is a collection of items such that the first item to enter
enter is the first one to exit (LIFO - Last In, First Out). They is the first one to exit (FIFO - First In, First Out). It is based on the
are based on the idea of stacks of books, or plates. idea of queueing at a bank or shop. Essential queue operations
Essential stack operations include: include:
• push() - add new item to stack • enqueue() - add new item to queue
• pop() - remove top item from stack • dequeue() - remove front item from queue
Additional stack operations are: Additional queue operations are:
• top() - fetch top item (but don't remove it) • front() - fetch the front item (but don't remove it)
• size() - no. of items • size() - no. of items
• is_empty() - check if stack is empty • is_empty() - check is queue is empty

Stack Applications Queue Applications

Stacks are used for: Queues are used for:
• Page-visited history in a web browser • Waiting lists, bureaucracy
• Undoing a sequence in a text editor • Access to shared resource (printers, etc.)
• Checking for balanced brackets • Phone call centres
• HTML tag matching • Multiple processors in a computer
• Postfix (RPN) calculator
• Chain of function calls in a program

Stack - Abstract Data Type - C Interface Queue - Abstract Data Type - C Interface
typedef struct stack_internals *stack; typedef struct queue_internals *queue;
stack stack_create(void); queue queue_create(void);
void stack_free(stack stack); void queue_free(queue queue);
void stack_push(stack stack, int item); void queue_enqueue(queue queue, int item);
int stack_pop(stack stack); int queue_dequeue(queue queue);
int stack_is_empty(stack stack); int queue_is_empty(queue queue);
int stack_top(stack stack); int queue_front(queue queue);
int stack_size(stack stack); int queue_size(queue queue);
Using stacks in C interface:
stack s; Using queues in C interface:
s = stack_create(); queue q;
stack_push(s, 10); q = queue_create();
stack_push(s, 11); queue_enqueue(q, 10);
stack_push(s, 12); queue_enqueue(q, 11);
printf("%d\n", stack_size(s)); // prints 3 queue_enqueue(q, 12);
printf("%d\n", stack_top(s)); // prints 12 printf("%d\n", queue_size(q)); // prints 3
printf("%d\n", stack_pop(s)); // prints 12 printf("%d\n", queue_front(q)); // prints 10
printf("%d\n", stack_pop(s)); // prints 11 printf("%d\n", queue_dequeue(q)); // prints 10
printf("%d\n", stack_pop(s)); // prints 10 printf("%d\n", queue_dequeue(q)); // prints 11
The implementation of a stack is opaque (hidden from the printf("%d\n", queue_dequeue(q)); // prints 12
user). Users program cannot depend on how stack is Like stacks the implementation of queues is opaque. Queue
implemented. Stack implementation can change without implementation can change without risk of breaking user
risk of breaking users program. This type of information programs.
hiding is crucial to managing complexity in large software
systems.
Implementing A Stack with a Linked List
A stack can be implemented using a linked list, by adding and removing at the head
[push() and pop()]. For a queue, we need to either add or remove at the tail.

Adding an item to the Tail of a list:

COMP1511 Page 28
Adding an item to the Tail of a list:

Adding an item to the tail is achieved by making the last node of the list point to the new
node. We first need to scan along the list to find the last item.
struct node *add_to_tail( *new_node, struct node *head) {
if (head == NULL) { // list is empty
head = new_node;
} else { // list not empty
struct node *node = head;
while (node->next != NULL) {
node = node->next; // scan to end
}
node->next = new_node;
}
return head;
}
Efficiency Issues:
Unfortunately, this implementation is very slow. Every time a new item is inserted, we
need to traverse the entire list (which could be very large). We can do the job more
efficiently if we retain a direct link to the last item of "tail" of the list
if (tail == NULL) { // list is empty
head = node;
} else { // list not empty
tail->next = node;
}
tail = node;
Note: There is no way to efficiently remove items from the tail

Postfix Calculator Example

Some early calculators and programming languages used convention known as Reverse
Polish Notation (RPN), where the operator comes after the two operands rather than
between them:
12+
result = 3
32*
result = 6
43+6*
result = 42
1234+*+
result = 15

A calculator using RPN is called a Postfix Calculator, it can be implemented using a stack:
• When a number is entered: push it onto the stack
• When an operator is entered: pop the two items from the stack, apply the
operator to them, and push the result back onto the stack
#include <stdio.h>
#include <ctype.h>
#include "stack.h"
int main(void) {
int ch;
stack s = stack_create();
while ((ch = getc(stdin)) != EOF) {
if (ch == '\n') {
printf("Result: %d\n", stack_pop(s));
} else if (isdigit(ch)) {
ungetc(ch, stdin); // put first digit back

COMP1511 Page 29
 ungetc(ch, stdin); // put first digit back
int num;
scanf("%d", &num); // now scan entire number
stack_push(s, num);
} else if (ch == '+' || ch == '-' || ch == '*') {
int a = stack_pop(s);
int b = stack_pop(s);
int result;
if (ch == '+') {
result = b + a;
} else if (ch == '-') {
result = b - a;
} else {
result = b * a;
}
stack_push(s, result);
}
}
}

COMP1511 Page 30
Illegal C
Wednesday, 23 May 2018 2:33 PM

Consequences of bugs
• A compilers gives syntax/semantic errors (if you're very lucky)
• The program halts with run-time error (if you're lucky)
• The program never halts (if you're lucky-ish)
• The program halts, but with incorrect results (if you're unlucky)
• The program appears correct, but has security holes (if you're unlucky)

Changed Variable
int a[10];
int b[10];
printf("a[0] is at address %p\n",&a[0]);
printf("a[9] is at address %p\n", &a[9]);
printf("b[0] is at address %p\n",&b[0]);
printf("b[9] is at address %p\n", &b[9]);
for (int i = 0; i < 10; i++) {
a[i] = 77;
}
for (int i = 0; i <= 12; i++) {
b[i] = 42;
}
for (int i = 0; i < 10; i++) {
printf("%d ", a[i]);
}
printf("\n");
The C program assigns to b[10]..b[12], which do not exist.
The consequence could be anything - a C implementation is permitted to behave in
any manner given an invalid program. On gcc 6.3 on Linux/x86 64 it happens to
change b[0] to 42:
$ gcc invalid_array_index0.c
$ a.out
a[0] is at address 0x7fffc9cbcbf0
a[9] is at address 0x7fffc9cbcc14
b[0] is at address 0x7fffc9cbcbc0
b[9] is at address 0x7fffc9cbcbe4
42 77 77 77 77 77 77 77 77 77

Changed Termination
int i;
int a[10];
printf("i is at address %p\n", &i);
printf("a[0] is at address %p\n", &a[0]);
printf("a[9] is at address %p\n", &a[9]);
printf("a[11] would be stored at address %p\n", &a[10]);
for (i = 0; i <= 11; i++) {
a[i] = 0;
}
Another invalid C program assigning to a non-existent array element.
On gcc 6.3 on Linux/x86 64 it happens to assigns to i and the loop doesn't terminate.
So a character error makes the program invalid, and seemingly certain termination
does not occur.
$ gcc invalid1.c
$ a.out

COMP1511 Page 31
$ a.out
i is at address 0x7fffbb72bfdc
a[0] is at address 0x7fffbb72bfb0
a[9] is at address 0x7fffbb72bfd4
a[10] is equivalent to address 0x7fffbb72bfd8
....

Changed Variable in Another Function

int main(void) {
int answer = 36;
f(5);
printf("answer=%d\n", answer); // prints 42
return 0;
}
void f(int x) {
int a[10];
// a[19] doesn't exist
// with gcc 6.3 on Linux/x86_64 variable answer
// in main happens to be where a[19] would be
a[19] = 42;
}
Yet another invalid C program assigning to a non-existent array
element. On gcc 6.3 on Linux/x86 64 it changes the variable answer in the
calling function main.
$ gcc invalid2.c
$ a.out
answer=42

Changed Function Return Location

void f() {
int a[10];
// on gcc-6.3/Linux x86
// change function's return address on stack
// causing function to return after the line
// where answer is assigned 24
a[14] += 7;
}

int main(void) {
int answer = 42;
f();
answer = 24;
printf("answer=%d\n", answer);
return 0;
}
Yet another invalid C program assigning to a non-existent array element.
With gcc 6.3 on Linux/x86 64 it changes where the function returns in main.
$ gcc invalid3.c
$ a.out
answer=42

Bypassing Authentication
int authenticated = 0;
char password[8];
printf("Enter your password: ");
gets(password);

COMP1511 Page 32
gets(password);
if (strcmp(password, "secret") == 0) {
authenticated = 1;
}
// a password longer than 8 characters will overflow
// array password on gcc 6.3 on Linux/x86_64 this can
// overwrite the variable authenticated and allow access
if (authenticated) {
printf("Welcome. You are authorized.\n");
} else {
printf("Welcome. You are unauthorized. ");
printf("Your death will now be implemented.\n");
printf("Welcome. You will experience ");
printf("a tingling sensation and then death. \n");
printf("Remain calm while your life is extracted.\n");
}
Yet another invalid C program assigning to a non-existent array element.
A password longer than 8 characters will overflow the array password. This is often
termed buffer-overflow.
$ gcc invalid4.c
$ a.out
Enter your password: secret
Welcome. You are authorized.
$ a.out
Enter your password: wrong
Welcome. You are unauthorized.
Your death will now be implemented.
Welcome. You will experience a
tingling sensation and then death.
Remain calm while your life is extracted.
$ a.out
Enter your password: longcorrectpassword
Welcome. You are authorized.

Implementation vs Language
C was designed for much smaller, slower computers (28K of RAM, 1mhz clock).
Program speed/size is much more important for programs than dominated language
choice.
Most C implementations still focus on maximising performance of valid programs.
Most C implementations do not check array bounds or for arithmetic overflow
because this have performance costs. The C definition does not entail this.
A C implementation can check array bounds and halt if invalid indexes are used.
A C implementation could check and halt if an uninitialised value is used - but is
difficult/expensive to track for arrays.

Address Sanitizer extension to gcc/clang

gcc -fsantitize=address gives a very different C implementation. Invalid array indexes,
pointer dereferences and some other invalid use of the string library function are
detected. Performance cost - execution from 1.2-10+x slower.
Information cryptic but note source code line indicated. e.g.:
$ cd /home/cs1511/public_html/lec/illegal_C/code/
$ gcc -g -fsanitize=address debug_examples.c
$ ./a.out 3
ASAN:DEADLYSIGNAL
==========================
==16917==ERROR: AddressSanitizer: SEGV on unknown address
0x000000000014 (pc 0x55819087cd2c bp 0x7ffd02a40bb0 ....
#0 0x55819087cd2b in test3 debug_examples.c:33

COMP1511 Page 33
 #0 0x55819087cd2b in test3 debug_examples.c:33
#1 0x55819087d19c in main debug_examples.c:96
#2 0x7fccf078d2b0 in __libc_start_main (/lib/...
#3 0x55819087caf9 in _start ...
....

dcc uses -fsanitize=address (with clang) but makes messages more comprehensible for
beginner programmers:
$ cd /home/cs1511/public_html/lec/illegal_C/code/
$ dcc debug_examples.c
$ ./a.out 3
ASAN:DEADLYSIGNAL

debug_examples.c:33 runtime error - illegal array, pointer

or other operation
Execution stopped in test3() in debug_examples.c line 33:

int *a = NULL;
// dereferencing NULL pointer
--> a[5] = 42;
}
Values when execution stopped:

Address Sanitizer does not detect the use of uninitialised values. e.g.:
% ./debug_examples 4
0
1
2
3
-2115323248
5
6
7
8
9

valgrind - another debugging/testing tool

valgrind works on x86 machine code - not C specific.
valgrind runs the code on a virtual machine and detects use of uninitialised memory.
It also picks up many invalid array indexes and pointer dereferences: Large
performance penalty - and slow start time.
% valgrind ./debug_examples 4
==1932== Memcheck, a memory error detector
==1932== Copyright (C) 2002-2010, and GNU GPL'd, ...
==1932== Using Valgrind-3.6.1 and LibVEX; rerun ...
==1932== Command: ./debug_examples 4
==1932==
0
1
2
3
==1932== Use of uninitialised value of size 8
==1932== at 0x521AF0B: _itoa_word (_itoa.c:195)
==1932== by 0x521D3B6: vfprintf (vfprintf.c:1619)
==1932== by 0x400FBF: test4 (debug_examples.c:45)
==1932== by 0x401317: main (debug_examples.c:92)
==1932==
...

COMP1511 Page 34
...

dcc -valgrind
dcc -valgrind causes valgrind to be used to run your program. It makes messages more
comprehensible for beginner programmers:
$ dcc --valgrind debug_examples.c
% ./a.out 4
Runtime error: uninitialized variable accessed.
Execution stopped in test4() debug_examples.c line 45:
// accessing uninitialized array element (a[4])
for (i = 0; i < 10; i++)
--> printf("%d\n", a[i]);
}

Values when execution stopped:

a = {0, 1, 2, 3, -16776544, 5, 6, 7, 8, 9}
i = 4
a[i] = -16776544

COMP1511 Page 35
Searching and Sorting
Wednesday, 23 May 2018 2:33 PM

Efficiency
COMP1511 focuses on writing programs, but efficiency is also important. We often need to consider:
• Execution time
• Memory use
A correct but slow program can be useless. Efficiency often depends on the size of the data being processed.
Understanding this dependency lets us predict program performance on larger data.

We can find out whether a program is efficient or not through:

• Empirical approach - run the program several times with different input sizes and measure the time taken
• Theoretical approach - try to count the number of operations performed by the algorithm on input of size n

Searching Sorting
Aim: to rearrange a sequence so it is in non-decreasing order
Linear Search Unordered Array
Advantages:
int linear_search(int array[], int length, int x) { • Sorted sequences can be searched efficiently
for (int i = 0; i < length; i = i + 1) { • Items with equal keys are located together
if (array[i] == x) { Disadvantages:
return 1; • Simple obvious algorithms are too slow at sorting large sequences
} • Better algorithms can sort very large sequences
} Sorting has been studied extensively and many algorithms have been
return 0;
proposed.
}
One slow obvious algorithm is bubblesort and one fast algorithm is:
An informal analysis: quicksort
Operations:
• Start at the first element Bubblesort Code:
• Inspect each element in turn
void bubblesort(int array[], int length) {
• Stop when you find X or reach the end
int swapped = 1;
If there are N elements to search:
while (swapped) {
• Best case scenario: we only check 1 element swapped = 0;
• Worst case scenario: we need to check N elements for (int i = 1; i < length; i = i + 1) {
• If the element is in the list we will check on average N/2 if (array[i] < array[i - 1]) {
elements int tmp = array[i];
• If it is not in the list, it will check N elements array[i] = array[i - 1];
array[i - 1] = tmp;
Linear Search Ordered Array swapped = 1;
int linear_ordered(int array[], int length, int x) }
{ }
for (int i = 0; i < length; i = i + 1) { }
if (array[i] == x) { }
return 1; Quicksort - Code
} else if (array[i] > x) { void quicksort(int array[], int length) {
return 0; quicksort1(array, 0, length - 1);
} }
} void quicksort1(int array[], int lo, int hi) {
return 0; if (lo >= hi) {
} return;
An informal analysis: }
Operations: int p = partition(array, lo, hi);
• Start at the first element // sort lower part of array
• Inspect each element in turn quicksort1(array, lo, p);
• Stop when you find X or find a value >X or reach the end // sort upper part of array
If there are N elements to search: quicksort1(array, p + 1, hi);
• Best case scenario: we only check 1 element }
• Worst case scenario: we need to check N elements
• If the element is in the list we will check on average N/2 int partition(int array[], int lo, int hi) {
int i = lo, j = hi;
elements
int pivotValue = array[(lo + hi) / 2];
• If it is not in the list, it will check N/2 elements
while (1) {
while (array[i] < pivotValue) {
Binary Search Ordered Array i = i + 1;
int binary_search(int array[], int length, int x) { }
int lower = 0; while (array[j] > pivotValue) {
int upper = length - 1; j = j - 1;
while (lower <= upper) { }
int mid = (lower + upper)/ 2; if (i >= j) {
if (array[mid] == x) { return j;
return 1; }

COMP1511 Page 36
 while (lower <= upper) { }
int mid = (lower + upper)/ 2; if (i >= j) {
if (array[mid] == x) { return j;
return 1; }
} else if (array[mid] > x) { int temp = array[i];
upper = mid - 1; array[i] = array[j];
} else { array[j] = temp;
lower = mid + 1; i = i + 1;
} j = j - 1;
} }
return 0; return j;
} }
An informal analysis:
Operations:
• Start with an entire array Quicksort and Bubblesort Compared
• At each step half the range the element may be in If we use quicksort and bubblesort code, we see:
• Stop when you find X or when the range is empty Array size (n) Bubblesort operations Quicksort operations
If there are N elements to search:
• Best case scenario: we only check 1 element 10 81 24
• Worst case scenario: we need to check log2N + 1 elements 100 8415 457
• If the element is in the list we will check on average log2N 1000 981018 9351
elements 10000 98790120 102807
• Bubblesort is proportional to n
log2(N) grows very slowly: • Quicksort is proportional to nlog2n
• log210 = 3.3 • If n is small, there is little difference between the algorithms
• log21000 = 10 • If n is large, there is a significant difference between the
• log21000000 = 20 algorithms
• log21000000000 = 30 • For large n, you need a good sorting algorithm like quicksort
• log21000000000000 = 40
Physicists estimate 1080 atoms in universe: log 2(1080) = 240
Binary search all atoms in universe in < 1 microsecond

COMP1511 Page 37