Week 02

Week 02

1/35

Things to Note ...

• Sample solutions to tute/lab exercises on the web on Monday of the following week

In This Lecture ...

• Arrays and Pointers (Slides, Ch.7)

Coming Up ...

• Testing, Debugging, Performance (Slides)

Nerds You Should Know #2

2/35

The next in a series on famous computer scientists ...
 
 

The guy in the middle invented something you used today ... what?

... Nerds You Should Know #2

3/35

Doug Engelbart

1940's, trained as Electrical Engineer 1950's, worked at NASA Ames and SRI labs 1960's and 1970's, invented ... the mouse 2-dimensional display editing   (e.g. vi, emacs) windows-based interfaces   (e.g. Mac, Windows) etc. etc. etc.

The First Mouse

4/35

 

 
Photo circa 1963

Arrays

5/35

An array is

• a collection of same-type variables
• arranged as a linear sequence
• accessed using an integer subscript
• for an array of size N, valid subscripts are 0..N-1

int a[20];     // array of 20 integer values/variables
char b[10];    // array of 10 character values/variables

printf("%d", a[0]);  // print first value in a[] array

... Arrays

6/35

Larger example:

#define SIZE 20

int i;         // integer value used as index
int v[SIZE];   // array of 20 integer values

for (i = 0; i < SIZE; i++) {
	v[i] = twoToPower(i);
}
for (i = 0; i < SIZE; i++) {
	printf("v[%02d] = %d\n", i, v[i]);
}

Exercise: Powers of Two

7/35

Convert code on previous slide into a complete program:

• define a suitable main program
• add necessary #include's
• define a function to compute 2 ⁱ

Iteration over Arrays

8/35

Two approaches to array iteration:

• count-controlled, using size of array, e.g.

  int i;
  int v[10];
  for (i = 0; i < 10; i++) {
  	printf("%d\n", v[i]);
  }
  

• using a terminating value, e.g.

  int i;
  char s[10];
  // somehow s[] is set to hold a C-string
  for (i = 0; s[i] != '\0'; i++) {
  	printf("%c", s[i]);
  }
  

Array Initialisation

9/35

Arrays can be initialised by code (as in previous example).

Alternatively, can specify an initial set of values in declaration.

Examples:

int v[8]  = {1, 2, 4, 8, 16, 32, 64, 128};

char s[4] = {'a', 'b', 'c', '\0'};

int w[]   = {5, 4, 3, 2, 1};

In the last case, C infers the array length   (as if we declared w[5]).

Exercise: Array Initialisation

10/35

Write a program to determine answers to the following:

• what happens if we declare v[8] but only give 6 values?
• what happens if we declare v[8] but give 10 values?
• is there an equivalent but simpler way to define s?
• how can we determine the number of elements in w[]?

Array Storage

11/35

What exactly does an array declaration produce in memory?

An array declaration   T a[N];    (where T is a type; N is a count)

• allocates a region of memory of size N×sizeof(T)
• associates name a with address of first element

... Array Storage

12/35

Example array declarations:

Exercise: Array Storage

13/35

Write a program to determine the storage size of the following:

Types:

char   int   float   double

Variables:

char s[10];  int n;  int a[6];  char *cp;  int *ip;

Also, work out in memory where each variable is located.

Do this for variables declared as local, then as global.

Arrays and Pointers

14/35

An analogy of how arrays are represented:

The array name acts like a pointer to the first element   (s == &s[0])

... Arrays and Pointers

15/35

On the previous slide, we suggested   s == &s[0]

Types:   s is char[],   &s[0] is char *

There is a close relationship between pointers and arrays

• an array's name is like a constant pointer to the first element
• using an unindexed array name gives you the address of the array

• pointers are often used to reference array elements
• arrays are passed to functions by reference

Arrays and Functions

16/35

When an array is passed as a parameter to a function

• a pointer to the start of the array is actually passed

int total, vec[20];
...
total = sum(vec);

Within the function ...

• the types of elements in the array are known
• the size of the array is unknown

... Arrays and Functions

17/35

Since functions do not know how large an array is:

• pass in the size of the array as an extra parameter, or
• include a "termination value" to mark the end of the array

int total, vec[20];
...
total = sum(vec,20);

Also, since the function doesn't know the array size, it can't check whether we've written an invalid subscript (e.g. in the above example 100 or 20).

Exercise: Array functions

18/35

Implement a C function to sum the values in an integer array.

Use the following function prototype:

int  sumArray(int array[], int nelems);

Implement an array to print a C string, character at a time.

Use the following function prototype:

void  printString(char string[]);

Aside: Pointer Arithmetic

19/35

A pointer variable holds a value which is an address.

C knows what type of object is being pointed to

• it knows the sizeof that object
• it can compute where the next/previous object is located

int a[6];   // address 0x1000
int *p;
p = &a[0];  // p contains 0x1000
p = p + 1;  // p now contains 0x1004

... Aside: Pointer Arithmetic

20/35

For a pointer declared as   T *p;   (where T is a type)

• if the pointer initially contains address A
  • executing   p = p + k;   (where k is a constant)
    • changes the value in  p  to   A + k*sizeof(T)

Example:

int a[6];   (addr 0x1000)       char s[10];   (addr 0x2000)
int *p;     (p == ?)           char *q;      (q == ?)
p = &a[0];  (p == 0x1000)       q = &s[0];    (q == 0x2000)
p = p + 2;  (p == 0x1008)       q++;          (q == 0x2001)

Arrays and Pointers

21/35

An alternative approach to iteration through an array:

• determine the address of the first element in the array
• determine the address of the last element in the array
• set a pointer variable to refer to the first element
• use pointer arithmetic to move from element to element
• terminate loop when address exceeds that of last element

int *p;
int a[6];
for (p = &a[0]; p < &a[6]; p++) {
	printf("%2d ", *p);
}

... Arrays and Pointers

22/35

Pointer-based scan written in more typical style

Note: because of pointer/array connection   a[i] == *(a+i)

Exercise: Pointer-based Array Scans

23/35

Rewrite the two function from before:

int  sumArray(int array[], int nelems);

void  printString(char string[]);

to use pointer-based scans.

Do the function interfaces need to change for this?

Arrays of Strings

24/35

One common type of multi-d array is e.g. char *argv[]

Each element of argv[] is

• a pointer to the start of a character array (char *)
  • containing a \0-terminated character sequence

... Arrays of Strings

25/35

More detail on how argv is represented:

argv[] is in the stack; the strings are in the global data area

... Arrays of Strings

26/35

Normally, we treat the strings as atomic values, e.g.

for (i = 0; i < argc; i++) {
    printf("%s ",argv[i]);
}

However, because argv is an array of arrays, we could access the individual characters in the strings, e.g.

for (i = 0; i < argc; i++) {
    for (j = 0; argv[i][j] != '\0'; i++) {
        putchar(argv[i][j]);
    }
    putchar(' ');
}

argv[i] is a character array; argv[i][j] is one character from that array.

Exercise: The ohce Command

27/35

Write a program called ohce that

• displays its command line arguments
• but writes them in the reverse order
• and reverses the characters in each word

$ ./ohce these are some args
sgra emos era eseht
$ ./ohce even 123 gets  reversed
desrever steg 321 neve
$ ./ohce able was I ere I saw elba
able was I ere I saw elba

2d Arrays (Matrices)

28/35

An array of N arrays, each of size M, is a matrix.

Examples:

Note:   q[0][1]==2.7    r[1][3]==8    q[1]=={3.1,0.1}

... 2d Arrays (Matrices)

29/35

Storage representation of matrices:

... 2d Arrays (Matrices)

30/35

Iteration can be done row-by-row or column-by-column:

int row, col;   int m[NROWS][NCOLS];

//row-by-row
for (row = 0; row < NROWS; row++) {
    for (col = 0; col < NCOLS; col++) {
        ... m[row][col] ...
    }
}
// colum-by-column
for (col = 0; col < NCOLS; col++) {
    for (row = 0; row < NROWS; row++) {
        ... m[row][col] ...
    }
}

Row-by-row is the most common style of iteration.

Exercise: Matrix Operations

31/35

Consider a datatype for square matrices defined as:

#define N 10
typedef int Matrix[N][N];

Write an operation on this datatype to:

• print values from a matrix variable onto stdout

3d Arrays

32/35

Arrays can be declared with more than two dimensions, e.g.

int cube[10][10][10];     // 3 dimensional
int hyper[4][5][6][7];    // 4 dimensional

The definition of cube can be interpreted as:

• cube is an object containing 10 elements
• each element is a 10×10 matrix
• each matrix contains 10 rows
• each row contains 10 int variables
• cube contains a total of 10×10×10 = 1000 int variables

... 3d Arrays

33/35

Iteration over an N-d array requires N nested loops:

int i, j, k;   int cube[N][N][N];

// for each element (matrix)
for (i = 0; i < N; i++) {
    // for each row of current matrix
    for (j = 0; j < N; j++) {
        // for each element in current row
        for (k = 0; k < N; k++) {
            // process cube[i][j][k] (an integer)
        }
    }
}

Exercise: Cube Operations

34/35

Consider a datatype for data "cubes" defined as:

#define N 4
typedef int Cube[N][N][N];

Write an operation on this datatype to:

• print values from a matrix variable onto stdout
    (as a sequence of matrices, each preceded by a label cube[i])

Tips for Next Week's Lab

35/35

Chars, strings and other arrays

• Search for substrings
  • String library function strstr(s,sp) returns pointer to first sp in s
• Use ctype.h library
  • isupper(c), islower(c), ...
• Process command line options

  
  prompt$ ./asci -upper
  65 A	66 B	67 C	68 D	...
  73 I	74 J	75 K	76 L	...
  ...
  

• Avoid magic numbers in your program code
• Implement various 2d matrix operations (read, copy, add, compare matrices)

Produced: 11 Aug 2016