Week8.pptx

7/27/2019 Week8.pptx

1/39

CS1010: Programming Methodologyhttp://www.comp.nus.edu.sg/~cs1010/
http://www.comp.nus.edu.sg/~cs1101/http://www.comp.nus.edu.sg/~cs1101/


2/39

Week 8: Multidimensional Arrays &

Testing and Debugging

Objectives:

Understand the concept and application of

multidimensional arrays Understand how to test and debug your program

CS1010 (AY2011/2 Semester 1)

References: Chapter 6 Numeric Arrays

Week8 - 2


3/39

Week 8: Outline (1/2)

1. Week 7 Exercise #2: Set Containment2. Multidimensional Arrays

Initializers

Examples

3. Matrices Demo #1: Matrix addition

Exercise #1: Matrix multiplication

4. Maze Problem Exercise #2 (take-home): Valid path

CS1010 (AY2011/2 Semester 1) Week8 - 3


4/39

Week 8: Outline (2/2)

5. Testing and Debugging5.1 Demo #2: Using printf()

5.2 Debugging with assert(): Demo #3: using assert()

5.3 Exercise #3: Using assert()

5.4 Testing thoroughly

5.5 Using a debugger gdb (Demo #4)

5.6 Find the bug

5.7 Famous programming errors



5/39

1. Week 7 Exercise #2: Set Containment

Consider two arrays, arrA and arrB, of int values, where theirsizes are sizeA and sizeB respectively.

Write a functionint isSubset(int arrA[], int sizeA, int arrB[], int sizeB)

to determine if the set of numbers in arrA is a subset of the set of

numbers in arrB. The function returns 1 ifarrA is a subset ofarrB, or 0 otherwise.

You may assume there are no duplicate numbers in each set.

Example: If arrA[ ] = {14, 5, 1, 9} and arrB[ ] = {2, 9, 3, 14, 5, 6, 1}

isSubset(arrA, 4, arrB, 7) returns 1 isSubset(arrA, 4, arrB, 6) returns 0

An incomplete program Week7_SetContainment.c is given.Complete the program. This is your take-home exercise.



6/39

2. Multidimensional Arrays (1/2)

In general, an array can have any number ofdimensions

Example of a 2-dimensional (2D) array:

// array with 3 rows, 5 columns

int a[3][5];a[0][0] = 2;a[2][4] = 9;a[1][0] = a[2][4] + 7;

Arrays are stored in row-major ordera[0][0] a[0][4] a[1][0] a[1][4] a[2][0] a[2][4]

row 0 row 1 row 2


0 1 2 3 4

01

2

2

9

16


7/39

2. Multidimensional Arrays (2/2)

Examples of applications


matrix[3][3]

1 2 3 30 31

Jan

Feb

Dec

:

32.1 31.8 31.9

32.3 32.4

32.6 32.6 33.0 0 0

31.8 32.3 30.9 31.6 32.2

Daily temperatures: temperatures[12][31]

Students lab marks: marks[4][5][3]

Ex1 Ex2 Ex3

Lab1

Lab2

Lab3

Lab4

52 50 45

57 60 63

52 59 66

33 42 37

Lab5 68 68 72

Suise

Ex1 Ex2 Ex3

Lab1

Lab2

Lab3

Lab4

79 75 66

90 83 77

81 73 79

58 64 52

Lab5 93 80 85

Ex1 Ex2 Ex3

Lab1

Lab2

Lab3

Lab4

59 68 60

0 0 0

67 71 75

38 52 35

Lab5 78 86 82

Emily

ZassJerna

Ex1 Ex2 Ex3

Lab1

Lab2

Lab3

Lab4

76 80 62

60 72 48

76 80 62

60 72 48

Lab5 58 79 73


8/39

2. Multidimensional Array Initializers

// nesting one-dimensional initializersint a[3][5] = { {4, 2, 1, 0, 0},

{8, 3, 3, 1, 6},{0, 0 ,0, 0, 0} };

// the first dimension can be unspecified

int b[][5] = { {4, 2, 1, 0, 0},{8, 3, 3, 1, 6},{0, 0, 0, 0, 0} };

// initializer with implicit zero valuesint d[3][5] = { {4, 2, 1},

{8, 3, 3, 1, 6} };


What happens tothe uninitializedelements?


9/39

2. Multidimensional Array: Example#include

#define N 5 // number of columns in arrayintsumArray(int[][N], int); // function prototype

int main(void) {int foo[][N] = { {3,7,1}, {2,1}, {4,6,2} };printf("sum is %d\n", sumArray(foo, 3));printf("sum is %d\n", sumArray(foo, 2));return 0;

}

// To sum all elements in arrintsumArray(intarr[][N], int rows) {

int i, j, total = 0;

for (i = 0; i < rows; i++) {for (j = 0; j < N; j++) {

total += arr[i][j];}

}return total;

}

Week8_2DArray.c

Second dimension must bespecified; first dimension isnot required.



10/39

3. Matrices


A two-dimensional array where all rows have the samelength is sometimes known as a matrix because itresembles that mathematical concept.

A matrixA with m rows and n columns is represented

mathematically in the following manner.a1 1 a1 2 a1 n

a2 1 a2 2 a2 n

am 1 am 2

am n

Note that in implementing the matrix as an array in C,the row number and column numberstart at 0 insteadof 1.


11/39

3. Demo #1: Matrix Addition (1/2)


To add two matrices, both must have the same numberof rows, and the same number of columns.

To compute C = A + B, where A, B, C are matrices

ci,j = ai,j+ bi,j

Example on 33 matrices:

021

122

020

120

012

001

101

110

021


12/39

3. Demo #1: Matrix Addition (2/2)


// To sum mtxA and mtxB to obtain mtxCvoidsumMatrix(float mtxA[][MAX_COL], float mtxB[][MAX_COL],

float mtxC[][MAX_COL], int row_size, int col_size){

int row, col;

for (row=0; row


13/39

3. Exercise #1: Matrix Multiplication (1/2)


To multiply two matrices A and B, the number ofcolumns in A must be the same as the number of rowsin B.

The resulting matrix has same number of rows as A and

number of columns as B. For example, multiplying a 45 matrix with a 53 matrix

gives a 43 matrix.


14/39

3. Exercise #1: Matrix Multiplication (2/2)


To compute C = A B, where A, B, C are matricesci,j= (ai,1 b1,j) + (ai,2 b2,j) + . . . + (ai,n bn,j)

ci,jis sum of terms produced by multiplying the elements of Asrow iwith Bs columnj.

Example on 33 matrices:

121

132

023

120

012

001

101

110

021

Complete the prodMatrix() function in Week8_matrixOps.c


15/39

4. Maze Problem (1/2)

Lets consider a maze that is represented by a two-dimensional NNinteger array.

The value of each array element is either0(representing a wall) or1 (representing a cell).

The starting and exit points in the maze are specifiedby the cells maze[0][0]and maze[N-1][N-1]respectively.

A path is represented by a single-dimensional

character array with four possible element valuesrepresenting the move directions: N (for north), S(for south), E (for east), and W (for west). Each pathis defined with respect to the starting cellmaze[0][0].



16/39

4. Maze Problem (2/2)

Example (N= 6)

0 1 2 3 4 5

0

1

2

3

4

5



17/39

4. Exercise #2 (take-home): Valid Path

A path in a maze is defined to be valid if the path iswithin the maze and does not visit any wall.

Examples:

Valid path: E, E, S, N

Invalid path: S, W

Invalid path: E, E, E, S

Write a functionint isValid (int maze[][N], char path[], int pathlen)

that takes as inputs a NNmaze and a path of lengthpathlen, and returns 1 ifpath is valid in maze; elsereturns 0.

This is a take-home exercise. Try it out before you referto the given incomplete program Week8_IsValid.c.

0 1 2 3 4 5

0

1

2

3

4

5



18/39

5. Testing and Debugging (1/7)


Where does the term debugging come from? Very early computers used mechanical relays for switching currents.

However, most likely the termbug existed beforethe moth-in-a-relaystory.

Why does a program core dump?

Early computers usedtiny magneticcores (rings) asmemory cells to hold0 and 1 values.

Electro-magnet

Moveable armature

Space(Bugs may getstuck here!)

Notebook with moth fromMark II computer:


19/39

Testing To determine if a code contains errors.

Debugging To locate the error(s) and fix it (them).

Documentation To improve maintainability of the code.

Include sensible comments, good coding style and clear logic.


Testing

Yes

Error?

Debug



20/39



Philosophical notes on program design, debugging andtesting

A good design is importantA good design results in a high quality program. A low qualitydesign cannot be debugged into high quality.

Dont optimize for speed too earlyIt is possible to make a correct program run faster.It is much more difficult to make a fast (but wrong) program runcorrectly.


21/39



A program should be tested with various input values to makesure that it performs correctly across all inputs.

A program should make as few assumptions about the input aspossible.

E.g.: Your program assumes that the user will type a number. But

she types a string crash! However, in CS1010 we assume that input follows the specification.

We do this to focus on the basics first. Writing robust programs isnot trivial.

Many of todays methods to hack into computers work by feeding

programs with unexpected inputs. This results in crashes, bufferoverflows, etc.


22/39


By user/programmer: Run program by hand multiple times.

By test program: Write a little test program that runs the program to be

tested with different inputs.

By test environments: Large-scale test suites that generate test cases, run them,

compare the expected output and provide a pass/failassessment.

E.g.: Mozillas Tinderbox


How to test?


23/39

Manual walkthroughs Tracing with pencil-and-paper.

Verbal walkthroughs.

printf() statements Easy to add

Provide information:

Which functions have been called

The value of parameters

The order in which functions have been called

The values of local variables and fields at strategic points Disadvantages

Not practical to add print statements in every function

Too many print statements lead to information overload

Removal of printf statements tedious




24/39

Tips and Techniques Start off with a working algorithm

Incremental coding/test early/fix bugs as you find them

Simplify the problem

Explain the bug to someone else Recognize common errors (such as using = instead of ==,

wrong addition of semi-colon, infinite loop, etc.)

Recompile everything (referring to a suite of programs)

Test boundaries

Eg: For primality test, did you test it with 1? 2?

Test exceptional conditions

Take a break!




25/39

Example on writing printf() statements


5.1 Demo #2: using printf( )

. . .

// compute weight of a single washerrim_area = circle_area(d2) - circle_area(d1);printf("Rim area = %lf\n", rim_area); // for checking

unit_weight = rim_area * thickness * density;

printf("Unit weight = %lf\n", unit_weight); // for checking

// compute weight of a batch of washerstotal_weight = unit_weight * qty;

// outputprintf("\nThe total weight of the batch of %dwashers is %.2f grams\n",

qty, total_weight);

. . .

Week8_WasherWeight_Printf.c


26/39

5.2 Debugging with assert( )

When writing large programs, it is often useful to know that acondition or set of conditions is true.

Such a statement is known as an assertion.

The C language provides an header file andcorresponding assert macro that the programmer can use to

make assertions.

If an assertion fails, the assert macro prints a diagnosticmessage describing the condition that should have been true butwas not, and then it kills the program.


#include

int my_function(int a, double b){

...assert(a = 17.1);

...}


27/39

5.2 Demo #3: Using assert( )


$ a.outAssertion failed: a = 17.1, file Week8_AssertDemo.c, line 11Abort (core dumped)$

#include #include

int main (void){

int a;double b;

a = 6;b = 18.0;assert(a = 17.1);

printf("The value of a is %dand the value of b is %lf\n", a, b);return 0;

}

Week8_AssertDemo.c


28/39

5.3 Exercise #3: Using assert( )


Caution: Do NOT use assignment statement insideassert().

Take our familiarWeek3_WashersWeight.c program andadd an assertion to make sure that the outer diameter is

larger than the inner diameter. Call it Week8_WashersWeight_Assert.c


29/39

Test your programs with your own data Do not rely on CodeCrunch to test your programs!

We discussed this in week 4. There is an error in thiscode:


// To find the maximum among 3 integer// values in variables num1, num2, num3.int max = 0;if (num1 > num2 && num1 > num3)max = num1;if (num2 > num1 && num2 > num3)max = num2;

if (num3 > num1 && num3 > num2)max = num3;

It works fine if it is tested on some sets of data: ,, , etc. and the program works for all these.

What is missing in his testing?

5.4 Testing Thoroughly (1/3)


30/39

In testing your programs thoroughly, do not forgetabout boundary or special cases! These are the cases where the program may give the wrong

answer a common error

In the Primality Test problem (checking if an integer isa prime), what are the boundary cases?




31/39

It is also important to test all the paths that yourprogram control flow can take Design test data to check all paths

Example


if (x != 3) {

y = 5;}else {z = z - x;

}

if (z > 1) {z = z / x;}else {z = 0;

}

if (x != 3)

y = 5 z = z - x

if (z > 1)

z = z / x z = 0

A

B

C

D

E

F

G

H

Test data:

to test path

A, B, G, H; to test pathE, F, C, D;

etc.



32/39

Debugger usually provides the following Stepping

Breakpoint

Watches (inspecting variables)

We will illustrate these with Gnu debuggergdb.

See videohttp://www.youtube.com/watch?v=Z6zMxp6r4mc


5.5 Using a Debugger
http://www.youtube.com/watch?v=Z6zMxp6r4mchttp://www.youtube.com/watch?v=Z6zMxp6r4mc


33/39

5.5 Using the gdb Debugger (Demo #4)

Step 1: add theg option when compiling and linking: gcc g Week8_WashersWeight_Printf.c

Step 2: start the gdb with your program: gdb a.out

Step 3: use the different gdb commands to step throughyour program:

Break at a specific function: break function_name

Start with: break main

Run program: run (orr)

Step to the next line of code: step (ors)

List source code: list(orl)

Examine the value of a variable:print variable_name

Quit the debugger: quit


Explore other gdb commands on your own!


34/39

5.6 Find the Bug

Why does this program produce the wrong result?


double half(double in_val)

{ return (in_val/2.0);}

#include

int main(void){

int val = 10;double result = 0.0;

result = half(val);printf("Half of %dis%lf\n", val, result);return 0;

}

Week8_Half.c

Week8_HalfMain.c

$ gcc Week8_HalfMain.c Week8_Half.c$ a.out$ Half of 10 is 10.000000

Compile it with

Wall option.


35/39

5.7 Famous Programming Errors (1/2)

Mariner Bugs Out (1962)

Cost: $18.5 million

Disaster: The Mariner 1 rocket with a space probe headedfor Venus diverted from its intended flight path shortly afterlaunch. Mission Control destroyed the rocket 293 secondsafter liftoff.

Cause: A programmer incorrectly transcribed a handwritten

formula into computer code, missing a single superscriptbar. Without the smoothing function indicated by the bar, thesoftware treated normal variations of velocity as if they wereserious, causing faulty corrections that sent the rocket offcourse.



36/39

5.7 Famous Programming Errors (1/2)

Mars Climate Crasher (1998)

Cost: $125 million

Disaster: After a 286-day journey from Earth, the MarsClimate Orbiter fired its engines to push into orbit aroundMars. The engines fired, but the spacecraft fell too far intothe planets atmosphere, likely causing it to crash on Mars.

Cause: The software that controlled the Orbiter thrustersused imperial units (pounds of force), rather than metric units(Newtons) as specified by NASA.



37/39

Summary for Today

Todays most important lessons Multidimensional arrays

Testing and Debugging

Using printf() to trace your program Using assert()

Using gdb debugger



38/39

Announcements/Things-to-do

Revise Chapter 6 Numeric Arrays

Term Test this Saturday

8 October 2011, Saturday

See module website for details

Next weeks lecture

Recursion: Lesson 8.10



39/39

End of File

Week8.pptx

Documents