TESTING,

DEBUGGING,
EXCEPTIONS,
(download slides and .py files and follow
along!)
ASSERTIONS
6.0001 LECTURE 7

6.0001 LECTURE 7 1
 WE AIM FOR HIGH
QUALITY – AN
ANALOGY
You are making soup butWITH SOUP
bugs keep falling in from the
ceiling. What do you do?
 check soup for bugs
• testing
 keep lid closed
• defensive

programmin
g
 clean kitchen
• eliminate
source of bugs 6.0001 LECTURE 7 2
 DEFENSIVE PROGRAMMING
• Write specifications for functions
• Modularize programs
• Check conditions on inputs/outputs (assertions)

TESTING/VALIDATION DEBUGGING
• Compare • Study events
input/output pairs to leading up to an error
specification • “Why is it not working?”
• “It’s not working!” • “How can I fix my
• “How can I break my program?”
program?”

6.0001 LECTURE 7 3
 SET YOURSELF UP FOR
EASY TESTING AND
DEBUGGING
from the start, design code to ease this part
break program up into modules that can be tested
and debugged individually
 document constraints on modules
• what do you expect the input to be?
• what do you expect the output to
be?
 document assumptions behind
code design

6.0001 LECTURE 7 4
WHEN ARE YOU READY
TO TEST?
 ensure code runs
• remove syntax errors
• remove static semantic errors
• Python interpreter can usually find these for you
 have a set of expected results
• an input set
• for each input, the expected output

6.0001 LECTURE 7 5
 CLASSES OF
TESTS
Unit testing
• validate each piece of program
• testing each function separately
 Regression testing
• add test for bugs as you find them
• catch reintroduced errors that were previously fixed
 Integration testing
• does overall program work?
• tend to rush to do this

6.0001 LECTURE 7 6
 TESTING
APPROACHES
intuition about natural boundaries to the problem
def is_bigger(x, y):
""" Assumes x and y are ints
Returns True if y is less than x, else False """
• can you come up with some natural partitions?
 if no natural partitions, might do random testing
• probability that code is correct increases with more tests
• better options below
 black box testing
• explore paths through specification
 glass box testing
• explore paths through code
6.0001 LECTURE 7 7
 BLACK BOX
TESTING
def sqrt(x, eps):
""" Assumes x, eps floats, x >= 0, eps > 0
Returns res such that x-eps <= res*res <= x+eps """

 designed without looking at the code

can be done by someone other than the implementer to
avoid some implementer biases
 testing can be reused if implementation changes
 paths through specification
• build test cases in different natural space partitions
• also consider boundary conditions (empty lists,
singleton list, large numbers, small numbers)
6.0001 LECTURE 7 8
 BLACK BOX TESTING
def sqrt(x, eps):
""" Assumes x, eps floats, x >= 0, eps > 0
Returns res such that x-eps <= res*res <= x+eps """
CASE x eps
boundary 0 0.0001
perfect square 25 0.0001
less than 1 0.05 0.0001
irrational square root 2 0.0001
extremes 2 1.0/2.0**64.0
extremes 1.0/2.0**64.0 1.0/2.0**64.0
extremes 2.0**64.0 1.0/2.0**64.0
extremes 1.0/2.0**64.0 2.0**64.0
extremes 2.0**64.0 2.0**64.0
6.0001 LECTURE 7 9
 GLASS BOX
TESTING
use code directly to guide design of test cases
called path-complete if every potential path through
code is tested at least once
 what are some drawbacks of this type of testing?
• can go through loops arbitrarily many times
• missing paths
 guidelines
• branches
• for loops
• while
loops
6.0001 LECTURE 7 10
GLASS BOX TESTING
def abs(x):
""" Assumes x is an int
Returns x if x>=0 and –x otherwise """
if x < -1:
return –x
else:
retur
n x
 a path-complete
test suite could
miss a bug
 path-complete
test suite: 2 and -
2 6.0001 LECTURE 7 11
 DEBUGGI
NG
steep learning curve
 goal is to have a bug-free program
 tools
• built in to IDLE and Anaconda
• Python Tutor
• print statement
• use your brain, be systematic in your hunt

6.0001 LECTURE 7 12
 PRINT
STATEMENTS
good way to test hypothesis
 when to print
• enter function
• parameters
• function results
 use bisection method
• put print halfway in code
• decide where bug may be depending on values

6.0001 LECTURE 7 13
 DEBUGGING
STEPS
study program code
• don’t ask what is wrong
• ask how did I get the unexpected result
• is it part of a family?

 scientific method
• study available data
• form hypothesis
• repeatable experiments
• pick simplest input to test with

6.0001 LECTURE 7 14
 ERROR MESSAGES –
 trying to access beyond the limits of a list
EASY
test = [1,2,3]
IndexError
then test[4]
 trying to convert an inappropriate type


int(test)  TypeError
 referencing a non-existent variable
a  NameError
 mixing data types without appropriate coercion
'3'/4  TypeError
 forgetting to close parenthesis, quotation, etc.
a = len([1,2,3]
print(a) 
SyntaxError
6.0001 LECTURE 7 15
 LOGIC ERRORS -
HARD
think before writing new code
 draw pictures, take a break
 explain the code to
• someone else
• a rubber ducky

6.0001 LECTURE 7 16
 DON’ D
T
• Write entire program
• Test entire program
•
• O
Write a function
Test the function, debug the function
• Debug entire program • Write a function
• Test the function, debug the function
• *** Do integration testing ***

• Change code • Backup code

• Remember where bug was • Change code
• Test code • Write down potential bug in a
• Forget where bug was or what change comment
you made • Test code
• Panic • Compare new version with old
version
6.0001 LECTURE 7 17
 EXCEPTIONS AND
ASSERTIONS
 what happens when procedure execution hits an
unexpected condition?
 get an exception… to what was expected
• trying to access beyond list limits
test = [1,7,4]
test[4]  IndexError
• trying to convert an inappropriate type
int(test)  TypeError
• referencing a non-existing variable
a  NameError
• mixing data types without coercion
'a'/4  TypeError

6.0001 LECTURE 7 18
 OTHER TYPES OF
EXCEPTIONS
already seen common error types:
• SyntaxError: Python can’t parse program
• NameError: local or global name not found
• AttributeError: attribute reference fails
• TypeError: operand doesn’t have correct type
• ValueError: operand type okay, but value is illegal
• IOError: IO system reports malfunction (e.g. file not
found)

6.0001 LECTURE 7 19
 DEALING WITH
EXCEPTIONS
Python code can provide handlers for exceptions
try:
a = int(input("Tell me one number:"))
b = int(input("Tell me another number:"))
print(a/b)
except:
print("Bug in user input.")

exceptions raised by any statement in body of try are

handled by the except statement and execution
continues with the body of the except statement
6.0001 LECTURE 7 20
HANDLING SPECIFIC
EXCEPTIONS
have separate except clauses to deal with a
particular type of exception
try:
a = int(input("Tell me one number: "))
b = int(input("Tell me another number: "))
print("a/b = ", a/b)
print("a+b = ", a+b)
except ValueError:
print("Could not convert to a number.")
except ZeroDivisionError:
print("Can't divide by zero")
except:
print("Something went very wrong.")

6.0001 LECTURE 7 21
OTHER EXCEPTIONS
 else:
• body of this is executed when execution of associated
try body completes with no exceptions
 finally:
• body of this is always executed after try, else
and except clauses, even if they raised another
error or executed a break, continue or return
• useful for clean-up code that should be run no
matter what else happened (e.g. close a file)

6.0001 LECTURE 7 22
WHAT TO DO WITH
EXCEPTIONS?
 what to do when encounter an error?
 fail silently:
• substitute default values or just continue
• bad idea! user gets no warning
 return an “error” value
• what value to choose?
• complicates code having to check for a special value
 stop execution, signal error condition
• in Python: raise an exception
raise Exception("descriptive string")

6.0001 LECTURE 7 23
EXCEPTIONS AS
CONTROL FLOW
don’t return special values when an error occurred
and then check whether ‘error value’ was returned
instead, raise an exception when unable to produce a
result consistent with function’s specification
raise <exceptionName>(<arguments>)
raise ValueError("something is wrong")

6.0001 LECTURE 7 24
 EXAMPLE: RAISING AN
EXCEPTION
def get_ratios(L1, L2):
""" Assumes: L1 and L2 are lists of equal length of numbers
Returns: a list containing L1[i]/L2[i] """
ratios = []
for index in range(len(L1)):
try:
ratios.append(L1[index]/L2[index])
except ZeroDivisionError:
ratios.append(float('nan')) #nan = not a number
except:
raise ValueError('get_ratios called with bad arg')
return ratios

6.0001 LECTURE 7 25
 EXAMPLE OF
EXCEPTIONS
assume we are given a class list for a subject: each
entry is a list of two parts
• a list of first and last name for a student
• a list of grades on assignments
test_grades = [[['peter', 'parker'], [80.0, 70.0,
85.0]],
[['bruce', 'wayne'], [100.0, 80.0,
74.0]]]

create a new class list, with name, grades, and an

average
[[['peter', 'parker'], [80.0, 70.0, 85.0],
78.33333],
[['bruce', 'wayne'], [100.0, 80.0, 74.0],
6.0001 LECTURE 7 26
EXAMP
COD
LE
[[['peter', 'parker'], [80.0, 70.0, 85.0]],
[['bruce', 'wayne'], [100.0, 80.0, 74.0]]]

E
def get_stats(class_list):
new_stats = []
for elt in class_list:
new_stats.append([elt[0], elt[1], avg(elt[1])])
return new_stats

def avg(grades):
return sum(grades)/len(grades)

6.0001 LECTURE 7 27
ERROR IF NO GRADE FOR
A STUDENT
if one or more students don’t have any grades,
get an error
test_grades = [[['peter', 'parker'], [10.0,
5.0, 85.0]],
[['bruce', 'wayne'], [10.0,
8.0, 74.0]],
[['captain', 'america'], [8.0,10.0,96.0]],
[['deadpool'], []]]

 get ZeroDivisionError: float division by zero

because try to
return sum(grades)/len(grades)

6.0001 LECTURE 7 28
 OPTION 1: FLAG THE
ERROR BY PRINTING A
MESSAGE
decide to notify that something went wrong with a msg
def avg(grades):
try:
return sum(grades)/len(grades)
except ZeroDivisionError:
print('warning: no grades
data')
warning:
runningnoongrades
some test
datadata gives
[[['peter', 'parker'], [10.0, 5.0, 85.0], 15.41666666],
[['bruce', 'wayne'], [10.0, 8.0, 74.0], 13.83333334],
[['captain', 'america'], [8.0, 10.0, 96.0], 17.5],
[['deadpool'], [], None]]

6.0001 LECTURE 7 29
 OPTION 2: CHANGE THE
POLICY
 decide that a student with no grades gets a zero
def avg(grades):
try:
return sum(grades)/len(grades)
except ZeroDivisionError:
print('warning: no grades
data') return 0.0
 running on some test data gives
warning: no grades data
[[['peter', 'parker'], [10.0, 5.0, 85.0], 15.41666666],
[['bruce', 'wayne'], [10.0, 8.0, 74.0], 13.83333334],
[['captain', 'america'], [8.0, 10.0, 96.0], 17.5],
[['deadpool'], [], 0.0]]
6.0001 LECTURE 7 30
 ASSERTIO
NS
want to be sure that assumptions on state of
computation are as expected
 use an assert statement to raise an
AssertionError exception if
assumptions not met
 an example of good defensive programming

6.0001 LECTURE 7 31
 EXAMP
LE
def avg(grades):
assert len(grades) != 0, 'no grades data'
return sum(grades)/len(grades)

raises an AssertionError if it is given an empty list for

grades
 otherwise runs ok

6.0001 LECTURE 7 32
 ASSERTIONS AS
DEFENSIVE
PROGRAMMING
assertions don’t allow a programmer to control
response to unexpected conditions
ensure that execution halts whenever an expected
condition is not met
typically used to check inputs to functions, but can be
used anywhere
can be used to check outputs of a function to avoid
propagating bad values
 can make it easier to locate a source of a bug

6.0001 LECTURE 7 33
 WHERE TO USE
ASSERTIONS?
goal is to spot bugs as soon as introduced and make
clear where they happened
 use as a supplement to testing
 raise exceptions if users supplies bad data input
 use assertions to
• check types of arguments or values
• check that invariants on data structures are met
• check constraints on return values
• check for violations of constraints on procedure (e.g.
no duplicates in a list)

6.0001 LECTURE 7 34
MIT OpenCourseWare
https://ocw.mit.edu

6.0001 Introduction to Computer Science and Programming in Python

Fall 2016

For information about citing these materials or our Terms of Use, visit: https://ocw.mit.edu/terms.