User Input Validation

1
User Input Validation

• CTEC1335/2008F

2
Why We Need To Validate Input

• Incorrect or invalid input can cause programs to
  produce incorrect or invalid outputs, behave
  erratically, or fail.
• We want user-friendly software, so that we can
  catch user input mistakes and give the user
  another chance
• DONT TRUST USER INPUT

3
Limitations of gtgt

• The insertion operators only work for a specific
  type of input (char, int, double) and tend to
  ignore invalid input
• Sometimes this behaviour makes it difficult to
  know whether the user has entered invalid data or
  just an invalid value.

4
Example

• int i 0
• cout ltlt Enter a number
• cin gtgt i
• If the user types in abc, the cin gtgt i statement
  will ignore it and i will be unchanged. But what
  if the user types in zero?

5
Strings To The Rescue

• A string is a sequence of char values, usually
  terminated with a zero byte.
• In C, a string is either an array of chars or a
  pointer to an array of chars.
• char name 20
• char line new char 80
• C strings are manipulated by functions contained
  in the cstring header file include ltcstringgt

6
C Strings

• In C, the ANSI/ISO Standard contains a string
  class, which contains several methods (functions)
  to manipulate string objectsinclude ltstringgt
• string name
• string p_title new string(
• C for Engineers and Scientists )

7
String Methods

• See Table 7.4, pp. 399-401.int length( ) //
  get string length
• // extract a substringstring substr( int
  startindex, int numchars ) // extract a single
  character
• char operator ( int index )
• // join two strings into a new onestring
  operator( const string s1, const string s2
  ) // append a string (on to the end)string
  operator( const string s )

8
String Input

• Use the getline( ) functionistream getline(
  istream stream, string line, char
  terminator \n )

9
Using getline( )

• This function, by default, reads all characters
  that the user types in, until Enter is pressed,
  and copies them, excluding the newline character
  generated by the Enter key, into the string
  object.
• You need to declare the string object (variable)
  for line beforehand, and use cin as the stream.

10
Using getline( )

• Examplestring name cout ltlt Enter your name
  getline( cin, name )
• Note that the terminator argument is not
  required, because it has a default value (a
  newline character \n)

11
Why not just use gtgt?

• The extraction operator gtgt is configured to read
  string data.
• For examplestring name cout ltlt Enter your
  name cin gtgt name

12
Limitations of gtgt (again)

• However, the gtgt looks for the first whitespace
  character in the input, either a space, tab or
  newline character, and stops reading there.
• So if the user types in his or her first and last
  names (separated by a space), only the first name
  is read into the string!
• You could get around this with an extra read and
  two string variables, but what if the user has
  three names?

13
Entering Numbers

• In order to do mathematical calculations in C,
  data has to be in the form of int or double
  variables.
• However, there is no getline( ) function for int
  or double variables, so we need a way to convert
  string data to numeric data.
• A few ways to go about this

14
C Standard Library Conversion Routines (1)

• The C Standard Library, which is part of the
  ANSI/ISO C Standard, has two functions to
  convert string data to numeric datainclude
  ltcstdlibgtint atoi( const char ) double atof(
  const char )

15
C Standard Library Conversion Routines (2)

• The C Standard I/O Library has another routine
  that can convert string data to numeric
  datainclude ltcstdiogtint sscanf( const
  char s, const char format, ... )

16
Limitations of Built-In Functions

• The three C Standard Library functions may have
  difficulty differentiating invalid input from
  invalid values.
• For example, both atoi( ) and atof( ) return 0
  (0.0) when invalid string data is encounted,
  i.e., non digit characters.
• There is no easy way to tell when an actual zero
  value has been converted!

17
Parsing

• Parsing is a fancy computer science buzzword
  that means making sense of string data.
• Using string methods, we can go through a string
  character by character and determine exactly what
  was entered.
• The cost of using parsing is that we typically
  have to write all or part of the parsing code

18
Writing A Parser

• To correctly write a parser, you need to
  determine what characters constitute valid input
  (i.e., the input format or grammar.)
• Second, you extract parts of the string called
  tokens and determine if the each token is
  acceptable where it is found. This process is
  called scanning or tokenizing.
• Then, you convert the extracted tokens into a
  return value that the rest of the program can
  process. This may involve eliminating tokens
  like whitespace or doing some calculations to
  generate a numeric value.

19
Parsers (continued)

• The parser will return both the value and a
  Boolean (bool) value indicating whether the input
  was valid or not.
• This way, the program can tell whether a zero
  return value actually means a zero was input.
• Often, the use of a Finite State Machine (FSM) is
  useful in writing a parser

20
Upcoming topics in Unit 2

• Using constants
• Exception handling
• Finite State Machines
• Writing a parser FSM
• File I/O (both text and binary)