Fundamentals of C

1
Fundamentals of C

• Program Structure
• Constants
• Variables
• Assignment Statements
• Standard Output
• Standard Input
• Math Functions
• Character Functions
• System Limitations
• Conditional Expressions
• Selection Statements
• Loop Structures

2
Simple Program Structure

• / ANSI-compliant comments /
• // Comment to end of line - common but not
  ANSI-compliant prior to C99
• include ltstdio.hgt // printf() Standard
  header File
• include myinclude.h // custom header file
• // Constant declaration - text replacement to
  comment of end of line
• define PI (3.14159)
• int main(void) // default declaration for main()
• // Variable Declarations
• int m,n,p12
• double area, r 9.0, z 9.2
• // executable code here
• area PI r r
• return(0) // main() declared as type int -
  must return int.

3
Constants

• Literal constants
• Values literally appearing in the code.
• Format determines how compiler interprets them.
• Interpreted as an integer value
• 42 - Interpreted as an integer value.
• 0x4F - Hexadecimal (0x prefix) - interpreted as
  an integer.
• 042 - Octal (0 prefix - watch out!) - interpreted
  as an integer.
• 84293L - Long integer - prevents compiler from
  truncating to int.
• Interpreted as a floating point value
• 42. - Decimal point is sufficient.
• 42.0 - But easy to miss - tacking on a zero is
  better.
• 1e9 - Scientific notation
• define statements
• Convenient way to define symbolic constants.
• Surrounding values with parentheses is good
  practice.

4
Variables

• A name associated with a memory location where a
  value can be stored.
• Can consist of letters, digits, underscore
  characters - thats it!
• Leading character cannot be a digit.
• Case sensitive - myvar, MYVAR, and MyVar are all
  different.
• Only first 32 characters are significant (more in
  newest ANSI Standard)
• Variable types
• Size and range system dependent - sizes below for
  Borland Turbo C/C V4.5
• integer types (can be prefixed with unsigned to
  double positive range)
• char - always one byte (but not always eight
  bits). One ASCII character.
• int - two bytes (-32768 to 32767)
• long (or long int) - four bytes (-2147483649 to
  2147483648)
• floating point types
• float - four bytes - 5 sig figs, exponent /- 38
  
• double - eight bytes - 15 sig figs, exponent /-
  308
• long double - ten bytes - 19 sig figs, exponent
  /- 4932

5
Operators

• Arithmetic Operators
• Relational Operators
• Logical Operators
• Assignment Operators
• Precedence
• Associativity
• Type Casts

6
Arithmetic Operators

• (Multiplication)
• / (Division)
• Integer Division if both operands are integers.
• 17 / 5 3, not 3.4 - common, common error!
• (Modulo)
• Remainder after Integer Division (175 2)
• (Addition)
• - (Subtraction)

7
Relational Operators

• Evaluate to 0 if FALSE and 1 if TRUE
• (equality) (not common, common mistake)
• ! (inequality)
• gt (greater than)
• lt (less than)
• gt (greater than or equal to)
• lt (less than or equal to)
• Complementary Pairs (always opposite result)
• (, !), (gt, lt), (lt, gt)

8
Logical Operators

• Evaluate to 0 if FALSE and 1 if TRUE
• ! (negation) (unary operator - operand to its
  right)
• (logical AND) (TRUE only if both operands are
  TRUE)
• (logical OR) (TRUE if either operand is TRUE)
• Logical Evaluation of Operands
• Operand is FALSE if it is exactly equal to zero.
• Operand is TRUE if it is ANYTHING other than
  zero.
• Guaranteed Short Circuiting
• If result is known after evaluating left operand,
  right will not be evaluated.
• Can by very useful in preventing errors.
• Care must be exercised if there are side-effects
  in right operand.
• (0 ! x) ( (y/x) gt 1)
• Useful cannot result in division by zero.
• (y gt x) (j lt k)
• Dangerous Will not increment k if x lt y!

9
Assignment Operators

• In expressions, evaluate to the value assigned.
• x y 9 (k 12 (x j 3)) // is valid
  code!
• Abbreviated Assignment Operators
• , /, , , -,
• k x is the same as k k x
• Increment/Decrement Operators
• k, k-- (post-increment/decrement)
• evaluates to original value of k in expressions.
• k, --k (pre-increment/decrement)
• evaluates to new value of k in expressions.

10
Precedence of Operators

• Largely follow familiar rules
• (, /, ) before (, -)
• Use parentheses liberally
• Easy to get trapped in subtleties.
• Prevents the compiler from making your decisions
  for you!
• Prevents common mistakes such as
• x ab / c instead of x (ab)/c

11
Associativity of Operators

• Determines order of evaluation when all else is
  equal.
• Left Associativity
• x y z w gt x ( (y z) w)
• Right Associativity
• x y z w gt x (y (z w))
• Use parens liberally - dont let the compiler
  decide!

12
Type Cast Operators

• Used to convert the representation of a value
  within an expression.
• Has NO effect on the variable that is cast.
• Operates on the value immediately to the right of
  the cast.
• Use name of type in parens
• x (double) j / (double) k will force floating
  point division.
• x (double) (j/k) will still perform integer
  division.
• Compiler will do some casting automatically.
• Only performed at points where necessary.
• k 12 (14/20) (7.5/3)
• (7.5/3) gt (7.5/3.0) 2.5 (integer promoted to
  floating point)
• (14/20) 0 (no cast necessary, integer division
  used)
• (12 0 2.5) gt (12.0 0.0 2.5) 14.5
  (promotion)
• k 14.5 gt k 14 (demotion if k is an int)

13
Standard Output

• include ltstdio.hgt
• printf(format string, arg1, arg2, .... argN)
  // to the screen
• fprintf(file_ptr, format string, arg1, arg2,
  .... argN) // to a file
• sprintf(str_ptr, format string, arg1, arg2,
  .... argN) // to a string
• each returns the number of characters printed.
• Format String
• All but two characters printed verbatim.
• character used to format and output the next
  argument.
• \ character used to print characters that cant
  be typed directly.
• Must be at least as many arguments as
  specifiers.

14
- printf() conversion specifiers

• Format flagswidth.precisionmodifiertype_
  character
• All except type_character are optional.
• Conversion specifier is everything from sign to
  the first type_character.
• flags - controls justification, leading spaces,
  sign, etc.
• -, , , , 0
• width - sets the minimum width of the field -
  may be longer.
• .precision - sets the number of digits
  following the decimal point.
• Usually used for floating points, but also
  affects character and integer types as well.
• modifier - combines with type_character to
  determine argument type.
• type_character - the basic type of that argument.
• c, d, e, E, f, g, G, i, o, p, s, u, x, X,
• In general, look up what you need.
• You will tend to remember the forms you use most
  often.

15
\ - printf() escape sequences

• Format \(character) or \(number)
• If a character follows the \, the action
  indicated by the character is performed.
• \n - newline
• \r - return w/o line feed.
• \ - print double quote
• \ - print single quote (aka apostrophe)
• \a - sound bell
• \b - backspace
• \\ - print backslash
• \? - question mark
• If a number follows the \, the ASCII character of
  the octal number is printed.

16
Standard Input

• include ltstdio.hgt
• scanf(format string, arg1, arg2, .... argN) //
  from the keyboard
• scanf(file_ptr, format string, arg1, arg2, ....
  argN) // from a file
• sscanf(str_ptr, format string, arg1, arg2, ....
  argN) // from a string
• each returns the number of successful
  conversions.
• Format String
• Literal characters in format string can be used
  to skip characters.
• conversion specifiers similar (but not
  identical) to printf().
• Arguments need to be memory locations where
  values will be stored.
• Big Time Caveat
• If input does not adequately match format,
  results can be VERY unpredictable.
• Many programmers avoid the use of scanf() at
  nearly any cost.
• Use of fscanf() for file I/O is generally safe
  provided the file format is adequately
  constrained.

17
Common mistakes with scanf()

• Passing values instead of memory locations.
• The scanf() function read values and store them
  at the memory locations you supply.
• k 10
• scanf(i, k)
• Tells scanf() to format the value as an integer
  and store it at location 10.
• But the memory location for variable k is almost
  certainly not 10.
• The address operator, , returns the memory
  location of the specified variable.
• scanf(i, k)
• Tells scanf() to format the value as an integer
  and store it at the memory address used for
  variable k.
• Using lf for doubles.
• printf() uses lf for both floats and doubles
  because the compiler promotes all arguments of
  type float to double.
• scanf() cannot due this. It must know which type
  the variable is so that the number and format of
  the bytes stored is correct.

18
Math functions

• include ltmath.hgt
• Approximately 30 function, most with a companion
  function for long ints.
• trigonometric
• sin(), cos(), tan() and their inverses.
• sinh(), cosh(), tanh() but NOT their inverses
  (use trig relations)
• atan2(y,x) - two argument version of atan()
  allowing four quadrant answer.
• arguments and/or return values in radians - not
  degrees.
• exponential/logorithmic
• base e exp(), log()
• base 10 pow10(), log10
• other sqrt(), pow(x, y) - returns xy
• absolute value
• abs() - for integers only! common mistake
• fabs() - for floating point values

19
ASCII Codes

• American Standard Code for Information
  Interchange - 7 bits (128 values)
• Used to represent characters and certain codes
  used to control their display.
• Originally developed for Teletypewriter
  applications.
• Many nonstandard extensions exist to extend the
  code to 8 (or more) bits.
• Codes lt32 are control codes, some of which are no
  longer used.
• Codes gt32 are printable characters (except 127
  which is the DELETE control code.
• ASCII 32 is the space character.
• In C, a character surrounded with single quotes
  is that characters ASCII code.
• 32
• 0 48
• A 65
• a 97 A 32 A

20
Character functions

• include ltctype.hgt
• Character test functions
• Return a logical value (0 or not-0) based on the
  properties of the character code.
• isdigit(k) returns 0 unless 0 lt k lt 9
• isupper(k) returns 0 unless A lt k lt Z
• Character manipulation functions
• Return a different code for some characters,
  otherwise return the value passed.
• tolower(k) returns k32 if isupper(k) is TRUE
  otherwise returns k.

21
System Limitations

• include ltlimits.hgt
• Symbolic constants that define the limits of
  integer representations.
• Examples
• INT_MAX Maximum value of an int
• INT_MIN Minimum value of an int
• ULONG_MAX Maximum value of an unsigned long int
• include ltfloat.hgt
• Symbolic constants that define the limits of
  floating point representations.
• Examples
• DBL_MAX Largest representable value for a
  double.
• DBL_MIN Smallest positive representable value
  for a double.
• DBL_EPSILON Smallest value x such that 1x is
  not equal to 1.

22
Selection Statements

• Selectively choose one path of execution.
• Based on the evaluation of a test.
• Test outcome is either TRUE or FALSE.
• FALSE if expression evaluates to ZERO.
• TRUE if expression evaluates to ANYTHING else.
• Three varieties of selection statements.
• if()/else
• switch()
• ()? (conditional or ternary operator)

23
Conditional Expressions

• (test)
• May be ANY expression that evaluates to a value.
• logical operator (!, , )
• relational operator (,!),(gt, lt), (lt,gt)
• assignment operator (equal to value assigned)
• single variable (or constant)
• function return value
• Test outcome is either TRUE or FALSE.
• FALSE if expression evaluates to ZERO - exactly.
• TRUE if expression evaluates to ANYTHING else.

24
if() and if()/else statements

• Syntax
• if(test)
• if_code
• else
• else_code

25
()? - conditional or ternary operator

• Syntax
• x (test)? if_exprelse_expr
• Shorthand version of if()/else

26
switch() statement

• Syntax
• switch(int_expr)
• case int_const1 code1
• break
• case int_const2 code2
• case int_const3 code3
• break
• default code4
• Compact way of writing certain types of complex
  but common if()/else blocks.

(int_expr) must evaluate to an integer value at
runtime. (int_constN) must evaluate to a unique
integer constant at compile time. execution
jumps to case where (int_constN int_expr) is
TRUE. break terminates switch()
execution. default case Optional. Executes only
if NO other case executes.
27
Testing for floating point equality

• if(x y) // BAD!!!
• Only true if EXACTLY equal. Because a great deal
  of care went into how floating point values are
  represented, you can frequently get away with it
  - but unless you REALLY know what is going on and
  what is preventing disaster from visiting you,
  you are just being lucky!
• DEFINE EPSILON (0.0001)
• if(EPSILON gt fabs(x-y)) // GOOD!!!
• Looking for the difference between two values to
  be less than some small amount. Will work most of
  the time but EPSILON must be matched to the size
  of the values to be compared and the range of
  values cant be too great.
• DEFINE EPSILON (0.0001)
• if(fabs(EPSILONy) gt fabs(x-y)) // EVEN
  BETTER!!!
• Looking for two values to vary by less than a
  certain percentage. Comparison written so as to
  avoid potential division by zero during test.

28
Loop Structures

• Special case of Selection Statement
• One branch eventually leads back to the original
  selection statement.
• Permits a block of code to be executed repeatedly
  as long as some test condition is satisfied.
• C provides three different looping structures
• while(), do/while(), for()
• Only one is needed and any one is sufficient.
• Different structures are better matches for the
  logic.
• Using the proper one aides the programmer and
  anyone else reading the code.
• The compiler doesnt care and will often
  implement the code identically regardless of
  which structure is used.

29
while() loop

• Syntax
• ini_code // not part of loop
• while(test_expr)
• loop_code
• increment_code
• next_code // not part of loop
• Features
• loop does not execute at all if test fails the
  first time.

30
do/while() loop

• Syntax
• ini_code // not part of loop
• do
• loop_code
• increment_code
• while(test_expr)
• next_code // not part of loop
• Features
• loop will always execute at least once, even if
  test fails the first time.

31
for() loop

• Syntax
• for(ini_code test_expr inc_code)
• loop_code
• next_code // not part of loop
• Features
• Just a while() loop with the initialization and
  increment code formally incorporated into the
  syntax.
• Can makes a cleaner divide between the loop logic
  and the housekeeping logic.