C Language

C Introduction & History

Birth of C •C is very often referred as a “System Programming Language” because it is used for writing compilers, editors and even operating systems. C was developed and implemented on the UNIX operating system on the DEC PDP-11, by Dennis M. Ritchie. •C was evolved during 1971-73 from B language of Ken Thompson, which was evolved (in 196970) from BCPL language of Martin Richards. “Every good tree produces good fruit.”

Programming Languages Phylogeny Fortran (1954) Algol (1958) LISP (1957)

Scheme (1975)

CPL (1963), U Cambridge Combined Programming Language Simula (1967) BCPL (1967), MIT Basic Combined Programming Language B (1969), Bell Labs C (1970), Bell Labs C++ (1983), Bell Labs

Java (1995), Sun

Objective C

Strengths „

Low-level nature makes it very efficient

„

Language itself is very readable

„

„ „

Many compilers available for a wide range of machines and platforms Standardization improves portability of code Efficient in terms of coding effort needed for a wide variety of programming applications

Weakness „

Parts of syntax could be better „

if ( i = 1 ) „

„

is probably error, should be „

if ( i == 1 ) „

„

„

stmt ...

stmt ...

Low-level nature makes it easy to get into trouble _ infinite loops

Weakness „

„ „

Illegal memory accesses / tromping over memory Obfuscated code can be written main(n,i,a,m) „

{ „ „ „

„

}

while(i=++n)for(a=0;a
C Language „

No support for: „ „ „ „ „

Array bounds checking Null dereferences checking Data abstraction, inheritance Exceptions Automatic memory management

„

Program crashes (or worse) when something bad happens

„

Lots of syntactically legal programs have undefined behavior

So, why would anyone use C today? „

Legacy Code „

„

Simple to write compiler „

„

Programming embedded systems, often only have a C compiler

Performance „

„

Linux, most open source applications are in C

Typically 50x faster than interpreted Java

Smaller, simpler, lots of experience

What is Embedded C? „ „ „

„

„

„

Embedded C is not any new language It is not a standard It is just the Tools that make the language dependent and today’s embedded applications are being developed using C (A shortcut to ASM) As Applications are target dependent standard language will never meet all the expectations Hence non standard keywords and concepts are added as a support for developers After all, Embedded C is defined as ‘variant’ by different Venders

C and Processor Target

„

C more suited for RISC „

„

Writing to a device is just by addressing

C is more challenging in CISC „

Need to extend the non standard key words]

C Programming Language „

Developed to build Unix operating system

„

Main design considerations: „

„

„ „

„

Compiler size: needed to run on PDP-11 with 24KB of memory (Algol60 was too big to fit) Code size: needed to implement the whole OS and applications with little memory Performance Portability

Little (if any consideration): „

Security, robustness, maintainability

Assembly Language „

Computer processors each have their own built-in assembly language „

„ „

Limited range of commands Each command typically does very little „

„

E.g., Intel CPUs have their own language that differs from the language of Motorola PowerPC CPUs

Arithmetic commands, load/store from memory

Code not totally unreadable, but close

C versus Assembly „

Here’s some simple C code: x = y + 2; if ( (x > 0) || ( (y-x) <= z) ) x = y + z;

„

And now the same three lines of code in Intel x86 assembly language...

MOV MOV MOV ADD CMP JG

AX, y BX, y CX, Z AX, 2 AX, 0 SET_X

MOV SUB CMP JLE

DX, BX DX, AX DX, CX SET_X

JMP DONE SET_X: MOV AX, BX ADD AX, CX DONE:

MOV x, AX RET

The C Programming Language „

C is a higher-level language compared to assembly „ „

„

Much more human-readable Many fewer lines of code for same task

C is a lower-level language compared to others (like Java) „

Direct control over memory allocation and cleanup (as we will see)

Compiled vs. Interpreted „

For a program to run, the source code must be translated into the assembly language of the machine the program will run on „

„

Compiled language: the source code is translated once and the executable is run. (C, Pascal) Interpreted language: an interpreter runs each time the program is started, and it does the translation on-the-fly (Perl, Tcl, other scripting languages)

Procedural vs. Object-Oriented „

Procedural Programming „

„

„

A program viewed as a series of instructions to the computer Instructions are organized into functions and libraries of functions

Object-Oriented Programming „

„

A program viewed as a set of objects that interact with each other An object encapsulates functions with the data that the object’s functions operate on

Procedural vs. Object-Oriented „

„

OOP good for complex systems – nice to break down into small independent objects Procedural perhaps a bit more intuitive, esp. for beginners

„

C is a procedural language

„

Java and C++ are object-oriented languages

Procedural Programming „

„

Program is a set of sequential steps to be executed by the computer, one after another Not necessarily the same steps every time the program runs „ „ „

May want to skip steps under certain conditions May want to repeat steps under certain conditions May need to save some information to use later in the program

Elements of a Program „

Individual steps/commands to the computer „

„

Technique to organize code so it’s easy to debug, fix, and maintain „

„

Functions

Way to save information that may change each time the program runs „

„

Statements

Variables

Way to change what steps get executed „

Control flow

Declarations vs. Definitions „

„

Both functions and variables MUST be declared before they can be used Declaration includes just: „ „ „

„

Type Name Args (for functions)

Definition can come later! „

For functions: body of function can be added later

Important Rules „

„

All variables and functions must be declared in the function/file where they are going to be used BEFORE they are used All variables must be initialized to some starting value before being used in calculations or being printed out

Statements „ „

A statement is a single line of code Can be: „ „

Variable or function declarations Arithmetic operations „

„

Function calls „

„ „

x = 5+3; printf(“Hello!”);

Control flow commands More to be seen…

Types „

„

„

„

“Type” in C refers to the kind of data or information Functions have return types Arguments to functions each have their own type Variables have types

Why Type? „

„ „

Why do we have to specify types of variables, functions, arguments? Has to do with computer memory Different kinds of data require different amounts of memory to store „

„

„

A single character can have only one of 128 values (a-z,AZ,0-9,punctuation, some others) An integer can have an infinite number of values, limited to ~65,000 values on a computer Therefore, more memory needed to store an integer than a character

Computer Memory „ „

„

„

„

Memory in computers made up of transistors Transistor: just a switch that can be either on or off “on” state corresponds to the value 1, “off” state corresponds to the value 0 Everything in memory has to be made up of 0s and 1s – i.e., has to be stored as a number in base 2 (binary) Important to understand different bases

Back to Types „

„

Types typically defined in pieces of memory that are powers of 2 Smallest piece of memory: 1 bit „

„

Can hold 0 or 1 (equiv to 1 transistor)

8 bits = 1 byte „

„

„

1 byte can have any value between 0000 0000 and 1111 1111 – ie, between 0 – 255. 1111 1111 binary hex digit: 1111 = 8+4+2+1 = 15 = E 0xEE = 15x161 + 15x160 = 240 + 15 = 255 More than number of values needed for characters – 1 byte typically used for character type

Numeric Types „

16 bits = 2 bytes can have any value from 0000 0000 0000 0000– 1111 1111 1111 1111 or 0 – 65,535. (Shortcut: 65,535 = 216-1) „ „

„

32-bit integers now more common „

„

Was used for a long time for integer values If used to store negative integers, could only store up to +/32,768 (approx) Ever heard of 32-bit operating system? Means that main data types used are 32-bit

Amount of memory given to each type dependent on the machine and operating system

Type Sizes „

Many different types in C (more to come) „

char: a single character „

„

int: an integer value „

„

Often 32 bits/4 bytes, 4 billion values

float: a floating-point number „

„

Often 1 byte, 128 values

Often 4 bytes

double: a double-precision floating-point number „

Double the size of a float, often 64 bits or 8 bytes

Operators „ „ „ „ „ „ „ „

Arithmetic: + - * / % -(unary) Increment/decrement: ++ -Relational: > >= < <= Equality: == != Assignment: = += -= *= Logical: || && Bitwise: | & ^ << >> ~ Reference/Dereference: & *

Type conversions „ „

Some types can automatically be converted to others int, float, double can often be treated the same for calculation purposes „

„

int values will be converted up into floats for calc.

But major differences between integer and floatingpoint arithmetic! „ „ „

Fractions just dropped in integer arithmetic Use mod (%) operator to get remainders Integer much faster to perform

Start here „ „ „ „ „

Hello, World – a first program Hello, World – line by line Basic rules for C Tips for good programming style Basic Input/Output

Hello, World /* Hello, World program */ #include <stdio.h> int main() { printf(“Hello, World!\n”); return 0; }

Programming Style „

„

„

„ „

Put brackets on separate lines by themselves (some variations on this) After each open bracket, increase indentation level by one tab After each close bracket, decrease indentation level by one tab Leave blank lines between sections Don’t let lines get too long

Control Flow „ „

Two basic types of control flow commands Conditionals „

„

„

Execute some commands if a given condition is true/false if, switch, ?: operator

Loops „

„

Execute some commands repeatedly until a given condition for stopping is met while, for, do-while

Functions „

„

Basic idea of functions is to divide code up into small, manageable chunks One way to get started designing funcs: „ „

„

Write out the entire program with no functions Look for sections of code that are almost exactly duplicated Create one function for each repeated section and replace each repetition with the function call

Function Design „

„ „

Look for places where several lines of code are used to accomplish a single task and move the code into a function No function too small Rule of thumb: no function (including main) should be longer than a page „

Goal: be able to see entire function at once when editing program

So,What is a function? „

Types of function „ „

„

Std UDF

Nature of function „ „ „ „

Stack use Recursive Chained …

Using variable number of arguments •Macros that implement variable argument lists • Declaration: void va_start(va_list ap, lastfix); type va_arg(va_list ap, type); void va_end(va_list ap); • Remarks: •Some C functions, such as vfprintf and vprintf, take variable argument lists in addition to taking a number of fixed (known) parameters. •The va_arg, va_end, and va_start macros provide a portable way to access these argument lists. •va_start sets ap to point to the first of the variable arguments being passed to the function. •va_arg expands to an expression that has the same type and value as the next argument being passed (one of the variable arguments). Because of default promotions, you can't use char, unsigned char, or float types with va_arg.va_end helps the called function perform a normal return.

Using variable number of arguments… • ap •Points to the va_list variable argument list being passed to the function. The ap to va_arg should be the same ap that va_start initialized. • lastfix •The name of the last fixed parameter being passed to the called function. • type •Used by va_arg to perform the dereference and to locate the following item.

•You use these macros to step through a list of arguments when the called function does not know the number and types of the arguments being passed.

Using variable number of arguments… • Calling Order •va_start must be used before the first call to va_arg or va_end. •va_end should be called after va_arg has read all the arguments. •Failure to do so might cause strange, undefined behavior in your program.

•Stepping Through the Argument List •The first time va_arg is used, it returns the first argument in the list. •Each successive time va_arg is used, it returns the next argument in the list. •It does this by first dereferencing ap, then incrementing ap to point to the next item.

Using variable number of arguments… void sum(char *msg, ...) { int total = 0;

va_list ap;

int arg;

va_start(ap, msg); while ((arg = va_arg(ap,int)) != 0) { total += arg;} printf(msg, total); va_end(ap); } int main(void) { sum("The total of 1+2+3+4 is %d\n", 1,2,3,4,0); return 0; }

Arrays „

„

An array variable is just a set of variables all of the same type, all linked together Can make an array variable of ANY type – built-in ones, or ones that you define with struct or enum char string[100]; // 100 chars int numArray[50]; // 50 ints Date dateArray[10]; // 10 Dates

Array Elements „

„

Each element of the array can be treated like an individual object of the underlying array type To access an element, use the subscript operator: <arrayname>[] : represents the element at position in the array named <arrayname>

Array Indices „

„

„

Indices into an array start at 0 and go up to (array length – 1). Very convenient when combined with for loops Can easily go through entire array to access every element in just a few lines of code, no matter how large the array

Example – with arrays #define ARRAYSIZE 10 int main() { int numArray[ARRAYSIZE]; int sum=0, i; for (i=0; i < ARRAYSIZE; i++) { numArray[i] = i; sum += numArray[i]; printf(“%d “, numArray[i]); } printf(“\nSum is: %d\n”, sum); return 0; }

Limitations „

Can’t assign or compare arrays

int numArray[10]; int numArrayCopy[10]; // Not OK: numArrayCopy = numArray; if (numArray == numArrayCopy) // OK: for(I=0; I<10; I++) numArrayCopy[I] = numArray[I];

What is a pointer?

Pointer Syntax „

„

Complicated – uses reference (&) and dereference (*) operators & „ „

„

returns the memory address of that variable result is a pointer to whatever type the variable is

*<pointer variable> „

„

accesses the memory address stored in the pointer result is of whatever type the original variable was

Pointers „

„

If all pointers are same size, why do we need to specify pointer type? Two reasons „

„

Need to know what type of variable we are going to get when we dereference the pointer Sometimes, pointer is not pointing to just one variable – can also point to first variable in an array

Overview „

Pointers „ „ „ „

Pointer to basic data types Pointers and arrays Pointers to function Function returning pointers

Overview „

Pointers „ „

„ „ „

Pointers and arrays Pointers and strings

Null pointer Generic pointer (void *) Dynamic Memory Allocation „ „

malloc() free()

NULL – the null pointer „

„

„

„

A pointer with the address 0x0 is a null pointer NULL: a constant equivalent to the address 0x0 Must NEVER dereference the null pointer – crash will occur Major source of bugs (common cause of ‘the blue screen of death’)

Protecting against null pointers „

„

Always initialize all pointers to NULL when declaring them NULL and the null character both equivalent to false „

„

Can be used in conditional statements

Typical technique: if (str) { // dereference str somehow } else { // str is NULL! }

Generic Pointer „ „

„ „

„

Pointers are all the same size underneath Sometimes convenient to treat all pointer types interchangeably Generic pointer type: void * Function that expects void * can take any pointer as argument Important in malloc/free

Dynamic Memory Allocation „

„

Declaring variables of fixed size allocates memory in a static way – variables do not change as program runs Can also declare memory dynamically „

„

Allocate different amounts of memory from run to run of the program Increase/reduce amount of memory as program runs

Dynamic Memory Allocation „

„

„

Have seen one example: using a variable to determine size of an array More flexible technique: combine pointers with the function malloc() malloc() and free() are in stdlib.h

Using Malloc char * str = NULL; // allocate 10 bytes to be used // to store char values str = (char *)malloc(10); // when finished, clean up free(str);

Malloc „

„ „

malloc() takes the number of bytes to allocate returns void * pointer Problems: „ „

need to calculate the number of bytes need to use the void * pointer as pointer of type you want to use (eg, int *, char *)

Calculating Bytes „

„

sizeof() – returns the number of bytes used by a single variable of that type multiply this value by however many variables of this type you want to store

int intArray[10]; int * iarray = (int *) malloc(sizeof(int)*10);

Casting „

„

„

Casting allows you to force a variable of one type to be treated as another Does NOT perform any conversion – just interprets the binary data stored in memory in a different way If you try to treat a variable of one type as an incompatible type, your program will likely crash

Casting Syntax: (<desired type>) „ Used most commonly with malloc() and other functions that return void * „ Can also be used to get integer values of character variables int valueOfA = (int) ‘A’; „

Overview „

Dynamic Memory Allocation „ „ „

„ „ „

malloc() free() realloc()

Dangling Pointers Memory Leaks Avoiding bugs

Dynamic Memory Allocation „

„

Declaring variables of fixed size allocates memory in a static way – variables do not change as program runs Can also declare memory dynamically „

„

Allocate different amounts of memory from run to run of the program Increase/reduce amount of memory as program runs

Malloc „

„

malloc() takes the number of bytes to allocate returns void * pointer holding the address of the chunk of memory that was allocated

Free „

„

free() takes a pointer of any type holding a memory address Deallocates the chunk of memory starting at that address and gives it back to the computer to let other programs use it

Realloc „

„

realloc() takes a pointer holding a memory address and the number of bytes to allocate Allocates the additional memory needed to get up to the specified number of bytes „

Will free up memory if fewer bytes than original specified!

Memory Management „

„

malloc() and free() must be used carefully to avoid bugs Potential problems „ „

dangling pointers memory leaks

Dangling Pointer „

„

„

Pointer that is not pointing to a valid, allocated chunk of memory Typically happens when you free a pointer and then forget to reset it to NULL, then try to use the pointer again Causes unpredictable behavior

Memory Leak „

„

„

Chunk of memory that has been allocated and will never be freed up in the course of the program Typically happens when you reset a pointer to NULL or to a new memory address without freeing it first Consumes unnecessary system resources and slows down program/computer

Avoiding Memory Mgmt Bugs „ „

Must match up calls to malloc() and free() Good technique – print out messages

#define DEBUG 1 ... if (DEBUG) printf(“malloc”); // malloc ... if (DEBUG) printf(“free”); // free ... if (DEBUG) printf(“resetting pointer”) // assigning pointer variable

Avoiding Bugs Cont. „

„

If more malloc() calls than free(), you have a memory leak If fewer pointer resets than free() calls, you may have a dangling pointer

Using Dynamic Allocation „

Only allocate the memory you need for strings

„

Grow a database on the fly

„

Shrink even!!

Dynamic Memory Allocation & Functions „

„

Often want to dynamically allocate memory within a function CAN be done „

„

Memory allocated with malloc() doesn’t get freed up automatically when a function exits the way local variables do

Often requires passing *pointers* by reference

Pointers & Pass-By-Reference „

„

„

Remember: when pointers are used to implement pass-by-reference, a copy is still being made! Copy holds same memory address as original pointer – so both can be used to access the original mem location However, when memory address stored in the copy is changed, the address stored in the original pointer is NOT changed!

Pointers & Pass-By-Reference „

„

Assuming a function that takes an argument named “ptr” WILL change ptr in the caller: *ptr = ; strcpy(ptr, ); ptr[0] = ;

„

Will NOT change ptr in the caller: ptr = (int *) malloc(...); ptr = ; ptr = &;

Pointers & Pass-By-Reference „

„

Rule of thumb: anything you do to a pointer argument inside a function MUST involve a dereference If pointer is treated as an array, that’s equivalent to a dereference

C Preprocessor „

„

Runs on a source code file before it gets to the compiler Gets the source into compilable form „ „ „ „ „

Strips out comments Loads in header files Replaces constant names with their values Expands macros Can chop out unneeded sections of code

C Preprocessor „ „

Has its own very simple language Three main kinds of statements: „ „ „

„

#include #define conditionals (#ifdef, #ifndef, #if, #else, #elif, #endif) #pragma

Header Files and #include „ „

Syntax: #include Syntax: #include “header file.h” „

„

Angle brackets used for shared, standard libraries (string.h, stdlib.h) Double-quotes used for header files you have written yourself

Header Files „

„

„ „

#include works by essentially copying & pasting in all the contents of the header file into your program Many source code files can all share the same header file Useful for reusing code Next time, will see use of header files

Constants with #define „ „

„ „ „

Syntax: #define NAME VALUE VALUE is pasted in wherever NAME appears in the code Can also just define a name Syntax: #define NAME Can later test this name with #ifdef statements

Macros with #define „ „

„

Can use same syntax to define macros Macros are like less-full-featured functions; just pasted into code Example: „

„

#define ADD(a,b) a+b

Usage: „

int x = ADD(5,3); => int x = 5+3;

Macros „ „ „ „

Can make extremely complex macros Faster to execute than a function Much harder to debug than functions Example: „ „ „ „

#define ADD(a,b) a+b int x = ADD(5,3)*2; =>int x = 5+3*2; Result: 5+6=11 NOT 8*2=16 #define ADD(a,b) ((a)+(b)) must be used

Macros „ „ „

„

Macro definitions must all be on one line Can get long and messy Use \ at very end of line to continue it onto the next line #define COMPLEX_MACRO(a,b) \ printf(“complex macro\n”); \ printf(“%d+%d=%d\n”, \ (a),(b),((a)+(b)));

Conditional Compilation „

„

„

Several conditional statements present in the preprocessor language Used to block out (or activate) chunks of code based on a single constant value Syntax: #ifdef /* conditional code goes here*/ #endif

Conditional Compilation Useful for debugging #define DEBUG #ifdef DEBUG printf(“Debugging info\n”); for (int i=0; i<10; i++) printf(“Array[%d] = \n”,i,array[i]); #endif

„

Conditional Compilation Also useful for compiling on different platforms // #define WINDOWS #define UNIX ... #ifdef UNIX #define END_OF_LINE “\n” #endif #ifdef WINDOWS #define END_OF_LINE “\r\n” #endif „