Brief Introduction to the C

Programming Language

Shashank Capoor
shashank_capoor@rediffmail.com

Washington
WASHINGTON UNIVERSITY IN ST LOUIS
 Introduction
• The C programming language was designed by Dennis
Ritchie at Bell Laboratories in the early 1970s
• Influenced by
– ALGOL 60 (1960),
– CPL (Cambridge, 1963),
– BCPL (Martin Richard, 1967),
– B (Ken Thompson, 1970)
• Traditionally used for systems programming, though
this may be changing in favor of C++
• Traditional C:
– The C Programming Language, by Brian Kernighan and Dennis
Ritchie, 2nd Edition, Prentice Hall
– Referred to as K&R

2
 Standard C
• Standardized in 1989 by ANSI (American National
Standards Institute) known as ANSI C
• International standard (ISO) in 1990 which was
adopted by ANSI and is known as C89
• As part of the normal evolution process the standard
was updated in 1995 (C95) and 1999 (C99)
• C++ and C
– C++ extends C to include support for Object Oriented
Programming and other features that facilitate large software
development projects
– C is not strictly a subset of C++, but it is possible to write
“Clean C” that conforms to both the C++ and C standards.

3
 Elements of a C Program
• A C development environment includes
– System libraries and headers: a set of standard libraries and
their header files. For example see /usr/include and glibc.
– Application Source: application source and header files
– Compiler: converts source to object code for a specific platform
– Linker: resolves external references and produces the
executable module
• User program structure
– there must be one main function where execution begins when the
program is run. This function is called main
• int main (void) { ... },
• int main (int argc, char *argv[]) { ... }
• UNIX Systems have a 3rd way to define main(), though it is not
POSIX.1 compliant
int main (int argc, char *argv[], char *envp[])
– additional local and external functions and variables
4
 A Simple C Program
• Create example file: try.c
• Compile using gcc:
gcc –o try try.c
• The standard C library libc is included
automatically /* you generally want to
* include stdio.h and
• Execute program * stdlib.h
./try * */
#include <stdio.h>
• Note, I always specify an absolute path #include <stdlib.h>
• Normal termination:
void exit(int status); int main (void)
{
– calls functions registered with printf(“Hello World\n”);
atexit() exit(0);
– flush output streams }
– close all open streams
– return status value and control to host
environment
5
 Source and Header files
• Just as in C++, place related code within the same module
(i.e. file).
• Header files (*.h) export interface definitions
– function prototypes, data types, macros, inline functions and other
common declarations
• Do not place source code (i.e. definitions) in the header
file with a few exceptions.
– inline’d code
– class definitions
– const definitions
• C preprocessor (cpp) is used to insert common definitions
into source files
• There are other cool things you can do with the
preprocessor

6
 Another Example C Program
#include directs the preprocessor
/usr/include/stdio.h to “include” the contents of the file
/* comments */
#ifndef _STDIO_H
at this point in the source file.
#define _STDIO_H #define directs preprocessor to
define macros.
... definitions and protoypes
example.c
#endif /* this is a C-style comment
* You generally want to palce
* all file includes at start of file
* */
/usr/include/stdlib.h #include <stdio.h>
/* prevents including file #include <stdlib.h>
* contents multiple
* times */ int
#ifndef _STDLIB_H main (int argc, char **argv)
#define _STDLIB_H {
// this is a C++-style comment
... definitions and protoypes // printf prototype in stdio.h
printf(“Hello, Prog name = %s\n”,
#endif argv[0]);
exit(0);
}

7
 Passing Command Line Arguments
• When you execute a program ./try –g 2 fred
you can include arguments on
the command line.
• The run time environment will argc = 4,
argv = <address0>
create an argument vector.
– argv is the argument vector
– argc is the number of
arguments
• Argument vector is an array of argv:
‘t’‘r’‘y’‘\0’

pointers to strings. [0] <addres1> ‘-’‘g’‘\0’

[1] <addres2>
• a string is an array of [2] <addres3>
characters terminated by a [3] <addres4> ‘2’‘\0’
binary 0 (NULL or ‘\0’). [4] NULL

• argv[0] is always the program ‘f’‘r’‘e’‘d’‘\0’

name, so argc is at least 1.

8
C Standard Header Files you may want to use

• Standard Headers you should know about:

– stdio.h – file and console (also a file) IO: perror, printf,
open, close, read, write, scanf, etc.
– stdlib.h - common utility functions: malloc, calloc,
strtol, atoi, etc
– string.h - string and byte manipulation: strlen, strcpy,
strcat, memcpy, memset, etc.
– ctype.h – character types: isalnum, isprint,
isupport, tolower, etc.
– errno.h – defines errno used for reporting system errors
– math.h – math functions: ceil, exp, floor, sqrt, etc.
– signal.h – signal handling facility: raise, signal, etc
– stdint.h – standard integer: intN_t, uintN_t, etc
– time.h – time related facility: asctime, clock, time_t,
etc.
9
 The Preprocessor
• The C preprocessor permits you to define simple macros
that are evaluated and expanded prior to compilation.

• Commands begin with a ‘#’. Abbreviated list:

– #define : defines a macro
– #undef : removes a macro definition
– #include : insert text from file
– #if : conditional based on value of expression
– #ifdef : conditional based on whether macro defined
– #ifndef : conditional based on whether macro is not defined
– #else : alternative
– #elif : conditional alternative
– defined() : preprocessor function: 1 if name defined, else 0
#if defined(__NetBSD__)

10
 Preprocessor: Macros
• Using macros as functions, exercise caution:
– flawed example: #define mymult(a,b) a*b
• Source: k = mymult(i-1, j+5);
• Post preprocessing: k = i – 1 * j + 5;
– better: #define mymult(a,b) (a)*(b)
• Source: k = mymult(i-1, j+5);
• Post preprocessing: k = (i – 1)*(j + 5);
• Be careful of side effects, for example what if we did
the following
– Macro: #define mysq(a) (a)*(a)
– flawed usage:
• Source: k = mysq(i++)
• Post preprocessing: k = (i++)*(i++)
• Alternative is to use inline’ed functions
– inline int mysq(int a) {return a*a};
– mysq(i++) works as expected in this case.
11
 Preprocessor: Conditional Compilation
• Its generally better to use inline’ed functions
• Typically you will use the preprocessor to define constants,
perform conditional code inclusion, include header files or to
create shortcuts
• #define DEFAULT_SAMPLES 100
• #ifdef __linux
static inline int64_t
gettime(void) {...}
• #elif defined(sun)
static inline int64_t
gettime(void) {return (int64_t)gethrtime()}
• #else
static inline int64_t
gettime(void) {... gettimeofday()...}
• #endif

12
 Another Simple C Program
int main (int argc, char **argv) {
int i;
printf(“There are %d arguments\n”, argc);
for (i = 0; i < argc; i++)
printf(“Arg %d = %s\n”, i, argv[i]);

return 0;
}
• Notice that the syntax is similar to Java
•What’s new in the above simple program?
– of course you will have to learn the new interfaces and utility
functions defined by the C standard and UNIX
– Pointers will give you the most trouble

13
 Arrays and Pointers
• A variable declared as an array represents a contiguous
region of memory in which the array elements are stored.
int x[5]; // an array of 5 4-byte ints.
little endian byte ordering

• All arrays begin with an index of 0 0

0 1 2 3

1
2
3
4

memory layout for array x

• An array identifier is equivalent to a pointer that

references the first element of the array
– int x[5], *ptr;
ptr = &x[0] is equivalent to ptr = x;
• Pointer arithmetic and arrays:
– int x[5];
x[2] is the same as *(x + 2), the compiler will assume you mean
2 objects beyond element x.
14
 Pointers
• For any type T, you may form a pointer type to T.
– Pointers may reference a function or an object.
– The value of a pointer is the address of the corresponding object or
function
– Examples: int *i; char *x; int (*myfunc)();
• Pointer operators: * dereferences a pointer, & creates a pointer
(reference to)
– int i = 3; int *j = &i;
*j = 4; printf(“i = %d\n”, i); // prints i = 4
– int myfunc (int arg);
int (*fptr)(int) = myfunc;
i = fptr(4); // same as calling myfunc(4);
• Generic pointers:
– Traditional C used (char *)
– Standard C uses (void *) – these can not be dereferenced or used in
pointer arithmetic. So they help to reduce programming errors
• Null pointers: use NULL or 0. It is a good idea to always initialize
pointers to NULL.
15
 Pointers in C (and C++)
Program Memory Address
Step 1:
int main (int argc, argv) {
int x = 4;
x 4 0x3dc
int *y = &x;
int *z[4] = {NULL, NULL, NULL, NULL}; y 0x3dc 0x3d8
int a[4] = {1, 2, 3, 4}; NA 0x3d4
... NA 0x3d0
z[3] 0 0x3cc
z[2] 0 0x3c8
Note: The compiler converts z[1] or *(z+1) to z[1] 0 0x3c4
Value at address (Address of z + sizeof(int)); z[0] 0 0x3c0
a[3] 4 0x3bc
In C you would write the byte address as: a[2] 0x3b8
3
(char *)z + sizeof(int);
a[1] 2 0x3b4
a[0] 1 0x3b0
or letting the compiler do the work for you
(int *)z + 1;

16
 Pointers Continued
Step 1: Program Memory Address
int main (int argc, argv) {
int x = 4;
int *y = &x; x 4 0x3dc
int *z[4] = {NULL, NULL, NULL, NULL}; 0x3d8
y 0x3dc
int a[4] = {1, 2, 3, 4};
NA 0x3d4
Step 2: Assign addresses to array Z
NA 0x3d0
z[0] = a; // same as &a[0];
z[1] = a + 1; // same as &a[1]; z[3] 0x3bc 0x3cc
z[2] = a + 2; // same as &a[2]; z[2] 0x3b8 0x3c8
z[3] = a + 3; // same as &a[3]; z[1] 0x3b4 0x3c4
z[0] 0x3b0 0x3c0
a[3] 4 0x3bc
a[2] 3 0x3b8
a[1] 2 0x3b4
a[0] 1 0x3b0

17
 Pointers Continued
Step 1: Program Memory Address
int main (int argc, argv) {
int x = 4;
int *y = &x; 0x3dc
x 4
int *z[4] = {NULL, NULL, NULL,
y 0x3dc 0x3d8
NULL};
NA 0x3d4
int a[4] = {1, 2, 3, 4};
Step 2: NA 0x3d0
z[0] = a; z[3] 0x3bc 0x3cc
z[1] = a + 1; z[2] 0x3b8 0x3c8
z[2] = a + 2; z[1] 0x3b4 0x3c4
z[3] = a + 3; z[0] 0x3b0 0x3c0
Step 3: No change in z’s values a[3] 4 0x3bc
z[0] = (int *)((char *)a); a[2] 3 0x3b8
z[1] = (int *)((char *)a a[1] 2 0x3b4
+ sizeof(int));
a[0] 1 0x3b0
z[2] = (int *)((char *)a
+ 2 * sizeof(int));
z[3] = (int *)((char *)a
+ 3 * sizeof(int));
18
 Getting Fancy with Macros
#define QNODE(type) \ #define QINSERT_BEFORE(loc, node, field) \
struct { \ do { \
struct type *next; \
struct type **prev; \ *(loc)->field.prev = (node); \
} (node)->field.prev = \
(loc)->field.prev; \
#define QNODE_INIT(node, field) \ (loc)->field.prev = \
do { \ &((node)->field.next); \
(node)->field.next = (node); \
(node)->field.prev = \ (node)->field.next = (loc); \
&(node)->field.next; \ } while (/* */0)
} while ( /* */ 0 );
#define QINSERT_AFTER(loc, node, field) \
#define QFIRST(head, field) \ do { \
((head)->field.next) ((loc)->field.next)->field.prev = \
&(node)->field.next; \
#define QNEXT(node, field) \ (node)->field.next = (loc)->field.next; \
((node)->field.next) (loc)->field.next = (node); \
(node)->field.prev = &(loc)->field.next; \
#define QEMPTY(head, field) \ } while ( /* */ 0)
((head)->field.next == (head))
#define QREMOVE(node, field) \
#define QFOREACH(head, var, field) \ do { \
for ((var) = (head)->field.next; \ *((node)->field.prev) = (node)->field.next; \
(var) != (head); \ ((node)->field.next)->field.prev = \
(var) = (var)->field.next) (node)->field.prev; \
(node)->field.next = (node); \
(node)->field.prev = &((node)->field.next); \
} while ( /* */ 0)

19
 After Preprocessing and Compiling
typedef struct wth_t {
int state;
typedef struct wth_t
struct {
{ CPP struct wth_t *next;
int state;
struct wth_t **prev;
QNODE(wth_t) alist;
} alist;
} wth_t;
#define QNODE_INIT(node, field) } wth_t; \
do {
#define QNODE(type) \ \
(node)->field.next
struct { \= (node); \
(node)->field.prev
struct type *next; \= &(node)->field.next;\memory layout after GCC
}struct
while type
( /* **prev;
*/ 0 ); \
}head: instance of wth_t 3 words in memory
0x100 0 QNODE_INIT(head, alist) <integer> state
0x104 0x00100 <address> next
0x108 0x00104 <address> prev

20
 QNODE Manipulations
before #define QINSERT_BEFORE(head, node, alist)\
do { \
head node0
*(head)->alist.prev = (node); \
0x100 0 0x1a0 0
(node)->alist.prev = (head)->alist.prev; \
0x104 0x100 0x1a4 0x1a0
(head)->alist.prev = &(node)->alist.next;\
0x108 0x104 0x1a8 0x1a4
(node)->alist.next = (head); \
} while (/* */0)

QINSERT_BEFORE(head, node0, alist);

?
21
 QNODE Manipulations
before #define QINSERT_BEFORE(head, node, alist)\
do { \
head node0
*(head)->alist.prev = (node); \
0x100 0 0x1a0 0
(node)->alist.prev = (head)->alist.prev; \
0x104 0x100 0x1a4 0x1a0
(head)->alist.prev = &(node)->alist.next;\
0x108 0x104 0x1a8 0x1a4
(node)->alist.next = (head); \
} while (/* */0)

QINSERT_BEFORE(head, node0, alist);

head node0
0x100 0 0x1a0 0
0x104 0x1a0 0x1a4 0x1a0
0x108 0x104 0x1a8 0x1a4

22
 QNODE Manipulations
before #define QINSERT_BEFORE(head, node, alist)\
do { \
head node0
*(head)->alist.prev = (node); \
0x100 0 0x1a0 0
(node)->alist.prev = (head)->alist.prev; \
0x104 0x100 0x1a4 0x1a0
(head)->alist.prev = &(node)->alist.next;\
0x108 0x104 0x1a8 0x1a4
(node)->alist.next = (head); \
} while (/* */0)

QINSERT_BEFORE(head, node0, alist);

head node0
0x100 0 0x1a0 0
0x104 0x1a0 0x1a4 0x1a0
0x108 0x104 0x1a8 0x104

23
 QNODE Manipulations
before #define QINSERT_BEFORE(head, node, alist)\
do { \
head node0
*(head)->alist.prev = (node); \
0x100 0 0x1a0 0
(node)->alist.prev = (head)->alist.prev; \
0x104 0x100 0x1a4 0x1a0
(head)->alist.prev = &(node)->alist.next;\
0x108 0x104 0x1a8 0x1a4
(node)->alist.next = (head); \
} while (/* */0)

QINSERT_BEFORE(head, node0, alist);

head node0
0x100 0 0x1a0 0
0x104 0x1a0 0x1a4 0x1a0
0x108 0x1a4 0x1a8 0x104

24
 QNODE Manipulations
before #define QINSERT_BEFORE(head, node, alist)\
do { \
head node0
*(head)->alist.prev = (node); \
0x100 0 0x1a0 0
(node)->alist.prev = (head)->alist.prev; \
0x104 0x100 0x1a4 0x1a0
(head)->alist.prev = &(node)->alist.next;\
0x108 0x104 0x1a8 0x1a4
(node)->alist.next = (head); \
} while (/* */0)

QINSERT_BEFORE(head, node0, alist);

head node0
0x100 0 0x1a0 0
0x104 0x1a0 0x1a4 0x100
0x108 0x1a4 0x1a8 0x104

25
 QNODE Manipulations
before #define QINSERT_BEFORE(head, node, alist)\
do { \
head node0
*(head)->alist.prev = (node); \
0x100 0 0x1a0 0
(node)->alist.prev = (head)->alist.prev; \
0x104 0x100 0x1a4 0x1a0
(head)->alist.prev = &(node)->alist.next;\
0x108 0x104 0x1a8 0x1a4
(node)->alist.next = (head); \
} while (/* */0)

QINSERT_BEFORE(head, node0, alist);

head node0
0x100 0 0x1a0 0
0x104 0x1a0 0x1a4 0x100
0x108 0x1a4 0x1a8 0x104

26
 Adding a Third Node
head node0 #define QINSERT_BEFORE(head, node, alist)\
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x100 *(head)->alist.prev = (node); \
0x108 0x1a4 0x1a8 0x104 (node)->alist.prev = (head)->alist.prev; \
(head)->alist.prev = &(node)->alist.next; \
node1 (node)->alist.next = (head); \
0x200 0
} while (/* */0)
0x204 0x200
0x208 0x204

QINSERT_BEFORE(head, node1, alist);

head node0 node1

0x100 0 0x1a0 0 0x200 0
0x104 0x1a0 0x1a4 0x100 0x204 0x200
0x108 0x1a4 0x1a8 0x104 0x208 0x204

27
 Adding a Third Node
head node0 #define QINSERT_BEFORE(head, node1, alist)\
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x100 (1) *(head)->alist.prev = (node1); \
0x108 0x1a4 0x1a8 0x104 (node1)->alist.prev = (head)->alist.prev; \
(head)->alist.prev = &(node1)->alist.next; \
node1 (node1)->alist.next = (head); \
0x200 0
} while (/* */0)
0x204 0x200
0x208 0x204

QINSERT_BEFORE(head, node1, alist);

head node0 (1) node1

0x100 0 0x1a0 0 0x200 0
0x104 0x1a0 0x1a4 0x200 0x204 0x200
0x108 0x1a4 0x1a8 0x104 0x208 0x204

28
 Adding a Third Node
head node0 #define QINSERT_BEFORE(head, node1, alist)\
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x100 *(head)->alist.prev = (node1); \
0x108 0x1a4 0x1a8 0x104 (2) (node1)->alist.prev = (head)->alist.prev; \
(head)->alist.prev = &(node1)->alist.next; \
node1 (node1)->alist.next = (head); \
0x200 0
} while (/* */0)
0x204 0x200
0x208 0x204

QINSERT_BEFORE(head, node1, alist);

head node0 node1

0x100 0 0x1a0 0 (1) 0x200 0
0x104 0x1a0 0x1a4 0x200 0x204 0x200
0x108 0x1a4 0x1a8 0x104 0x208 0x1a4

(2)
29
 Adding a Third Node
head node0 #define QINSERT_BEFORE(head, node1, alist)\
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x100 (1) *(head)->alist.prev = (node1); \
0x108 0x1a4 0x1a8 0x104 (2) (node1)->alist.prev = (head)->alist.prev; \
(3) (head)->alist.prev = &(node1)->alist.next; \
node1 (node1)->alist.next = (head); \
0x200 0
} while (/* */0)
0x204 0x200
0x208 0x204

QINSERT_BEFORE(head, node1, alist);

head node0 node1

0x100 0 0x1a0 0 (1) 0x200 0
0x104 0x1a0 0x1a4 0x200 0x204 0x200
0x108 0x204 0x1a8 0x104 0x208 0x1a4
(3)
(2)
30
 Adding a Third Node
head node0 #define QINSERT_BEFORE(head, node1, alist)\
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x100 (1) *(head)->alist.prev = (node1); \
0x108 0x1a4 0x1a8 0x104 (2) (node1)->alist.prev = (head)->alist.prev; \
(3) (head)->alist.prev = &(node1)->alist.next; \
node1 (4) (node1)->alist.next = (head); \
0x200 0
} while (/* */0)
0x204 0x200
0x208 0x204

QINSERT_BEFORE(head, node1, alist);

(4)

head node0 node1

0x100 0 0x1a0 0 (1) 0x200 0
0x104 0x1a0 0x1a4 0x200 0x204 0x100
0x108 0x204 0x1a8 0x104 0x208 0x1a4
(3)
(2)
31
 Removing a Node
head node0 #define QREMOVE(node, alist) \
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x200 (1) *((node)->alist.prev) = (node)->alist.next; \
0x108 0x204 0x1a8 0x104 (2) ((node)->alist.next)->alist.prev = (node)->alist.prev;\
(3) (node)->alist.next = (node); \
node1 (4) (node)->alist.prev = &((node)->alist.next); \
0x200 0
} while ( /* */ 0)
0x204 0x100
0x208 0x1a4

QREMOVE(node0, alist);

head node0 node1

0x100 0 0x1a0 0 0x200 0
0x104 ?? 0x1a4 ?? 0x204 ??
0x108 ?? 0x1a8 ?? 0x208 ??

32
 Removing a Node
head node0 #define QREMOVE(node, alist) \
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x200 *((node)->alist.prev) = (node)->alist.next; \
0x108 0x204 0x1a8 0x104 ((node)->alist.next)->alist.prev = (node)->alist.prev;\
(node)->alist.next = (node); \
node1 (node)->alist.prev = &((node)->alist.next); \
0x200 0
} while ( /* */ 0)
0x204 0x100
0x208 0x1a4
QREMOVE(node0, alist);

head node0 node1

0x100 0 0x1a0 0 0x200 0
0x104 0x1a0 0x1a4 0x200 0x204 0x100
0x108 0x204 0x1a8 0x104 0x208 0x1a4

33
 Removing a Node
head node0 #define QREMOVE(node0, alist) \
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x200 (1) *((node0)->alist.prev) = (node0)->alist.next; \
0x108 0x204 0x1a8 0x104 ((node0)->alist.next)->alist.prev = (node0)->alist.prev;\
(node0)->alist.next = (node0); \
node1 (node0)->alist.prev = &((node0)->alist.next);\
0x200 0
} while ( /* */ 0)
0x204 0x100
0x208 0x1a4
QREMOVE(node0, alist);

head node0 node1

0x100 0 (1) 0x1a0 0 0x200 0
0x104 0x200 0x1a4 0x200 0x204 0x100
0x108 0x204 0x1a8 0x104 0x208 0x1a4

34
 Removing a Node
head node0 #define QREMOVE(node0, alist) \
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x200 *((node0)->alist.prev) = (node0)->alist.next; \
0x108 0x204 0x1a8 0x104 (2) ((node0)->alist.next)->alist.prev = (node0)->alist.prev;\
(node0)->alist.next = (node0); \
node1 (node0)->alist.prev = &((node0)->alist.next); \
0x200 0
} while ( /* */ 0)
0x204 0x100
0x208 0x1a4
QREMOVE(node0, alist);

head node0 node1

0x100 0 0x1a0 0 0x200 0
0x104 0x200 0x1a4 0x200 0x204 0x100
0x108 0x204 0x1a8 0x104 0x208 0x104
(2)

35
 Removing a Node
head node0 #define QREMOVE(node0, alist) \
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x200 *((node0)->alist.prev) = (node0)->alist.next; \
0x108 0x204 0x1a8 0x104 ((node0)->alist.next)->alist.prev = (node0)->alist.prev;\
(3) (node0)->alist.next = (node0); \
node1 (node0)->alist.prev = &((node0)->alist.next); \
0x200 0
} while ( /* */ 0)
0x204 0x100
0x208 0x1a4
QREMOVE(node0, alist);

head node0 node1

(3)
0x100 0 0x1a0 0 0x200 0
0x104 0x200 0x1a4 0x1a0 0x204 0x100
0x108 0x204 0x1a8 0x104 0x208 0x104

36
 Removing a Node
head node0 #define QREMOVE(node0, alist) \
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x200 *((node0)->alist.prev) = (node0)->alist.next; \
0x108 0x204 0x1a8 0x104 ((node0)->alist.next)->alist.prev = (node0)->alist.prev;\
(node0)->alist.next = (node0); \
node1 (4) (node0)->alist.prev = &((node0)->alist.next); \
0x200 0
} while ( /* */ 0)
0x204 0x100
0x208 0x1a4
QREMOVE(node0, alist);

head node0 node1

0x100 0 0x1a0 0 0x200 0
0x104 0x200 0x1a4 0x1a0 0x204 0x100
(4)
0x108 0x204 0x1a8 0x1a4 0x208 0x104

37
 Solution to Removing a Node
head node0 #define QREMOVE(node, alist) \
0x100 0 0x1a0 0 do { \
0x104 0x1a0 0x1a4 0x200 (1) *((node)->alist.prev) = (node)->alist.next; \
0x108 0x204 0x1a8 0x104 (2) ((node)->alist.next)->alist.prev = (node)->alist.prev;\
(3) (node)->alist.next = (node); \
node1 (4) (node)->alist.prev = &((node)->alist.next); \
0x200 0
} while ( /* */ 0)
0x204 0x100
0x208 0x1a4
QREMOVE(node0, alist);

head node0 node1

0x100 0 0x1a0 0 0x200 0
0x104 0x200 0x1a4 0x1a0 0x204 0x100
0x108 0x204 0x1a8 0x1a4 0x208 0x104

38
 Functions
• Always use function prototypes
int myfunc (char *, int, struct MyStruct *);
int myfunc_noargs (void);
void myfunc_noreturn (int i);
• C and C++ are call by value, copy of parameter passed to function
– C++ permits you to specify pass by reference
– if you want to alter the parameter then pass a pointer to it (or use
references in C++)
• If performance is an issue then use inline functions, generally
better and safer than using a macro. Common convention
– define prototype and function in header or name.i file
– static inline int myinfunc (int i, int j);
– static inline int myinfunc (int i, int j) { ... }

39
 Basic Types and Operators
• Basic data types
– Types: char, int, float and double
– Qualifiers: short, long, unsigned, signed, const
• Constant: 0x1234, 12, “Some string”
• Enumeration:
– Names in different enumerations must be distinct
– enum WeekDay_t {Mon, Tue, Wed, Thur, Fri};
enum WeekendDay_t {Sat = 0, Sun = 4};
• Arithmetic: +, -, *, /, %
– prefix ++i or --i ; increment/decrement before value is used
– postfix i++, i--; increment/decrement after value is used
• Relational and logical: <, >, <=, >=, ==, !=, &&, ||
• Bitwise: &, |, ^ (xor), <<, >>, ~(ones complement)

40
 Operator Precedence (from “C a Reference Manual”, 5 th
Edition)

Tokens

Operator

Class

Precedence

Associates

Tokens

Operator

Class

Precedence

Associates
names, simple tokens primary 16 n/a (type) casts unary 14 right-to-left
literals
* / % multiplicative binary 13 left-to-right
a[k] subscripting postfix left-to-right
+ - additive binary 12 left-to-right
f(...) function call postfix left-to-right
<< >> left, right shift binary 11 left-to-right
. direct selection postfix left-to-right
< <= > >= relational binary 10 left-to-right
-> indirect selection postfix left to right
== != equality/ineq. binary 9 left-to-right
++ -- increment, decrement postfix left-to-right
& bitwise and binary 8 left-to-right
(type){init} compound literal postfix left-to-right
^ bitwise xor binary 7 left-to-right
++ -- increment, decrement prefix 15 right-to-left
| bitwise or binary 6 left-to-right
sizeof size unary right-to-left
&& logical and binary 5 left-to-right
~ bitwise not unary right-to-left
|| logical or binary 4 left-to-right
! logical not unary right-to-left
?: conditional ternary 3 right-to-left
- + negation, plus unary right-to-left
= += -= assignment binary 2 right-to-left
& address of unary right-to-left *= /= %=
* indirection unary right-to-left &= ^= |=
(dereference) <<= >>=

, sequential eval. binary 1 left-to-right

41
 Structs and Unions
• structures
– struct MyPoint {int x, int y};
– typedef struct MyPoint MyPoint_t;
– MyPoint_t point, *ptr;
– point.x = 0;point.y = 10;
– ptr = &point; ptr->x = 12; ptr->y = 40;

• unions
– union MyUnion {int x; MyPoint_t pt; struct {int
3; char c[4]} S;};
– union MyUnion x;
– Can only use one of the elements. Memory will be allocated for
the largest element

42
 Conditional Statements (if/else)
if (a < 10)
printf(“a is less than 10\n”);
else if (a == 10)
printf(“a is 10\n”);
else
printf(“a is greater than 10\n”);
• If you have compound statements then use brackets (blocks)
– if (a < 4 && b > 10) {
c = a * b; b = 0;
printf(“a = %d, a\’s address = 0x%08x\n”, a, (uint32_t)&a);
} else {
c = a + b; b = a;
}
• These two statements are equivalent:
– if (a) x = 3; else if (b) x = 2; else x = 0;
– if (a) x = 3; else {if (b) x = 2; else x = 0;}
• Is this correct?
– if (a) x = 3; else if (b) x = 2;
else (z) x = 0; else x = -2;

43
 Conditional Statements (switch)
int c = 10;
switch (c) {
case 0:
printf(“c is 0\n”);
break;
...
default:
printf(“Unknown value of c\n”);
break;
}
• What if we leave the break statement out?
• Do we need the final break statement on the default case?

44
 Loops
for (i = 0; i < MAXVALUE; i++) {
dowork();
}
while (c != 12) {
dowork();
}
do {
dowork();
} while (c < 12);

• flow control
– break – exit innermost loop
– continue – perform next iteration of loop
• Note, all these forms permit one statement to be executed. By
enclosing in brackets we create a block of statements.

45
 Building your program
• For all labs and programming assignments:
– you must supply a make file
– you must supply a README file that describes the assignment
and results. This must be a text file, no MS word.
– of course the source code and any other libraries or utility
code you used
– you may submit plots, they must be postscript or pdf

46
 make and Makefiles, Overview
• Why use make?
– convenience of only entering compile directives once
– make is smart enough (with your help) to only compile and link modules
that have changed or which depend on files that have changed
– allows you to hide platform dependencies
– promotes uniformity
– simplifies my (and hopefully your) life when testing and verifying your
code
• A makefile contains a set of rules for building a program
target ... : prerequisites ...
command
...
• Static pattern rules.
– each target is matched against target-pattern to derive stem which is
used to determine prereqs (see example)
targets ... : target-pattern : prereq-patterns ...
command
...
47
 Makefiles
• Defining variables
MyOPS := -DWTH
MyDIR ?= /home/fred
MyVar = $(SHELL)
• Using variables
MyFLAGS := $(MyOPS)
• Built-in Variables
– $@ = filename of target
– $< = name of the first prerequisites
• Patterns
– use % character to determine stem
– foo.o matches the pattern %.o with foo as the stem.
– foo.o moo.o : %.o : %.c # says that foo.o depends on foo.c and
moo.o depends on moo.c
48
 Example Makefile for wulib
Makefile.inc Makefile
# Makefile.inc # Project specific
# Contains common definitions include ../Makefile.inc
INCLUDES = ${WUINCLUDES} –I.
MyOS := $(shell uname -s) LIBS = ${WILIBS} ${OSLIBS}
MyID := $(shell whoami) CFLAGS = ${WUCLFAGS} –DWUDEBUG
MyHost := $(shell hostname) CC = ${WUCC}
WARNSTRICT := -W \
-Wstrict- HDRS := util.h
prototypes \ CSRCS := testapp1.c testapp2.c
-Wmissing-prototypes SRCS := util.c callout.c
WARNLIGHT := -Wall COBJS = $(addprefix ${OBJDIR}/, \
WARN := ${WARNLIGHT} $(patsubst %.c,%.o,$(CSRCS)))
ALLFLGS := -D_GNU_SOURCE \ OBJS = $(addprefix ${OBJDIR}/, \
-D_REENTRANT \ $(patsubst %.c,%.o,$(SRCS)))
-D_THREAD_SAFE CMDS = $(addprefix ${OBJDIR}/, $(basename $(CSRCS)))

APPCFLGS = $(ALLFLGS) \ all : $(OBJDIR) $(CMDS)

$(WARN)
install : all
WUCC := gcc
WUCFLAGS := -DMyOS=$(MyOS) \ $(OBJDIR) :
$(OSFLAGS) \ mkdir $(OBJDIR)
$(ALLFLGS) $(WARN)
$(OBJS) $(COBJS) : ${OBJDIR}/%.o : %.c $(HDRS)
WUINCLUDES := ${CC} ${CFLAGS} ${INCLUDES} –o $@ -c $<
WULIBS := -lm
$(CMDS) : ${OBJDIR}/% : ${OBJDIR}/%.o $(OBJS)
ifeq (${MyOS), SunOS) ${CC} ${CFLAGS} -o $@ $@.o ${LIBS}
OSLIBS+= -lrt chmod 0755 $@
endif
clean :
/bin/rm -f $(CMDS) $(OBJS)

49
 Project Documentation
• README file structure
– Section A: Introduction
describe the project, paraphrase the requirements and state your
understanding of the assignments value.
– Section B: Design and Implementation
List all files turned in with a brief description for each. Explain your
design and provide simple psuedo-code for your project. Provide a simple
flow chart of you code and note any constraints, invariants, assumptions
or sources for reused code or ideas.
– Section C: Results
For each project you will be given a list of questions to answer, this is
where you do it. If you are not satisfied with your results explain why
here.
– Section D: Conclusions
What did you learn, or not learn during this assignment. What would you
do differently or what did you do well.

50
 Attacking a Project
• Requirements and scope: Identify specific requirements and or goals.
Also note any design and/or implementation environment
requirements.
– knowing when you are done, or not done
– estimating effort or areas which require more research
– programming language, platform and other development environment
issues
• Approach: How do you plan to solve the problem identified in the first
step. Develop a prototype design and document. Next figure out how
you will verify that you did satisfy the requirements/goals. Designing
the tests will help you to better understand the problem domain and
your proposed solution
• Iterative development: It is good practice to build your project in
small pieces. Testing and learning as you go.
• Final Touches: Put it all together and run the tests identified in the
approach phase. Verify you met requirements. Polish you code and
documentation.
• Turn it in:
51