Java Chapter Wise Tutorial

Chapter 1: Language Fundamentals (Next: Chapter 2)

The Java programming language has includes five simple arithmetic operators like are + (addition), - (subtraction), * (multiplication), / (division), and % (modulo). The following table summarizes them:
1. Source file's elements (in order)

· Package declaration

· Import statements

· Class definitions

2. Importing packages doesn't recursively import sub-packages.

3. Sub-packages are really different packages, happen to live within an enclosing package. Classes in sub-packages cannot access classes in enclosing package with default access.

4. Comments can appear anywhere. Can't be nested. No matter what type of comments.

5. At most one public class definition per file. This class name should match the file name. If there are more than one public class definitions, compiler will accept the class with the file's name and give an error at the line where the other class is defined.

6. It's not required having a public class definition in a file. Strange, but true. J In this case, the file's name should be different from the names of classes and interfaces (not public obviously).

7. Even an empty file is a valid source file.

8. An identifier must begin with a letter, dollar sign ($) or underscore (_). Subsequent characters may be letters, $, _ or digits.

9. An identifier cannot have a name of a Java keyword. Embedded keywords are OK. true, false and null are literals (not keywords), but they can't be used as identifiers as well.

10. const and goto are reserved words, but not used.

11. Unicode characters can appear anywhere in the source code. The following code is valid.

ch\u0061r a = 'a';

char \u0062 = 'b';

char c = '\u0063';

12. Java has 8 primitive data types.

Data Type	Size (bits)	Initial Value	Min Value	Max Value
boolean	1	false	false	true
byte	8	0	-128 (-2⁷)	127 (2⁷ - 1)
short	16	0	-2¹⁵	2¹⁵ - 1
char	16	'\u0000'	'\u0000' (0)	'\uFFFF' (2¹⁶ - 1)
int	32	0	-2³¹	2³¹ - 1
long	64	0L	-2⁶³	2⁶³ - 1
float	32	0.0F	1.4E-45	3.4028235E38
double	64	0.0	4.9E-324	1.7976931348623157E308

13. All numeric data types are signed. char is the only unsigned integral type.

14. Object reference variables are initialized to null.

15. Octal literals begin with zero. Hex literals begin with 0X or 0x.

16. Char literals are single quoted characters or unicode values (begin with \u).

17. A number is by default an int literal, a decimal number is by default a double literal.

18. 1E-5d is a valid double literal, E2d is not (since it starts with a letter, compiler thinks that it's an identifier)

19. Two types of variables.

a. Member variables

· Accessible anywhere in the class.

· Automatically initialized before invoking any constructor.

· Static variables are initialized at class load time.

· Can have the same name as the class.

b. Automatic variables method local

· Must be initialized explicitly. (Or, compiler will catch it.) Object references can be initialized to null to make the compiler happy. The following code won't compile. Specify else part or initialize the local variable explicitly.

public String testMethod ( int a) {

String tmp;

if ( a > 0 ) tmp = "Positive";

return tmp;

}

· Can have the same name as a member variable, resolution is based on scope.

20. Arrays are Java objects. If you create an array of 5 Strings, there will be 6 objects created.

21. Arrays should be

· Declared. (int[] a; String b[]; Object []c; Size should not be specified now)

· Allocated (constructed). ( a = new int[10]; c = new String[arraysize] )

· Initialized. for (int i = 0; i < a.length; a[i++] = 0)

22. The above three can be done in one step.

int a[] = { 1, 2, 3 }; (or )

int a[] = new int[] { 1, 2, 3 }; But never specify the size with the new statement.

23. Java arrays are static arrays. Size has to be specified at compile time. Array.length returns array's size. (Use Vectors for dynamic purposes).

24. Array size is never specified with the reference variable, it is always maintained with the array object. It is maintained in array.length, which is a final instance variable.

25. Anonymous arrays can be created and used like this: new int[] {1,2,3} or new int[10]

26. Arrays with zero elements can be created. args array to the main method will be a zero element array if no command parameters are specified. In this case args.length is 0.

27. Comma after the last initializer in array declaration is ignored.

int[] i = new int[2] { 5, 10}; // Wrong

int i[5] = { 1, 2, 3, 4, 5}; // Wrong

int[] i[] = {{}, new int[] {} }; // Correct

int i[][] = { {1,2}, new int[2] }; // Correct

int i[] = { 1, 2, 3, 4, } ; // Correct

28. Array indexes start with 0. Index is an int data type.

29. Square brackets can come after datatype or before/after variable name. White spaces are fine. Compiler just ignores them.

30. Arrays declared even as member variables also need to be allocated memory explicitly.

static int a[];

static int b[] = {1,2,3};

public static void main(String s[]) {

System.out.println(a[0]); // Throws a null pointer exception

System.out.println(b[0]); // This code runs fine

System.out.println(a); // Prints 'null'

System.out.println(b); // Prints a string which is returned by toString

}

31. Once declared and allocated (even for local arrays inside methods), array elements are automatically initialized to the default values.

32. If only declared (not constructed), member array variables default to null, but local array variables will not default to null.

33. Java doesn't support multidimensional arrays formally, but it supports arrays of arrays. From the specification - "The number of bracket pairs indicates the depth of array nesting." So this can perform as a multidimensional array. (no limit to levels of array nesting)

34. In order to be run by JVM, a class should have a main method with the following signature.

public static void main(String args[])

static public void main(String[] s)

35. args array's name is not important. args[0] is the first argument. args.length gives no. of arguments.

36. main method can be overloaded.

37. main method can be final.

38. A class with a different main signature or w/o main method will compile. But throws a runtime error.

39. A class without a main method can be run by JVM, if its ancestor class has a main method. (main is just a method and is inherited)

40. Primitives are passed by value.

41. Objects (references) are passed by reference. The object reference itself is passed by value. So, it can't be changed. But, the object can be changed via the reference.

42. Garbage collection is a mechanism for reclaiming memory from objects that are no longer in use, and making the memory available for new objects.

43. An object being no longer in use means that it can't be referenced by any 'active' part of the program.

44. Garbage collection runs in a low priority thread. It may kick in when memory is too low. No guarantee.

45. It's not possible to force garbage collection. Invoking System.gc may start garbage collection process.

46. The automatic garbage collection scheme guarantees that a reference to an object is always valid while the object is in use, i.e. the object will not be deleted leaving the reference "dangling".

47. There are no guarantees that the objects no longer in use will be garbage collected and their finalizers executed at all. gc might not even be run if the program execution does not warrant it. Thus any memory allocated during program execution might remain allocated after program termination, unless reclaimed by the OS or by other means.

48. There are also no guarantees on the order in which the objects will be garbage collected or on the order in which the finalizers are called. Therefore, the program should not make any decisions based on these assumptions.

49. An object is only eligible for garbage collection, if the only references to the object are from other objects that are also eligible for garbage collection. That is, an object can become eligible for garbage collection even if there are references pointing to the object, as long as the objects with the references are also eligible for garbage collection.

50. Circular references do not prevent objects from being garbage collected.

51. We can set the reference variables to null, hinting the gc to garbage collect the objects referred by the variables. Even if we do that, the object may not be gc-ed if it's attached to a listener. (Typical in case of AWT components) Remember to remove the listener first.

52. All objects have a finalize method. It is inherited from the Object class.

53. finalize method is used to release system resources other than memory. (such as file handles and network connections) The order in which finalize methods are called may not reflect the order in which objects are created. Don't rely on it. This is the signature of the finalize method.

protected void finalize() throws Throwable { }

In the descendents this method can be protected or public. Descendents can restrict the exception list that can be thrown by this method.

54. finalize is called only once for an object. If any exception is thrown in finalize, the object is still eligible for garbage collection (at the discretion of gc)

55. gc keeps track of unreachable objects and garbage-collects them, but an unreachable object can become reachable again by letting know other objects of its existence from its finalize method (when called by gc). This 'resurrection' can be done only once, since finalize is called only one for an object.

56. finalize can be called explicitly, but it does not garbage collect the object.

57. finalize can be overloaded, but only the method with original finalize signature will be called by gc.

58. finalize is not implicitly chained. A finalize method in sub-class should call finalize in super class explicitly as its last action for proper functioning. But compiler doesn't enforce this check.

59. System.runFinalization can be used to run the finalizers (which have not been executed before) for the objects eligible for garbage collection.

Chapter 2 Operators and assignments (Next: Chapter 3)

The Java programming language has included five simple arithmetic operators like + (addition), - (subtraction), * (multiplication), / (division)
1. Unary operators

1.1 Increment and Decrement operators ++ --

We have postfix and prefix notation. In post-fix notation value of the variable/expression is modified after the value is taken for the execution of statement. In prefix notation, value of the variable/expression is modified before the value is taken for the execution of statement.

x = 5; y = 0; y = x++; Result will be x = 6, y = 5

x = 5; y = 0; y = ++x; Result will be x = 6, y = 6

Implicit narrowing conversion is done, when applied to byte, short or char.

1.2 Unary minus and unary plus + -

+ has no effect than to stress positivity.

- negates an expression's value. (2's complement for integral expressions)

1.3 Negation !

Inverts the value of a boolean expression.

1.4 Complement ~

Inverts the bit pattern of an integral expression. (1's complement - 0s to 1s and 1s to 0s)

Cannot be applied to non-integral types.

1.5 Cast ()

Persuades compiler to allow certain assignments. Extensive checking is done at compile and runtime to ensure type-safety.

2. Arithmetic operators - *, /, %, +, -

· Can be applied to all numeric types.

· Can be applied to only the numeric types, except '+' - it can be applied to Strings as well.

· All arithmetic operations are done at least with 'int'. (If types are smaller, promotion happens. Result will be of a type at least as wide as the wide type of operands)

· Accuracy is lost silently when arithmetic overflow/error occurs. Result is a nonsense value.

· Integer division by zero throws an exception.

· % - reduce the magnitude of LHS by the magnitude of RHS. (continuous subtraction)

· % - sign of the result entirely determined by sign of LHS

· 5 % 0 throws an ArithmeticException.

· Floating point calculations can produce NaN (square root of a negative no) or Infinity ( division by zero). Float and Double wrapper classes have named constants for NaN and infinities.

· NaN's are non-ordinal for comparisons. x == Float.NaN won't work. Use Float.IsNaN(x) But equals method on wrapper objects(Double or Float) with NaN values compares Nan's correctly.

· Infinities are ordinal. X == Double.POSITIVE_INFINITY will give expected result.

· + also performs String concatenation (when any operand in an expression is a String). The language itself overloads this operator. toString method of non-String object operands are called to perform concatenation. In case of primitives, a wrapper object is created with the primitive value and toString method of that object is called. ("Vel" + 3 will work.)

· Be aware of associativity when multiple operands are involved.

System.out.println( 1 + 2 + "3" ); // Prints 33

System.out.println( "1" + 2 + 3 ); // Prints 123

3. Shift operators - <<, >>, >>>

· << performs a signed left shift. 0 bits are brought in from the right. Sign bit (MSB) is preserved. Value becomes old value * 2 ^ x where x is no of bits shifted.

· >> performs a signed right shift. Sign bit is brought in from the left. (0 if positive, 1 if negative. Value becomes old value / 2 ^ x where x is no of bits shifted. Also called arithmetic right shift.

· >>> performs an unsigned logical right shift. 0 bits are brought in from the left. This operator exists since Java doesn't provide an unsigned data type (except char). >>> changes the sign of a negative number to be positive. So don't use it with negative numbers, if you want to preserve the sign. Also don't use it with types smaller than int. (Since types smaller than int are promoted to an int before any shift operation and the result is cast down again, so the end result is unpredictable.)

· Shift operators can be applied to only integral types.

· -1 >> 1 is -1, not 0. This differs from simple division by 2. We can think of it as shift operation rounding down.

· 1 << 31 will become the minimum value that an int can represent. (Value becomes negative, after this operation, if you do a signed right shift sign bit is brought in from the left and the value remains negative.)

· Negative numbers are represented in two's complement notation. (Take one's complement and add 1 to get two's complement)

· Shift operators never shift more than the number of bits the type of result can have. ( i.e. int 32, long 64) RHS operand is reduced to RHS % x where x is no of bits in type of result.

int x;

x = x >> 33; // Here actually what happens is x >> 1

4. Comparison operators - all return boolean type.

4.1 Ordinal comparisons - <, <=, > , >=

· Only operate on numeric types. Test the relative value of the numeric operands.

· Arithmetic promotions apply. char can be compared to float.

4.2 Object type comparison - instanceof

· Tests the class of an object at runtime. Checking is done at compile and runtime same as the cast operator.

· Returns true if the object denoted by LHS reference can be cast to RHS type.

· LHS should be an object reference expression, variable or an array reference.

· RHS should be a class (abstract classes are fine), an interface or an array type, castable to LHS object reference. Compiler error if LHS & RHS are unrelated.

· Can't use java.lang.Class or its String name as RHS.

· Returns true if LHS is a class or subclass of RHS class

· Returns true if LHS implements RHS interface.

· Returns true if LHS is an array reference and of type RHS.

· x instanceof Component[] - legal.

· x instanceof [] - illegal. Can't test for 'any array of any type'

· Returns false if LHS is null, no exceptions are thrown.

· If x instanceof Y is not allowed by compiler, then Y y = (Y) x is not a valid cast expression. If x instanceof Y is allowed and returns false, the above cast is valid but throws a ClassCastException at runtime. If x instanceof Y returns true, the above cast is valid and runs fine.

4.3 Equality comparisons - ==, !=

· For primitives it's a straightforward value comparison. (promotions apply)

· For object references, this doesn't make much sense. Use equals method for meaningful comparisons. (Make sure that the class implements equals in a meaningful way, like for X.equals(Y) to be true, Y instance of X must be true as well)

· For String literals, == will return true, this is because of compiler optimization.

5. Bit-wise operators - &, ^, |

· Operate on numeric and boolean operands.

· & - AND operator, both bits must be 1 to produce 1.

· | - OR operator, any one bit can be 1 to produce 1.

· ^ - XOR operator, any one bit can be 1, but not both, to produce 1.

· In case of booleans true is 1, false is 0.

· Can't cast any other type to boolean.

6. Short-circuit logical operators - &&, ||

· Operate only on boolean types.

· RHS might not be evaluated (hence the name short-circuit), if the result can be determined only by looking at LHS.

· false && X is always false.

· true || X is always true.

· RHS is evaluated only if the result is not certain from the LHS.

· That's why there's no logical XOR operator. Both bits need to be known to calculate the result.

· Short-circuiting doesn't change the result of the operation. But side effects might be changed. (i.e. some statements in RHS might not be executed, if short-circuit happens. Be careful)

7. Ternary operator

· Format a = x ? b : c ;

· x should be a boolean expression.

· Based on x, either b or c is evaluated. Both are never evaluated.

· b will be assigned to a if x is true, else c is assigned to a.

· b and c should be assignment compatible to a.

· b and c are made identical during the operation according to promotions.

8. Assignment operators.

· Simple assignment =.

· op= calculate and assign operators extended assignment operators.

· *=, /=, %=, +=, -=

· x += y means x = x + y. But x is evaluated only once. Be aware.

· Assignment of reference variables copies the reference value, not the object body.

· Assignment has value, value of LHS after assignment. So a = b = c = 0 is legal. c = 0 is executed first, and the value of the assignment (0) assigned to b, then the value of that assignment (again 0) is assigned to a.

· Extended assignment operators do an implicit cast. (Useful when applied to byte, short or char)

byte b = 10;

b = b + 10; // Won't compile, explicit cast required since the expression evaluates to an int

b += 10; // OK, += does an implicit cast from int to byte

9. General

· In Java, No overflow or underflow of integers happens. i.e. The values wrap around. Adding 1 to the maximum int value results in the minimum value.

· Always keep in mind that operands are evaluated from left to right, and the operations are executed in the order of precedence and associativity.

· Unary Postfix operators and all binary operators (except assignment operators) have left to right assoiciativity.

· All unary operators (except postfix operators), assignment operators, ternary operator, object creation and cast operators have right to left assoiciativity.

· Inspect the following code.

public class Precedence {

final public static void main(String args[]) {

int i = 0;

i = i++;

i = i++;

i = i++;

System.out.println(i); // prints 0, since = operator has the lowest precedence.

int array[] = new int[5];

int index = 0;

array[index] = index = 3; // 1st element gets assigned to 3, not the 4th element

for (int c = 0; c < array.length; c++)

System.out.println(array[c]);

System.out.println("index is " + index); // prints 3

}

}

Type of Operators	Operators	Associativity
Postfix operators	[] . (parameters) ++ --	Left to Right
Prefix Unary operators	++ -- + - ~ !	Right to Left
Object creation and cast	new (type)	Right to Left
Multiplication/Division/Modulus	* / %	Left to Right
Addition/Subtraction	+ -	Left to Right
Shift	>> >>> <<	Left to Right
Relational	< <= > >= instanceof	Left to Right
Equality	== !=	Left to Right
Bit-wise/Boolean AND	&	Left to Right
Bit-wise/Boolean XOR	^	Left to Right
Bit-wise/Boolean OR	\|	Left to Right
Logical AND (Short-circuit or Conditional)	&&	Left to Right
Logical OR (Short-circuit or Conditional)	\|\|	Left to Right
Ternary	? :	Right to Left
Assignment	= += -= *= /= %= <<= >>= >>>= &= ^= \|=	Right to Left

Chapter 3 - Modifiers

1. Modifiers are Java keywords that provide information to compiler about the nature of the code, data and classes.

2. Access modifiers - public, protected, private

· Only applied to class level variables. Method variables are visible only inside the method.

· Can be applied to class itself (only to inner classes declared at class level, no such thing as protected or private top level class)

· Can be applied to methods and constructors.

· If a class is accessible, it doesn't mean, the members are also accessible. Members' accessibility determines what is accessible and what is not. But if the class is not accessible, the members are not accessible, even though they are declared public.

· If no access modifier is specified, then the accessibility is default package visibility. All classes in the same package can access the feature. It's called as friendly access. But friendly is not a Java keyword. Same directory is same package in Java's consideration.

· 'private' means only the class can access it, not even sub-classes. So, it'll cause access denial to a sub-class's own variable/method.

· These modifiers dictate, which classes can access the features. An instance of a class can access the private features of another instance of the same class.

· 'protected' means all classes in the same package (like default) and sub-classes in any package can access the features. But a subclass in another package can access the protected members in the super-class via only the references of subclass or its subclasses. A subclass in the same package doesn't have this restriction. This ensures that classes from other packages are accessing only the members that are part of their inheritance hierarchy.

· Methods cannot be overridden to be more private. Only the direction shown in following figure is permitted from parent classes to sub-classes.

private à friendly (default) à protected à public

Parent classes Sub-classes

3. final

· final features cannot be changed.

· The final modifier applies to classes, methods, and variables.

· final classes cannot be sub-classed.

· You can declare a variable in any scope to be final.

· You may, if necessary, defer initialization of a final local variable. Simply declare the local variable and initialize it later (for final instance variables. You must initialize them at the time of declaration or in constructor).

· final variables cannot be changed (result in a compile-time error if you do so )

· final methods cannot be overridden.

· Method arguments marked final are read-only. Compiler error, if trying to assign values to final arguments inside the method.

· Member variables marked final are not initialized by default. They have to be explicitly assigned a value at declaration or in an initializer block. Static finals must be assigned to a value in a static initializer block, instance finals must be assigned a value in an instance initializer or in every constructor. Otherwise the compiler will complain.

· A blank final is a final variable whose declaration lacks an initializer.

· Final variables that are not assigned a value at the declaration and method arguments that are marked final are called blank final variables. They can be assigned a value at most once.

· Local variables can be declared final as well.

· If a final variable holds a reference to an object, then the state of the object may be changed by operations on the object, but the variable will always refer to the same object.

· This applies also to arrays, because arrays are objects; if a final variable holds a reference to an array, then the components of the array may be changed by operations on the array, but the variable will always refer to the same array

· A blank final instance variable must be definitely assigned at the end of every constructor of the class in which it is declared; otherwise a compile-time error occurs.

· A class can be declared final if its definition is complete and no subclasses are desired or required.

· A compile-time error occurs if the name of a final class appears in the extends clause of another class declaration; this implies that a final class cannot have any subclasses.

· A compile-time error occurs if a class is declared both final and abstract, because the implementation of such a class could never be completed.

· Because a final class never has any subclasses, the methods of a final class are never overridden

4. abstract

· Can be applied to classes and methods.

· For deferring implementation to sub-classes.

· Opposite of final, final can't be sub-classed, abstract must be sub-classed.

· A class should be declared abstract,

1. if it has any abstract methods.

2. if it doesn't provide implementation to any of the abstract methods it inherited

3. if it doesn't provide implementation to any of the methods in an interface that it says implementing.

· Just terminate the abstract method signature with a ';', curly braces will give a compiler error.

· A class can be abstract even if it doesn't have any abstract methods.

5. static

· Can be applied to nested classes, methods, variables, free floating code-block (static initializer)

· Static variables are initialized at class load time. A class has only one copy of these variables.

· Static methods can access only static variables. (They have no this)

· Access by class name is a recommended way to access static methods/variables.

· Static initializer code is run at class load time.

· Static methods may not be overridden to be non-static.

· Non-static methods may not be overridden to be static.

· Abstract methods may not be static.

· Local variables cannot be declared as static.

· Actually, static methods are not participating in the usual overriding mechanism of invoking the methods based on the class of the object at runtime. Static method binding is done at compile time, so the method to be invoked is determined by the type of reference variable rather than the actual type of the object it holds at runtime.

Let's say a sub-class has a static method which 'overrides' a static method in a parent class. If you have a reference variable of parent class type and you assign a child class object to that variable and invoke the static method, the method invoked will be the parent class method, not the child class method. The following code explains this.

public class StaticOverridingTest {

public static void main(String s[]) {

Child c = new Child();

c.doStuff(); // This will invoke Child.doStuff()

Parent p = new Parent();

p.doStuff(); // This will invoke Parent.doStuff()

p = c;

p.doStuff(); // This will invoke Parent.doStuff(), rather than Child.doStuff()

}

class Parent {

static int x = 100;

public static void doStuff() {

System.out.println("In Parent..doStuff");

System.out.println(x);

}

class Child extends Parent {

static int x = 200;

public static void doStuff() {

System.out.println("In Child..doStuff");

System.out.println(x);

}

6. native

· Can be applied to methods only. (static methods also)

· Written in a non-Java language, compiled for a single machine target type.

· Java classes use lot of native methods for performance and for accessing hardware Java is not aware of.

· Native method signature should be terminated by a ';', curly braces will provide a compiler error.

· native doesn't affect access qualifiers. Native methods can be private.

· Can pass/return Java objects from native methods.

· System.loadLibrary is used in static initializer code to load native libraries. If the library is not loaded when the static method is called, an UnsatisfiedLinkError is thrown.

7. transient

· Can be applied to class level variables only.(Local variables cannot be declared transient)

· Transient variables may not be final or static.(But compiler allows the declaration, since it doesn't do any harm. Variables marked transient are never serialized. Static variables are not serialized anyway.)

· Not stored as part of object's persistent state, i.e. not written out during serialization.

· Can be used for security.

8. synchronized

· Can be applied to methods or parts of methods only.

· Used to control access to critical code in multi-threaded programs.

9. volatile

· Can be applied to variables only.

· Can be applied to static variables.

· Cannot be applied to final variables.

· Declaring a variable volatile indicates that it might be modified asynchronously, so that all threads will get the correct value of the variable.

· Used in multi-processor environments.

Modifier	Class	Inner classes (Except local and anonymous classes)	Variable	Method		Constructor	Free floating Code block
public	Y	Y	Y	Y	Y		N
protected	N	Y	Y	Y	Y		N
(friendly) No access modifier	Y	Y (OK for all)	Y	Y	Y		N
private	N	Y	Y	Y	Y		N
final	Y	Y (Except anonymous classes)	Y	Y	N		N
abstract	Y	Y (Except anonymous classes)	N	Y	N		N
static	N	Y	Y	Y	N		Y (static initializer)
native	N	N	N	Y	N		N
transient	N	N	Y	N	N		N
synchronized	N	N	N	Y	N		Y (part of method, also need to specify an object on which a lock should be obtained)
volatile	N	N	Y	N	N		N

Chapter 4 Converting and Casting

Unary Numeric Promotion

Contexts:

· Operand of the unary arithmetic operators + and -

· Operand of the unary integer bit-wise complement operator ~

· During array creation, for example new int[x], where the dimension expression x must evaluate to an int value.

· Indexing array elements, for example table['a'], where the index expression must evaluate to an int value.

· Individual operands of the shift operators.

Binary numeric promotion

Contexts:

· Operands of arithmetic operators *, / , %, + and -

· Operands of relational operators <, <= , > and >=

· Numeric Operands of equality operators == and !=

· Integer Operands of bit-wise operators &, ^ and |

Conversion of Primitives

1. 3 types of conversion - assignment conversion, method call conversion and arithmetic promotion

2. boolean may not be converted to/from any non-boolean type.

3. Widening conversions accepted. Narrowing conversions rejected.

4. byte, short can't be converted to char and vice versa. ( but can be cast )

5. Arithmetic promotion

5.1 Unary operators

· if the operand is byte, short or char {

convert it to int;

}

else {

do nothing; no conversion needed;

}

5.2 Binary operators

· if one operand is double {

all double; convert the other operand to double;

}

else if one operand is float {

all float; convert the other operand to float;

}

else if one operand is long {

all long; convert the other operand to long;

}

else {

all int; convert all to int;

}

6. When assigning a literal value to a variable, the range of the variable's data type is checked against the value of the literal and assignment is allowed or compiler will produce an error.

char c = 3; // this will compile, even though a numeric literal is by default an int since the range of char will accept the value

int a = 3;

char d = a; // this won't compile, since we're assigning an int to char

char e = -1; // this also won't compile, since the value is not in the range of char

float f = 1.3; // this won't compile, even though the value is within float range. Here range is not important, but precision is. 1.3 is by default a double, so a specific cast or f = 1.3f will work.

float f = 1/3; // this will compile, since RHS evaluates to an int.

Float f = 1.0 / 3.0; // this won't compile, since RHS evaluates to a double.

7. Also when assigning a final variable to a variable, even if the final variable's data type is wider than the variable, if the value is within the range of the variable an implicit conversion is done.

byte b;

final int a = 10;

b = a; // Legal, since value of 'a' is determinable and within range of b

final int x = a;

b = x; // Legal, since value of 'x' is determinable and within range of b

int y;

final int z = y;

b = z; // Illegal, since value of 'z' is not determinable

8. Method call conversions always look for the exact data type or a wider one in the method signatures. They will not do narrowing conversions to resolve methods, instead we will get a compile error.

Here is the figure of allowable primitive conversion.

byte à short à int à long à float à double

char

Casting of Primitives

9. Needed with narrowing conversions. Use with care - radical information loss. Also can be used with widening conversions, to improve the clarity of the code.

10. Can cast any non-boolean type to another non-boolean type.

11. Cannot cast a boolean or to a boolean type.

Conversion of Object references

12. Three types of reference variables to denote objects - class, interface or array type.

13. Two kinds of objects can be created - class or array.

14. Two types of conversion - assignment and method call.

15. Permitted if the direction of the conversion is 'up' the inheritance hierarchy. Means that types can be assigned/substituted to only super-types - super-classes or interfaces. Not the other way around, explicit casting is needed for that.

16. Interfaces can be used as types when declaring variables, so they participate in the object reference conversion. But we cannot instantiate an interface, since it is abstract and doesn't provide any implementation. These variables can be used to hold objects of classes that implement the interface. The reason for having interfaces as types may be, I think, several unrelated classes may implement the same interface and if there's a need to deal with them collectively one way of treating them may be an array of the interface type that they implement.

17. Primitive arrays can be converted to only the arrays of the same primitive type. They cannot be converted to another type of primitive array. Only object reference arrays can be converted / cast.

18. Primitive arrays can be converted to an Object reference, but not to an Object[] reference. This is because all arrays (primitive arrays and Object[]) are extended from Object.

Casting of Object references

19. Allows super-types to be assigned to subtypes. Extensive checks done both at compile and runtime. At compile time, class of the object may not be known, so at runtime if checks fail, a ClassCastException is thrown.

20. Cast operator, instanceof operator and the == operator behave the same way in allowing references to be the operands of them. You cannot cast or apply instanceof or compare unrelated references, sibling references or any incompatible references.

Compile-time Rules

· When old and new types are classes, one class must be the sub-class of the other.

· When old and new types are arrays, both must contain reference types and it must be legal to cast between those types (primitive arrays cannot be cast, conversion possible only between same type of primitive arrays).

· We can always cast between an interface and a non-final object.

Run-time rules

· If new type is a class, the class of the expression being converted must be new type or extend new type.

· If new type is an interface, the class of the expression being converted must implement the interface.

An Object reference can be converted to: (java.lang.Object)

· an Object reference

· a Cloneable interface reference, with casting, with runtime check

· any class reference, with casting, with runtime check

· any array referenece, with casting, with runtime check

· any interface reference, with casting, with runtime check

A Class type reference can be converted to:

· any super-class type reference, (including Object)

· any sub-class type reference, with casting, with runtime check

· an interface reference, if the class implements that interface

· any interface reference, with casting, with runtime check (except if the class is final and doesn't implement the interface)

An Interface reference can be converted to:

· an Object reference

· a super-interface reference

· any interface/class reference with casting, with runtime check (except if the class is final and doesn't implement the interface)

A Primitive Array reference can be converted to:

· an Object reference

· a Cloneable interface reference

· a primitive array reference of the same type

An Object Array reference can be converted to:

· an Object reference

· a Cloneable interface reference

· a super-class Array reference, including an Object Array reference

· any sub-class Array reference with casting, with runtime check

Chapter 5 Flow Control and Exceptions

Unreachable statements produce a compile-time error.

while (false) { x = 3; } // won't compile

for (;false;) { x =3; } // won't compile

if (false) {x = 3; } // will compile, to provide the ability to conditionally compile the code.

· Local variables already declared in an enclosing block, therefore visible in a nested block cannot be re-declared inside the nested block.

· A local variable in a block may be re-declared in another local block, if the blocks are disjoint.

· Method parameters cannot be re-declared.

1. Loop constructs

· 3 constructs - for, while, do

· All loops are controlled by a boolean expression.

· In while and for, the test occurs at the top, so if the test fails at the first time, body of the loop might not be executed at all.

· In do, test occurs at the bottom, so the body is executed at least once.

· In for, we can declare multiple variables in the first part of the loop separated by commas, also we can have multiple statements in the third part separated by commas.

· In the first section of for statement, we can have a list of declaration statements or a list of expression statements, but not both. We cannot mix them.

· All expressions in the third section of for statement will always execute, even if the first expression makes the loop condition false. There is no short -circuit here.

2. Selection Statements

· if takes a boolean arguments. Parenthesis required. else part is optional. else if structure provides multiple selective branching.

· switch takes an argument of byte, short, char or int.(assignment compatible to int)

· case value should be a constant expression that can be evaluated at compile time.

· Compiler checks each case value against the range of the switch expression's data type. The following code won't compile.

byte b;

switch (b) {

case 200: // 200 not in range of byte

default:

}

· We need to place a break statement in each case block to prevent the execution to fall through other case blocks. But this is not a part of switch statement and not enforced by the compiler.

· We can have multiple case statements execute the same code. Just list them one by one.

· default case can be placed anywhere. It'll be executed only if none of the case values match.

· switch can be nested. Nested case labels are independent, don't clash with outer case labels.

· Empty switch construct is a valid construct. But any statement within the switch block should come under a case label or the default case label.

3. Branching statements

· break statement can be used with any kind of loop or a switch statement or just a labeled block.

· continue statement can be used with only a loop (any kind of loop).

· Loops can have labels. We can use break and continue statements to branch out of multiple levels of nested loops using labels.

· Names of the labels follow the same rules as the name of the variables.(Identifiers)

· Labels can have the same name, as long as they don't enclose one another.

· There is no restriction against using the same identifier as a label and as the name of a package, class, interface, method, field, parameter, or local variable.

4. Exception Handling

· An exception is an event that occurs during the execution of a program that disrupts the normal flow of instructions.

· There are 3 main advantages for exceptions:

1. Separates error handling code from "regular" code

2. Propagating errors up the call stack (without tedious programming)

3. Grouping error types and error differentiation

· An exception causes a jump to the end of try block. If the exception occurred in a method called from a try block, the called method is abandoned.

· If there's a catch block for the occurred exception or a parent class of the exception, the exception is now considered handled.

· At least one 'catch' block or one 'finally' block must accompany a 'try' statement. If all 3 blocks are present, the order is important. (try/catch/finally)

· finally and catch can come only with try, they cannot appear on their own.

· Regardless of whether or not an exception occurred or whether or not it was handled, if there is a finally block, it'll be executed always. (Even if there is a return statement in try block).

· System.exit() and error conditions are the only exceptions where finally block is not executed.

· If there was no exception or the exception was handled, execution continues at the statement after the try/catch/finally blocks.

· If the exception is not handled, the process repeats looking for next enclosing try block up the call hierarchy. If this search reaches the top level of the hierarchy (the point at which the thread was created), then the thread is killed and message stack trace is dumped to System.err.

· Use throw new xxxException() to throw an exception. If the thrown object is null, a NullPointerException will be thrown at the handler.

· If an exception handler re-throws an exception (throw in a catch block), same rules apply. Either you need to have a try/catch within the catch or specify the entire method as throwing the exception that's being re-thrown in the catch block. Catch blocks at the same level will not handle the exceptions thrown in a catch block - it needs its own handlers.

· The method fillInStackTrace() in Throwable class throws a Throwable object. It will be useful when re-throwing an exception or error.

· The Java language requires that methods either catch or specify all checked exceptions that can be thrown within the scope of that method.

· All objects of type java.lang.Exception are checked exceptions. (Except the classes under java.lang.RuntimeException) If any method that contains lines of code that might throw checked exceptions, compiler checks whether you've handled the exceptions or you've declared the methods as throwing the exceptions. Hence the name checked exceptions.

· If there's no code in try block that may throw exceptions specified in the catch blocks, compiler will produce an error. (This is not the case for super-class Exception)

· Java.lang.RuntimeException and java.lang.Error need not be handled or declared.

· An overriding method may not throw a checked exception unless the overridden method also throws that exception or a super-class of that exception. In other words, an overriding method may not throw checked exceptions that are not thrown by the overridden method. If we allow the overriding methods in sub-classes to throw more general exceptions than the overridden method in the parent class, then the compiler has no way of checking the exceptions the sub-class might throw. (If we declared a parent class variable and at runtime it refers to sub-class object) This violates the concept of checked exceptions and the sub-classes would be able to by-pass the enforced checks done by the compiler for checked exceptions. This should not be allowed.

Here is the exception hierarchy.

Object

Throwable
| |

| |

| Error

Exception-->ClassNotFoundException, ClassNotSupportedException, IllegalAccessException, InstantiationException, IterruptedException, NoSuchMethodException, RuntimeException, AWTException, IOException

RuntimeException-->EmptyStackException, NoSuchElementException, ArithmeticException, ArrayStoreException, ClassCastException, IllegalArgumentException, IllegalMonitorStateException,

IndexOutOfBoundsException, NegativeArraySizeException, NullPointerException, SecurityException.

IllegalArgumentException-->IllegalThreadStateException, NumberFormatException

IndexOutOfBoundsException-->ArrayIndexOutOfBoundsException, StringIndexOutOfBoundsException

IOException-->EOFException, FileNotFoundException, InterruptedIOException, UTFDataFormatException, MalformedURLException, ProtocolException, SockException, UnknownHostException, UnknownServiceException.

Chapter 6 - Objects and Classes (Part 1)

Part 2 Part 3 Part 4

Implementing OO relationships

· "is a" relationship is implemented by inheritance (extends keyword)

· "has a" relationship is implemented by providing the class with member variables.

Overloading and Overriding

· Overloading is an example of polymorphism. (operational / parametric)

· Overriding is an example of runtime polymorphism (inclusive)

· A method can have the same name as another method in the same class, provided it forms either a valid overload or override

Overloading	Overriding
Signature has to be different. Just a difference in return type is not enough.	Signature has to be the same. (including the return type)
Accessibility may vary freely.	Overriding methods cannot be more private than the overridden methods.
Exception list may vary freely.	Overriding methods may not throw more checked exceptions than the overridden methods.
Just the name is reused. Methods are independent methods. Resolved at compile-time based on method signature.	Related directly to sub-classing. Overrides the parent class method. Resolved at run-time based on type of the object.
Can call each other by providing appropriate argument list.	Overriding method can call overridden method by super.methodName(), this can be used only to access the immediate super-class's method. super.super won't work. Also, a class outside the inheritance hierarchy can't use this technique.
Methods can be static or non-static. Since the methods are independent, it doesn't matter. But if two methods have the same signature, declaring one as static and another as non-static does not provide a valid overload. It's a compile time error.	static methods don't participate in overriding, since they are resolved at compile time based on the type of reference variable. *A static method in a sub-class can't use 'super' (for the same reason that it can't use 'this' for)* *Remember that a static method can't be overridden to be non-static and a non-static method can't be overridden to be static*. In other words, a static method and a non-static method cannot have the same name and signature (if signatures are different, it would have formed a valid overload)
There's no limit on number of overloaded methods a class can have.	Each parent class method may be overridden at most once in any sub-class. (That is, you cannot have two identical methods in the same class)

· Variables can also be overridden, it's known as shadowing or hiding. But, member variable references are resolved at compile-time. So at the runtime, if the class of the object referred by a parent class reference variable, is in fact a sub-class having a shadowing member variable, only the parent class variable is accessed, since it's already resolved at compile time based on the reference variable type. Only methods are resolved at run-time.

public class Shadow {

public static void main(String s[]) {

S1 s1 = new S1();

S2 s2 = new S2();

System.out.println(s1.s); // prints S1

System.out.println(s1.getS()); // prints S1

System.out.println(s2.s); // prints S2

System.out.println(s2.getS()); // prints S2

s1 = s2;

System.out.println(s1.s); // prints S1, not S2 -

// since variable is resolved at compile time

System.out.println(s1.getS()); // prints S2 -

// since method is resolved at run time

}

class S1 {

public String s = "S1";

public String getS() {

return s;

}

class S2 extends S1{

public String s = "S2";

public String getS() {

return s;

}

In the above code, if we didn't have the overriding getS() method in the sub-class and if we call the method from sub-class reference variable, the method will return only the super-class member variable value. For explanation, see the following points.

· Also, methods access variables only in context of the class of the object they belong to. If a sub-class method calls explicitly a super class method, the super class method always will access the super-class variable. Super class methods will not access the shadowing variables declared in subclasses because they don't know about them. (When an object is created, instances of all its super-classes are also created.) But the method accessed will be again subject to dynamic lookup. It is always decided at runtime which implementation is called. (Only static methods are resolved at compile-time)

public class Shadow2 {

String s = "main";

public static void main(String s[]) {

S2 s2 = new S2();

s2.display(); // Produces an output - S1, S2

S1 s1 = new S1();

System.out.println(s1.getS()); // prints S1

System.out.println(s2.getS()); // prints S1 - since super-class method

// always accesses super-class variable

}

class S1 {

String s = "S1";

public String getS() {

return s;

}

void display() {

System.out.println(s);

}

class S2 extends S1{

String s = "S2";

void display() {

super.display(); // Prints S1

System.out.println(s); // prints S2

}

· With OO languages, the class of the object may not be known at compile-time (by virtue of inheritance). JVM from the start is designed to support OO. So, the JVM insures that the method called will be from the real class of the object (not with the variable type declared). This is accomplished by virtual method invocation (late binding). Compiler will form the argument list and produce one method invocation instruction - its job is over. The job of identifying and calling the proper target code is performed by JVM.

· JVM knows about the variable's real type at any time since when it allocates memory for an object, it also marks the type with it. Objects always know 'who they are'. This is the basis of instanceof operator.

· Sub-classes can use super keyword to access the shadowed variables in super-classes. This technique allows for accessing only the immediate super-class. super.super is not valid. But casting the 'this' reference to classes up above the hierarchy will do the trick. By this way, variables in super-classes above any level can be accessed from a sub-class, since variables are resolved at compile time, when we cast the 'this' reference to a super-super-class, the compiler binds the super-super-class variable. But this technique is not possible with methods since methods are resolved always at runtime, and the method gets called depends on the type of object, not the type of reference variable. So it is not at all possible to access a method in a super-super-class from a subclass.

public class ShadowTest {

public static void main(String s[]){

new STChild().demo();

}

class STGrandParent {

double wealth = 50000.00;

public double getWealth() {

System.out.println("GrandParent-" + wealth);

return wealth;

}

class STParent extends STGrandParent {

double wealth = 100000.00;

public double getWealth() {

System.out.println("Parent-" + wealth);

return wealth;

}

class STChild extends STParent {

double wealth = 200000.00;

public double getWealth() {

System.out.println("Child-" + wealth);

return wealth;

}

public void demo() {

getWealth(); // Calls Child method

super.getWealth(); // Calls Parent method

// Compiler error, GrandParent method cannot be accessed

//super.super.getWealth();

// Calls Child method, due to dynamic method lookup

((STParent)this).getWealth();

// Calls Child method, due to dynamic method lookup

((STGrandParent)this).getWealth();

System.out.println(wealth); // Prints Child wealth

System.out.println(super.wealth); // Prints Parent wealth

// Prints Parent wealth

System.out.println(((STParent)(this)).wealth);

// Prints GrandParent wealth

System.out.println(((STGrandParent)(this)).wealth);

}

· An inherited method, which was not abstract on the super-class, can be declared abstract in a sub-class (thereby making the sub-class abstract). There is no restriction.

· In the same token, a subclass can be declared abstract regardless of whether the super-class was abstract or not.

· Private members are not inherited, but they do exist in the sub-classes. Since the private methods are not inherited, they cannot be overridden. A method in a subclass with the same signature as a private method in the super-class is essentially a new method, independent from super-class, since the private method in the super-class is not visible in the sub-class.

public class PrivateTest {

public static void main(String s[]){

new PTSuper().hi(); // Prints always Super

new PTSub().hi(); // Prints Super when subclass doesn't have hi method

// Prints Sub when subclass has hi method

PTSuper sup;

sup = new PTSub();

sup.hi(); // Prints Super when subclass doesn't have hi method

// Prints Sub when subclass has hi method

}

class PTSuper {

public void hi() { // Super-class implementation always calls superclass hello

hello();

}

private void hello() { // This method is not inherited by subclasses, but exists in them.

// Commenting out both the methods in the subclass show this.

// The test will then print "hello-Super" for all three calls

// i.e. Always the super-class implementations are called

System.out.println("hello-Super");

}

class PTSub extends PTSuper {

public void hi() { // This method overrides super-class hi, calls subclass hello

try {

hello();

}

catch(Exception e) {}

}

void hello() throws Exception { // This method is independent from super-class hello

// Evident from, it's allowed to throw Exception

System.out.println("hello-Sub");

}

· Private methods are not overridden, so calls to private methods are resolved at compile time and not subject to dynamic method lookup. See the following example.

public class Poly {

public static void main(String args[]) {

PolyA ref1 = new PolyC();

PolyB ref2 = (PolyB)ref1;

System.out.println(ref2.g()); // This prints 1

// If f() is not private in PolyB, then prints 2

}

class PolyA {

private int f() { return 0; }

public int g() { return 3; }

}

class PolyB extends PolyA {

private int f() { return 1; }

public int g() { return f(); }

}

class PolyC extends PolyB {

public int f() { return 2; }

}

Chapter 6 - Objects and Classes (Part 2)

Part 1 Part 3 Part 4

Constructors and Sub-classing
· Constructors are not inherited as normal methods, they have to be defined in the class itself.

· If you define no constructors at all, then the compiler provides a default constructor with no arguments. Even if, you define one constructor, this default is not provided.

· We can't compile a sub-class if the immediate super-class doesn't have a no argument default constructor, and sub-class constructors are not calling super or this explicitly (and expect the compiler to insert an implicit super() call )

· A constructor can call other overloaded constructors by 'this (arguments)'. If you use this, it must be the first statement in the constructor. This construct can be used only from within a constructor.

· A constructor can't call the same constructor from within. Compiler will say ' recursive constructor invocation'

· A constructor can call the parent class constructor explicitly by using 'super (arguments)'. If you do this, it must be first the statement in the constructor. This construct can be used only from within a constructor.

· Obviously, we can't use both this and super in the same constructor. If compiler sees a this or super, it won't insert a default call to super().

· Constructors can't have a return type. A method with a class name, but with a return type is not considered a constructor, but just a method by compiler. Expect trick questions using this.

· Constructor body can have an empty return statement. Though void cannot be specified with the constructor signature, empty return statement is acceptable.

· Only modifiers that a constructor can have are the accessibility modifiers.

· Constructors cannot be overridden, since they are not inherited.

· Initializers are used in initialization of objects and classes and to define constants in interfaces. These initializers are :

1. Static and Instance variable initializer expressions.

§ Literals and method calls to initialize variables.

§ Static variables can be initialized only by static method calls.

§ Cannot pass on the checked exceptions. Must catch and handle them.

2. Static initializer blocks.

§ Used to initialize static variables and load native libraries.

§ Cannot pass on the checked exceptions. Must catch and handle them.

3. Instance initializer blocks.

§ Used to factor out code that is common to all the constructors.

§ Also useful with anonymous classes since they cannot have constructors.

§ All constructors must declare the uncaught checked exceptions, if any.

§ Instance Initializers in anonymous classes can throw any exception.

· In all the initializers, forward referencing of variables is not allowed. Forward referencing of methods is allowed.

· Order of code execution (when creating an object) is a bit tricky.

1. static variables initialization.

2. static initializer block execution. (in the order of declaration, if multiple blocks found)

3. constructor header ( super or this - implicit or explicit )

4. instance variables initialization / instance initializer block(s) execution

5. rest of the code in the constructor

Chapter 6 - Objects and Classes (Part 3)

Part 1 Part 2 Part 4

Interfaces:

· All methods in an interface are implicitly public, abstract, and never static.

· All variables in an interface are implicitly static, public, final. They cannot be transient or volatile. A class can shadow the variables it inherits from an interface, with its own variables.

· A top-level interface itself cannot be declared as static or final since it doesn't make sense.

· Declaring parameters to be final is at method's discretion, this is not part of method signature.

· Same case with final, synchronized, native. Classes can declare the methods to be final, synchronized or native whereas in an interface they cannot be specified like that. (These are implementation details, interface need not worry about this)

· But classes cannot implement an interface method with a static method.

· If an interface specifies an exception list for a method, then the class implementing the interface need not declare the method with the exception list. (Overriding methods can specify sub-set of overridden method's exceptions, here none is a sub-set). But if the interface didn't specify any exception list for a method, then the class cannot throw any exceptions.

· All interface methods should have public accessibility when implemented in class.

· Interfaces cannot be declared final, since they are implicitly abstract.

· A class can implement two interfaces that have a method with the same signature or variables with the same name.

Inner Classes

· A class can be declared in any scope. Classes defined inside of other classes are known as nested classes. There are four categories of nested classes.

1. Top-level nested classes / interfaces

· Declared as a class member with static modifier.

· Just like other static features of a class. Can be accessed / instantiated without an instance of the outer class. Can access only static members of outer class. Can't access instance variables or methods.

· Very much like any-other package level class / interface. Provide an extension to packaging by the modified naming scheme at the top level.

· Classes can declare both static and non-static members.

· Any accessibility modifier can be specified.

· Interfaces are implicitly static (static modifier also can be specified). They can have any accessibility modifier. There are no non-static inner, local or anonymous interfaces.

2. Non-static inner classes

· Declared as a class member without static.

· An instance of a non-static inner class can exist only with an instance of its enclosing class. So it always has to be created within a context of an outer instance.

· Just like other non-static features of a class. Can access all the features (even private) of the enclosing outer class. Have an implicit reference to the enclosing instance.

· Cannot have any static members.

· Can have any access modifier.

3. Local classes

· Defined inside a block (could be a method, a constructor, a local block, a static initializer or an instance initializer). Cannot be specified with static modifier.

· Cannot have any access modifier (since they are effectively local to the block)

· Cannot declare any static members.(Even declared in a static context)

· Can access all the features of the enclosing class (because they are defined inside the method of the class) but can access only final variables defined inside the method (including method arguments). This is because the class can outlive the method, but the method local variables will go out of scope - in case of final variables, compiler makes a copy of those variables to be used by the class. (New meaning for final)

· Since the names of local classes are not visible outside the local context, references of these classes cannot be declared outside. So their functionality could be accessed only via super-class references (either interfaces or classes). Objects of those class types are created inside methods and returned as super-class type references to the outside world. This is the reason that they can only access final variables within the local block. That way, the value of the variable can be always made available to the objects returned from the local context to outside world.

· Cannot be specified with static modifier. But if they are declared inside a static context such as a static method or a static initializer, they become static classes. They can only access static members of the enclosing class and local final variables. But this doesn't mean they cannot access any non-static features inherited from super classes. These features are their own, obtained via the inheritance hierarchy. They can be accessed normally with 'this' or 'super'.

4. Anonymous classes

· Anonymous classes are defined where they are constructed. They can be created wherever a reference expression can be used.

· Anonymous classes cannot have explicit constructors. Instance initializers can be used to achieve the functionality of a constructor.

· Typically used for creating objects on the fly.

· Anonymous classes can implement an interface (implicit extension of Object) or explicitly extend a class. Cannot do both.

Syntax: new interface name() { } or new class name() { }

· Keywords implements and extends are not used in anonymous classes.

· Abstract classes can be specified in the creation of an anonymous class. The new class is a concrete class, which automatically extends the abstract class.

· Discussion for local classes on static/non-static context, accessing enclosing variables, and declaring static variables also holds good for anonymous classes. In other words, anonymous classes cannot be specified with static, but based on the context, they could become static classes. In any case, anonymous classes are not allowed to declare static members. Based on the context, non-static/static features of outer classes are available to anonymous classes. Local final variables are always available to them.

· One enclosing class can have multiple instances of inner classes.

· Inner classes can have synchronous methods. But calling those methods obtains the lock for inner object only not the outer object. If you need to synchronize an inner class method based on outer object, outer object lock must be obtained explicitly. Locks on inner object and outer object are independent.

· Nested classes can extend any class or can implement any interface. No restrictions.

· All nested classes (except anonymous classes) can be abstract or final.

· Classes can be nested to any depth. Top-level static classes can be nested only within other static top-level classes or interfaces. Deeply nested classes also have access to all variables of the outer-most enclosing class (as well the immediate enclosing class's)

· Member inner classes can be forward referenced. Local inner classes cannot be.

· An inner class variable can shadow an outer class variable. In this case, an outer class variable can be referred as (outerclassname.this.variablename).

· Outer class variables are accessible within the inner class, but they are not inherited. They don't become members of the inner class. This is different from inheritance. (Outer class cannot be referred using 'super', and outer class variables cannot be accessed using 'this')

· An inner class variable can shadow an outer class variable. If the inner class is sub-classed within the same outer class, the variable has to be qualified explicitly in the sub-class. To fully qualify the variable, use classname.this.variablename. If we don't correctly qualify the variable, a compiler error will occur. (Note that this does not happen in multiple levels of inheritance where an upper-most super-class's variable is silently shadowed by the most recent super-class variable or in multiple levels of nested inner classes where an inner-most class's variable silently shadows an outer-most class's variable. Problem comes only when these two hierarchy chains (inheritance and containment) clash.)

· If the inner class is sub-classed outside of the outer class (only possible with top-level nested classes) explicit qualification is not needed (it becomes regular class inheritance)

Chapter 6 - Objects and Classes (Part 4)

Part 1 Part 2 Part 3

// Example 1

public class InnerInnerTest {

public static void main(String s[]) {

new Outer().new Inner().new InnerInner().new InnerInnerInner().doSomething();

new Outer().new InnerChild().doSomething();

new Outer2().new Inner2().new InnerInner2().doSomething();

new InnerChild2().doSomething();

}

class Outer {

String name = "Vel";

class Inner {

String name = "Sharmi";

class InnerInner {

class InnerInnerInner {

public void doSomething() {

// No problem in accessing without full qualification,

// inner-most class variable shadows the outer-most class variable

System.out.println(name); // Prints "Sharmi"

System.out.println(Outer.this.name); // Prints "Vel", explicit reference to Outer

// error, variable is not inherited from the outer class, it can be just accessible

// System.out.println(this.name);

// System.out.println(InnerInner.this.name);

// System.out.println(InnerInnerInner.this.name);

// error, super cannot be used to access outer class.

// super will always refer the parent, in this case Object

// System.out.println(super.name);

System.out.println(Inner.this.name); // Prints "Sharmi", Inner has declared 'name'

}

/* This is an inner class extending an inner class in the same scope */

class InnerChild extends Inner {

public void doSomething() {

// compiler error, explicit qualifier needed

// 'name' is inherited from Inner, Outer's 'name' is also in scope

// System.out.println(name);

System.out.println(Outer.this.name); // prints "Vel", explicit reference to Outer

System.out.println(super.name); // prints "Sharmi", Inner has declared 'name'

System.out.println(this.name); // prints "Sharmi", name is inherited by InnerChild

}

class Outer2 {

static String name = "Vel";

static class Inner2 {

static String name = "Sharmi";

class InnerInner2 {

public void doSomething() {

System.out.println(name); // prints "Sharmi", inner-most hides outer-most

System.out.println(Outer2.name); // prints "Vel", explicit reference to Outer2's static variable

// System.out.println(this.name); // error, 'name' is not inherited

// System.out.println(super.name); // error, super refers to Object

}

/* This is a stand-alone class extending an inner class */

class InnerChild2 extends Outer2.Inner2 {

public void doSomething() {

System.out.println(name); // prints "Sharmi", Inner2's name is inherited

System.out.println(Outer2.name); // prints "Vel", explicit reference to Outer2's static variable

System.out.println(super.name); // prints "Sharmi", Inner2 has declared 'name'

System.out.println(this.name); // prints "Sharmi", name is inherited by InnerChild2

}

// Example 2

public class InnerTest2 {

public static void main(String s[]) {

new OuterClass().doSomething(10, 20);

// This is legal

// OuterClass.InnerClass ic = new OuterClass().new InnerClass();

// ic.doSomething();

// Compiler error, local inner classes cannot be accessed from outside

// OuterClass.LocalInnerClass lic = new OuterClass().new LocalInnerClass();

// lic.doSomething();

new OuterClass().doAnonymous();

}

class OuterClass {

final int a = 100;

private String secret = "Nothing serious";

public void doSomething(int arg, final int fa) {

final int x = 100;

int y = 200;

System.out.println(this.getClass() + " - in doSomething");

System.out.print("a = " + a + " secret = " + secret + " arg = " + arg + " fa = " + fa);

System.out.println(" x = " + x + " y = " + y);

// Compiler error, forward reference of local inner class

// new LocalInnerClass().doSomething();

abstract class AncestorLocalInnerClass { } // inner class can be abstract

final class LocalInnerClass extends AncestorLocalInnerClass { // can be final

public void doSomething() {

System.out.println(this.getClass() + " - in doSomething");

System.out.print("a = " + a );

System.out.print(" secret = " + secret);

// System.out.print(" arg = " + arg); // Compiler error, accessing non-final argument

System.out.print(" fa = " + fa);

System.out.println(" x = " + x);

// System.out.println(" y = " + y); // Compiler error, accessing non-final variable

}

new InnerClass().doSomething(); // forward reference fine for member inner class

new LocalInnerClass().doSomething();

}

abstract class AncestorInnerClass { }

interface InnerInterface { final int someConstant = 999;} // inner interface

class InnerClass extends AncestorInnerClass implements InnerInterface {

public void doSomething() {

System.out.println(this.getClass() + " - in doSomething");

System.out.println("a = " + a + " secret = " + secret + " someConstant = " + someConstant);

}

public void doAnonymous() {

// Anonymous class implementing the inner interface

System.out.println((new InnerInterface() { }).someConstant);

// Anonymous class extending the inner class

( new InnerClass() {

public void doSomething() {

secret = "secret is changed";

super.doSomething();

}

} ).doSomething();

}

Entity	Declaration Context	Accessibility Modifiers	Outer instance	Direct Access to enclosing context	Defines static or non-static members
Package level class	As package member	Public or default	No	N/A	Both static and non-static
Top level nested class (static)	As static class member	All	No	Static members in enclosing context	Both static and non-static
Non static inner class	As non-static class member	All	Yes	All members in enclosing context	Only non-static
Local class (non-static)	In block with non-static context	None	Yes	All members in enclosing context + local final variables	Only non-static
Local class (static)	In block with static context	None	No	Static members in enclosing context + local final variables	Only non-static
Anonymous class (non-static)	In block with non-static context	None	Yes	All members in enclosing context + local final variables	Only non-static
Anonymous class (static)	In block with static context	None	No	Static members in enclosing context + local final variables	Only non-static
Package level interface	As package member	Public or default	No	N/A	Static variables and non-static method prototypes
Top level nested interface (static)	As static class member	All	No	Static members in enclosing context	Static variables and non-static method prototypes

Chapter 7 Threads

· Java is fundamentally multi-threaded.

· Every thread corresponds to an instance of java.lang.Thread class or a sub-class.

· A thread becomes eligible to run, when its start() method is called. Thread scheduler co-ordinates between the threads and allows them to run.

· When a thread begins execution, the scheduler calls its run method.

Signature of run method - public void run ()

· When a thread returns from its run method (or stop method is called - deprecated in 1.2), its dead. It cannot be restarted, but its methods can be called. (it's just an object no more in a running state)

· If start is called again on a dead thread, IllegalThreadStateException is thrown.

· When a thread is in running state, it may move out of that state for various reasons. When it becomes eligible for execution again, thread scheduler allows it to run.

· There are two ways to implement threads.

1. Extend Thread class

· Create a new class, extending the Thread class.

· Provide a public void run method, otherwise empty run in Thread class will be executed.

· Create an instance of the new class.

· Call start method on the instance (don't call run - it will be executed on the same thread)

2. Implement Runnable interface

· Create a new class implementing the Runnable interface.

· Provide a public void run method.

· Create an instance of this class.

· Create a Thread, passing the instance as a target - new Thread(object)

· Target should implement Runnable, Thread class implements it, so it can be a target itself.

· Call the start method on the Thread.

· JVM creates one user thread for running a program. This thread is called main thread. The main method of the class is called from the main thread. It dies when the main method ends. If other user threads have been spawned from the main thread, program keeps running even if main thread dies. Basically a program runs until all the user threads (non-daemon threads) are dead.

· A thread can be designated as a daemon thread by calling setDaemon(boolean) method. This method should be called before the thread is started, otherwise IllegalThreadStateException will be thrown.

· A thread spawned by a daemon thread is a daemon thread.

· Threads have priorities. Thread class have constants MAX_PRIORITY (10), MIN_PRIORITY (1), NORM_PRIORITY (5)

· A newly created thread gets its priority from the creating thread. Normally it'll be NORM_PRIORITY.

· getPriority and setPriority are the methods to deal with priority of threads.

· Java leaves the implementation of thread scheduling to JVM developers. Two types of scheduling can be done.

1. Pre-emptive Scheduling.

Ways for a thread to leave running state -

· It can cease to be ready to execute ( by calling a blocking i/o method)

· It can get pre-empted by a high-priority thread, which becomes ready to execute.

· It can explicitly call a thread-scheduling method such as wait or suspend.

· Solaris JVM's are pre-emptive.

· Windows JVM's were pre-emptive until Java 1.0.2

2. Time-sliced or Round Robin Scheduling

· A thread is only allowed to execute for a certain amount of time. After that, it has to contend for the CPU (virtual CPU, JVM) time with other threads.

· This prevents a high-priority thread mono-policing the CPU.

· The drawback with this scheduling is - it creates a non-deterministic system - at any point in time, you cannot tell which thread is running and how long it may continue to run.

· Macintosh JVM's

· Windows JVM's after Java 1.0.2

· Different states of a thread:

1. Yielding

· Yield is a static method. Operates on current thread.

· Moves the thread from running to ready state.

· If there are no threads in ready state, the yielded thread may continue execution, otherwise it may have to compete with the other threads to run.

· Run the threads that are doing time-consuming operations with a low priority and call yield periodically from those threads to avoid those threads locking up the CPU.

2. Sleeping

· Sleep is also a static method.

· Sleeps for a certain amount of time. (passing time without doing anything and w/o using CPU)

· Two overloaded versions - one with milliseconds, one with milliseconds and nanoseconds.

· Throws an InterruptedException. (must be caught)

· After the time expires, the sleeping thread goes to ready state. It may not execute immediately after the time expires. If there are other threads in ready state, it may have to compete with those threads to run. The correct statement is the sleeping thread would execute some time after the specified time period has elapsed.

· If interrupt method is invoked on a sleeping thread, the thread moves to ready state. The next time it begins running, it executes the InterruptedException handler.

3. Suspending

· Suspend and resume are instance methods and are deprecated in 1.2

· A thread that receives a suspend call, goes to suspended state and stays there until it receives a resume call on it.

· A thread can suspend it itself, or another thread can suspend it.

· But, a thread can be resumed only by another thread.

· Calling resume on a thread that is not suspended has no effect.

· Compiler won't warn you if suspend and resume are successive statements, although the thread may not be able to be restarted.

4. Blocking

· Methods that are performing I/O have to wait for some occurrence in the outside world to happen before they can proceed. This behavior is blocking.

· If a method needs to wait an indeterminable amount of time until some I/O takes place, then the thread should graciously step out of the CPU. All Java I/O methods behave this way.

· A thread can also become blocked, if it failed to acquire the lock of a monitor.

5. Waiting

· wait, notify and notifyAll methods are not called on Thread, they're called on Object. Because the object is the one which controls the threads in this case. It asks the threads to wait and then notifies when its state changes. It's called a monitor.

· Wait puts an executing thread into waiting state.(to the monitor's waiting pool)

· Notify moves one thread in the monitor's waiting pool to ready state. We cannot control which thread is being notified. notifyAll is recommended.

· NotifyAll moves all threads in the monitor's waiting pool to ready.

· These methods can only be called from synchronized code, or an IllegalMonitorStateException will be thrown. In other words, only the threads that obtained the object's lock can call these methods.

Chapter 8

Locks, Monitors and Synchronization

· Every object has a lock (for every synchronized code block). At any moment, this lock is controlled by at most one thread.

· A thread that wants to execute an object's synchronized code must acquire the lock of the object. If it cannot acquire the lock, the thread goes into blocked state and comes to ready only when the object's lock is available.

· When a thread, which owns a lock, finishes executing the synchronized code, it gives up the lock.

· Monitor (a.k.a Semaphore) is an object that can block and revive threads, an object that controls client threads. Asks the client threads to wait and notifies them when the time is right to continue, based on its state. In strict Java terminology, any object that has some synchronized code is a monitor.

· 2 ways to synchronize:

1. Synchronize the entire method

· Declare the method to be synchronized - very common practice.

· Thread should obtain the object's lock.

2. Synchronize part of the method

· Have to pass an arbitrary object which lock is to be obtained to execute the synchronized code block (part of a method).

· We can specify "this" in place object, to obtain very brief locking - not very common.

· wait - points to remember

§ calling thread gives up CPU

§ calling thread gives up the lock

§ calling thread goes to monitor's waiting pool

§ wait also has a version with timeout in milliseconds. Use this if you're not sure when the current thread will get notified, this avoids the thread being stuck in wait state forever.

· notify - points to remember

§ one thread gets moved out of monitor's waiting pool to ready state

§ notifyAll moves all the threads to ready state

§ Thread gets to execute must re-acquire the lock of the monitor before it can proceed.

· Note the differences between blocked and waiting.

Blocked	Waiting
Thread is waiting to get a lock on the monitor. (or waiting for a blocking i/o method)	Thread has been asked to wait. (by means of wait method)
Caused by the thread tried to execute some synchronized code. (or a blocking i/o method)	The thread already acquired the lock and executed some synchronized code before coming across a wait call.
Can move to ready only when the lock is available. ( or the i/o operation is complete)	Can move to ready only when it gets notified (by means of notify or notifyAll)

· Points for complex models:

1. Always check monitor's state in a while loop, rather than in an if statement.

2. Always call notifyAll, instead of notify.

· Class locks control the static methods.

· wait and sleep must be enclosed in a try/catch for InterruptedException.

· A single thread can obtain multiple locks on multiple objects (or on the same object)

· A thread owning the lock of an object can call other synchronous methods on the same object. (this is another lock) Other threads can't do that. They should wait to get the lock.

· Non-synchronous methods can be called at any time by any thread.

· Synchronous methods are re-entrant. So they can be called recursively.

· Synchronized methods can be overrided to be non-synchronous. Synchronized behavior affects only the original class.

· Locks on inner/outer objects are independent. Getting a lock on outer object doesn't mean getting the lock on an inner object as well, that lock should be obtained separately.

· wait and notify should be called from synchronized code. This ensures that while calling these methods the thread always has the lock on the object. If you have wait/notify in non-synchronized code compiler won't catch this. At runtime, if the thread doesn't have the lock while calling these methods, an IllegalMonitorStateException is thrown.

· Deadlocks can occur easily. e.g, Thread A locked Object A and waiting to get a lock on Object B, but Thread B locked Object B and waiting to get a lock on Object A. They'll be in this state forever.

· It's the programmer's responsibility to avoid the deadlock. Always get the locks in the same order.

· While 'suspended', the thread keeps the locks it obtained - so suspend is deprecated in 1.2

· Use of stop is also deprecated; instead use a flag in run method. Compiler won't warn you, if you have statements after a call to stop, even though they are not reachable.