1. Portability and Safety Mahdi Milani Fard Dec, 2006 Java

2. Outline ● Language Overview  History and design goals ● Design Goals  Portability, Reliability, Safety, Simplicity, Efficiency ● Java Programming Language  Objects in Java, Type system, Generics ● Java's Runtime Environment  Virtual machine, Loader, Bytecodes, Security issues

3. History ● Primarily Designed by James Gosling and others at Sun, 1990 – 95 ● Was called “Oak” at first ● Language for “set-top box” ● Was to be used for embedded computers in home appliances  This was too soon for that time  Java was a side product

4. Design Goals ● Portability  Internet-wide distribution: PC, Unix, Mac ● Reliability  Avoid program crashes and error messages ● Safety  Programmer may be malicious ● Simplicity and familiarity  Appeal to average programmer; less complex than C++ ● Efficiency  Important but secondary

5. Java's Success ● What did make Java so successful?  Portability? ● There are lots of other languages that are as portable as Java. Even C++ has portable standards now.  Reliability? ● Reliable exception handling mechanism is also implemented in C++ and many other languages  Safety? ● Again there are many tools and mechanisms in other languages to support this, such as digital signature of packages and components and access rights.  Simplicity? ● There are many languages easer and simpler than Java, such as Visual Basic.

6. Java's Success ● The true answer:  None ● Tremendous marketing effort by sun Microsystems ● Java is now on you cellphones, in your browser and on your everyday used servers ● Will it keep wining the market?

7. Simplicity ● Almost everything is an object ● All objects on heap, accessed through references ● Omitted troublesome features  No functions, no multiple inheritance, no goto, no operator overloading, few automatic coercions

8. Portability ● Bytecode interpreter on many platforms  Windows, Linux, Mac,... ● Maybe the most important portability feature is Java's portable GUI library  There are many portable GUI libraries for C++, so what's the deal? ● Only few C++ geeks can cross compile such codes under different platforms ● These libraries are not standard, they need to be installed ● Many “cool” platform depended features are not implemented for portability reasons

9. Reliability and Safety ● Typed source ● Typed bytecode language ● Run-time type and bounds checks ● Garbage collection ● Security mechanisms  Sandboxing  Digital signatures

10. Java Programming Language ● Syntax similar to C++ ● Simple and easy to learn and understand ● Fewer confusing syntax and programing features  Better syntactic consistency  Many samples of “responsible design” principle ● Enhanced mechanisms for exception handling ● Embedded mechanisms for documentation  Java Doc

11. Java Classes and Objects ● Syntax similar to C++ ● Object  has fields and methods  is allocated on heap, not run-time stack  accessible through reference (only ptr assignment)  garbage collected (no explicit delete or free) ● There are also primitive types  for efficiency  int, double, boolean ● No header files  A class will be defined only in a single file

12. A Java Sample Class class Point { private int x; protected void setX (int y) { x = y; } public int getX() { return x; } Point(int xval) { //constructor x = xval; } };

13. Object and Class Load and Destruction ● Java guarantees constructor call for each object  Memory allocated  Constructor called to initialize memory  Issues related to inheritance ● We’ll discuss later… ● Static fields of class initialized at class load time  Class loader is responsible for loading classes ● We’ll discuss later… ● Objects are garbage collected  No explicit free, Avoids resulting type errors

14. Encapsulation and Packages ● Every field or method belongs to a class  no global functions of variables ● Every class is part of some package  Package can be unnamed (default)  File declares which package code belongs to PackageA ClassX field method PackageB ClassY field method

15. Visibility and Access ● Four visibility distinctions (both methods and fields)  Public: ● accessible by any class in any package  Package (not declared) ● accessible by any class in the same package  protected ● accessible by subclasses in the any package  private ● accessible by methods of the same class

16. Java Types ● Two general kinds of times  Primitive types – not objects ● Integers, Booleans, etc  Reference types ● Classes, interfaces, arrays ● Static type checking  Every expression has type, determined from its parts ● Dynamic type checking  Downcast checked at run-time, may raise exception  Method invocation (dynamic binding, polymorphism)

17. Subtyping and Inheritance ● Interface  the external view of an object (code) ● Implementation  the internal representation of an object (code) ● Subtyping  relation between interfaces  only inherits the signatures ● Inheritance  relation between implementations  the code is inherited

18. Java's Interfaces ● Using the “impelements” keyword for “interface” types (Only Subtyping) ● Similar to C++ completely abstract classes ● Flexibility  Multiple subtyping can be done for a single class  Allows subtype graph instead of tree  Avoids problems with multiple inheritance of implementations (remember C++ “diamond”) ● Cost  Offset in method lookup table not known at compile  Different bytecodes for method lookup (we'll see)

19. Interfacing Sample interface Shape { public float area(); }; interface Drawable { public void draw(); }; class Circle implements Shape, Drawable { private float r; private Point c; public float draw() { /*implementation*/ } public void draw() { /*implementation*/ } };

20. Java's Inheritance ● Similar to Smalltalk, C++  Subclass inherits from superclass  Using the “extends” keyword: ● class ColorPoint extends Point {... }; ● Single inheritance only ● Every class extends another class  Superclass is Object if no other class named ● Methods of class Object:  getClass  toString  equals  hashCode  clone ...

21. Java's Inheritance (cont) ● Java guarantees constructor call for each object  This must be preserved by inheritance  Subclass constructor must call super constructor ● Explicit (First line of the subclass constructor) ● Implicit (Only when there is a default constructor for the superclass) ● Restrict inheritance  Final classes and methods cannot be redefined ● Important for security

22. Arrays Types ● Automatically defined  Array type T[ ], T[ ][ ],... exists for each class or interface type T  Cannot extend array types (array types are final) ● Treated as reference type  An array variable is a pointer to an array, can be null  Example: Circle[ ] x = new Circle[array_size]  Anonymous array expression: new int[ ] {1,2,3} ● Every array type is a subtype of Object[ ] ● Length of array is not part of its static type

23. Array Subtyping ● Covariance  if S <: T then S[ ] <: T[ ] ● Standard type error class A {…}; class B extends A {…}; B[] bArray = new B[10] A[] aArray = bArray // OK since B[] <: A[] aArray[0] = new A() // compiles // but run-time error // raises ArrayStoreException

24. Java Generics class Stack { void push(Object o) {... } Object pop() {... }... }; String s = "Hello"; Stack st = new Stack();... st.push(s);... s = (String) st.pop(); class Stack { void push(A a) {... } A pop() {... }... }; String s = "Hello"; Stack st = new Stack (); st.push(s);... s = st.pop();

25. Java's Runtime Environment ● Virtual machine ● Loader ● Bytecodes ● Security issues

26. Java Virtual Machine ● Three different issues when you use the term “Java Virtual Machine”:  The abstract specification ● Sun's specification  Concrete implementation ● Sun's implementation ● Open source implementation  Runtime instance ● The process named Java

27. JVM Architecture

28. Class Loader ● Loading: finding and importing the binary data for a type ● Linking:  Verification  Preparation  Resolution ● Initialization: invoking Java code that initializes class variables to their proper starting values.

29. Runtime Areas

30. Heap

31. Method Area ● The fully qualified name of the type ● The fully qualified name of the type’s direct superclass ● Whether or not the type is a class or an interface ● The type’s modifiers (public,…) ● An ordered list of the fully qualified names of any direct superinterfaces ● The constant pool for the type ● Field information ● Method information ● All class (static) variables declared in the type ● A reference to class ClassLoader ● A reference to class Class

32. Activation Record ● The activation record has three parts:  Local variables  Operand stack  Frame data ● The sizes of the local variables and operand stack are determined at compile time

33. Java Bytecode ● Java Class A extends Object { int i void f(int val) { i = val + 1; } ● Bytecode Method void f(int) aload 0 ;object ref this iload 1 ; int val iconst 1 iadd ; add val +1 putfield #4 return val This data area local variables operand stack JVM Activation Record Return addr, exception info, Const pool res.

34. Field and Method Access ● Instruction includes index into constant pool  Constant pool stores symbolic names ● First execution  Use symbolic name to find field or method ● Second execution  Use modified “quick” instruction to simplify search  Putfield_quick 6

35. invokevirtual ● Search for method  find the method entry in the constant pool  pop arguments  find method with the given name and signature using the reference ● Java Object x; … x.equals(”test”); ● ByteCode aload_1 ldc ”test” invokevirtual java/lang/Object/equals

36. Method Lookup 3 5 blue Point object ColorPoint object x mptr x c Point Method table ColorPoint Method table Code for move Code for darken Point p = new ColorPoint(3, 2, “RED”); p.move(2, 3); // (p.mptr[0])(p,2, 3)

37. Bytecode Rewriting: invokevirtual inv_virt_quick vtable offset “A.foo()” Bytecode invokevirtual After search, rewrite bytcode to use fixed offset into the vtable. No search on second execution.

38. invokeinterface ● Interfaces ● Problem with multiple interface ● Solutions:  Multiple method tables  Search

39. Bytecode Rewriting: invokeinterface Cache address of method; check class on second use inv_int_quick Constant pool “A.foo()” Bytecode invokeinterface vtable offset “A.foo()”

40. CPP Approach C object vptr B data vptr A data C data B object A object & C::f0 C-as-A vtbl C-as-B vtbl & B::g0 & C::f   pa, pc pb ● C Extends A,B ● More memory usage ● Less dereferencing C AB

41. Java Security ● Security  Prevent unauthorized use of computational resources ● Java security  A code can read input from careless user or malicious attacker (buffer overflow,...)  A code may be written by careless friend or malicious attacker (trajans, viruses,...) ● Java is designed to reduce many security risks

42. Java Security Mechanisms ● Sandboxing  Run program in restricted environment ● Analogy: child’s sandbox with only safe toys  This term refers to ● Features of loader, verifier, interpreter that restrict program ● Java Security Manager, a special object that acts as access control “gatekeeper” ● Code signing  Use cryptography to establish origin of class file ● This info can be used by security manager

43. Java Sandbox ● Four complementary mechanisms  Class loader ● Separate namespaces for separate class loaders ● Associates protection domain with each class  Verifier ● Checks code consistency  JVM run-time checks ● No unchecked casts or other type errors, No array overflows,... ● Preserves private, protected visibility levels  Security Manager ● Limits the access

44. Why is typing a security feature? ● Java library functions call security manager ● Security manager object answers at run time  Decide if calling code is allowed to do operation  Examine protection domain of calling class ● Signer: organization that signed code before loading ● Location: URL where the Java classes came from  Uses the system policy to decide access permission

45. Exception Handling ● Similar to C++ with a few new mechanism ● Advantages of Exceptions  Separating Error-Handling Code from "Regular" Code  Propagating Errors Up the Call Stack  Grouping and Differentiating Error Types ● Using subtyping and inheritance

46. Exception Handling (cont) ● Differences with C++  Only “throwable” Objects can be thrown  Every exception should either be catched or re- thrown by the method  Method explicitly state the type of exceptions they throw  Many error checking “checks” can be done at compile time ● Compile error is reported if an exception is not catched  Unlike C++ exceptions are heavily used in Java's standard library

47. Exception Handling (cont) method1 { try { call method2; } catch (exception e) { doErrorProcessing; } method2 throws exception { call method3; } method3 throws exception { call readFile; }