Polymorphism:
Polymorphism enables one common interface for many implementations, and for objects to act differently under different circumstances.
C++ supports several kinds of static (compile-time) and dynamic (run-time) polymorphisms. Compile-time polymorphism does not allow for certain run-time decisions, while run-time polymorphism typically incurs a performance penalty.
Static polymorphism:
Function overloading allows programs to declare multiple functions having the same name (but with different arguments). The functions are distinguished by the number and/or types of their formal parameters. Thus, the same function name can refer to different functions depending on the context in which it is used. The type returned by the function is not used to distinguish overloaded functions.
When declaring a function, a programmer can specify default arguments for one or more parameters. Doing so allows the parameters with defaults to optionally be omitted when the function is called, in which case the default arguments will be used. When a function is called with fewer arguments than there are declared parameters, explicit arguments are matched to parameters in left-to-right order, with any unmatched parameters at the end of the parameter list being assigned their default arguments. In many cases, specifying default arguments in a single function declaration is preferable to providing overloaded function definitions with different numbers of parameters.
Templates in C++ provide a sophisticated mechanism for writing generic, polymorphic code. In particular, through the Curiously Recurring Template Pattern it`s possible to implement a form of static polymorphism that closely mimics the syntax for overriding virtual functions. Since C++ templates are type-aware and Turing-complete they can also be used to let the compiler resolve recursive conditionals and generate substantial programs through template metaprogramming.
Dynamic polymorphism:
Inheritance:
Variable pointers (and references) to a base class type in C++ can refer to objects of any derived classes of that type in addition to objects exactly matching the variable type. This allows arrays and other kinds of containers to hold pointers to objects of differing types. Because assignment of values to variables usually occurs at run-time, this is necessarily a run-time phenomenon.
C++ also provides a dynamic_cast operator, which allows the program to safely attempt conversion of an object into an object of a more specific object type (as opposed to conversion to a more general type, which is always allowed). This feature relies on run-time type information (RTTI). Objects known to be of a certain specific type can also be cast to that type with static_cast, a purely compile-time construct which is faster and does not require RTTI.
Virtual member functions:
Ordinarily when a function in a derived class overrides a function in a base class, the function to call is determined by the type of the object. A given function is overridden when there exists no difference, in the number or type of parameters, between two or more definitions of that function. Hence, at compile time it may not be possible to determine the type of the object and therefore the correct function to call, given only a base class pointer; the decision is therefore put off until runtime. This is called dynamic dispatch. Virtual member functions or methods allow the most specific implementation of the function to be called, according to the actual run-time type of the object. In C++, this is commonly done using virtual function tables. If the object type is known, this may be bypassed by prepending a fully qualified class name before the function call, but in general calls to virtual functions are resolved at run time.
In addition to standard member functions, operator overloads and destructors can be virtual. A general rule of thumb is that if any functions in the class are virtual, the destructor should be as well. As the type of an object at its creation is known at compile time, constructors, and by extension copy constructors, cannot be virtual. Nonetheless a situation may arise where a copy of an object needs to be created when a pointer to a derived object is passed as a pointer to a base object. In such a case a common solution is to create a clone() (or similar) function and declare that as virtual. The clone() method creates and returns a copy of the derived class when called.
A member function can also be made "pure virtual" by appending it with = 0 after the closing parenthesis and before the semicolon. Objects cannot be created of a class with a pure virtual function and are called abstract data types. Such abstract data types can only be derived from. Any derived class inherits the virtual function as pure and must provide a non-pure definition of it (and all other pure virtual functions) before objects of the derived class can be created. A program that attempts to create an object of a class with a pure virtual member function or inherited pure virtual member function is ill-formed.
Parsing and processing C++ source code:
It is relatively difficult to write a good C++ parser with classic parsing algorithms such as LALR(1). This is partly because the C++ grammar is not LALR. Because of this, there are very few tools for analyzing or performing non-trivial transformations (e.g., refactoring) of existing code. One way to handle this difficulty is to choose a different syntax, such as Significantly Prettier and Easier C++ Syntax, which is LALR(1) parsable. More powerful parsers, such as GLR parsers, can be substantially simpler (though slower).
Parsing (in the literal sense of producing a syntax tree) is not the most difficult problem in building a C++ processing tool. Such tools must also have the same understanding of the meaning of the identifiers in the program as a compiler might have. Practical systems for processing C++ must then not only parse the source text, but be able to resolve for each identifier precisely which definition applies (e.g. they must correctly handle C++`s complex scoping rules) and what its type is, as well as the types of larger expressions.
Finally, a practical C++ processing tool must be able to handle the variety of C++ dialects used in practice (such as that supported by the GNU Compiler Collection and that of Microsoft`s Visual C++) and implement appropriate analyzers, source code transformers, and regenerate source text. Combining advanced parsing algorithms such as GLR with symbol table construction and program transformation machinery can enable the construction of arbitrary C++ tools.
Compatibility:
Producing a reasonably standards-compliant C++ compiler has proven to be a difficult task for compiler vendors in general. For many years, different C++ compilers implemented the C++ language to different levels of compliance to the standard, and their implementations varied widely in some areas such as partial template specialization. Recent releases of most popular C++ compilers support almost all of the C++ 1998 standard. Apparently there is no implementation that fully supports the entire standard.
One particular point of contention is the export keyword, intended to allow template definitions to be separated from their declarations. The first compiler to implement export was Comeau C/C++, in early 2003 (5 years after the release of the standard); in 2004, the beta compiler of Borland C++ Builder X was also released with export. Both of these compilers are based on the EDG C++ front end. It should also be noted that many C++ books provide example code using the keyword export (for example, Beginning ANSI C++ by Ivor Horton) which will not compile in most compilers, but there is no reference to the problem with the keyword export mentioned. Other compilers such as GCC do not support it at all. Herb Sutter, former convener of the C++ standards committee, recommended that export be removed from future versions of the C++ standard,but finally the decision was made to retain it.
In order to give compiler vendors greater freedom, the C++ standards committee decided not to dictate the implementation of name mangling, exception handling, and other implementation-specific features. The downside of this decision is that object code produced by different compilers is expected to be incompatible. There are, however, third party standards for particular machines or operating systems which attempt to standardize compilers on those platforms (for example C++ ABI); some compilers adopt a secondary standard for these items.