Main Page | Modules | Namespace List | Class Hierarchy | Class List | Namespace Members | Class Members

Common Definitions
[Foundation Classes]


Namespaces

namespace  Constants
 The constants namespace.
namespace  Exception
 Exception.
namespace  RTTI
 Simplified RunTime Type Identification.

Classes

class  Exception::GeneralException
 General exception class. More...
class  HashFunction< T >
 General Hash Function Template. More...
class  Limits< T >
 Numeric limits class. More...
class  VersionInfo
 Version information class. More...

[NOHEADER]

#define BALL_CREATE_DEEP(name)
 Virtual construction macro.
#define BALL_CREATE(name)
 Virtual construction macro.
#define BALL_DEFINE_CREATE(name)
 Virtual cloning method definition macro.

Type aliases defined by BALL.

These predefined types are used in BALL for portability and comprehensibility.

typedef BALL_INDEX_TYPE Distance
 Distance type.
typedef BALL_SIZE_TYPE Handle
 Handle type.
typedef BALL_INDEX_TYPE Index
 Index type.
typedef BALL_SIZE_TYPE Size
 Size type.
typedef BALL_LONG64_TYPE LongIndex
 Long signed int type.
typedef BALL_ULONG64_TYPE LongSize
 Long unsigned int type.
typedef time_t Time
 Time type.
typedef BALL_SIZE_TYPE HashIndex
 HashIndex type.
typedef BALL_SIZE_TYPE Position
 Position type.
typedef float Real
 Real type.
typedef double DoubleReal
 Double-precision real type.
typedef BALL_SIZE_TYPE Property
 Unnamed property type.
typedef BALL_INDEX_TYPE ErrorCode
 Error code property type.
typedef unsigned char Byte
 Byte type.
typedef BALL_POINTERSIZEUINT_TYPE PointerSizeUInt
 Unsigned int with the same size as a pointer.

Functions

BALL_EXPORT std::ostream & operator<< (std::ostream &os, const Exception::GeneralException &e)
 Output operator for exceptions.
BALL_EXPORT string streamClassName (const std::type_info &t)
 Returns a unique name for a class.

Detailed Description

This chapter contains all relevant information on very fundamental definitions and classes used throughout BALL (part of the Foundation Classes).

Define Documentation

#define BALL_CREATE (  name   ) 
 

Value:

\
  virtual void* create(bool /* deep */ = true, bool empty = false) const\
  {\
    void* ptr;\
    if (empty == true)\
    {\
      ptr = (void*)new name;\
    }\
    else\
    {\
      ptr = (void*)new name(*this);\
    }\
    \
    return ptr;\
  }\
  \
  static void* createDefault()\
  {\
    return static_cast<void*>(new name);\
  }
Virtual construction macro.

This macro is used to define the virtual create method for classes that do not define a copy constructor taking a second argument (boolean, deep or shallow copy). On inclusion of this macro in the public interface of a class, the virtual creation method becomes available. The create method's signature is as follows: virtual void* create(bool deep = true, bool empty = false) const

The create method either creates an empty default object of the class (empty == true) or a copy of the object. The use of the create method requires a (public) default constructor (when creating an empty copy) and a copy constructor taking a reference to an object. The macro also implements a static method createDefault that returns a void pointer to a new instance of name.
Parameters:
name the class name

#define BALL_CREATE_DEEP (  name   ) 
 

Value:

\
  virtual void* create(bool deep = true, bool empty = false) const\
  {\
    void* ptr;\
    if (empty == true)\
    {\
      ptr = (void*)new name;\
    }\
    else\
    {\
      ptr = (void*)new name(*this, deep);\
    }\
    \
    return ptr;\
  }\
  \
  static void* createDefault()\
  {\
    return static_cast<void*>(new name);\
  }
Virtual construction macro.

This macro is used to define the virtual create method. On inclusion of this macro in the public interface of a class, the virtual creation method becomes available. The create method's signature is as follows: virtual void* create(bool deep = true, bool empty = false) const

The create method either creates an empty default object of the class (empty == true) or a copy of the object. The copy is either deep (deep == true) or shallow (deep == false). By default, the create methods returns a pointer to a deep copy of the object. The use of the create method requires a (public) default constructor (when creating an empty copy) or a copy constructor.
The macro also implements a static method createDefault that returns a void pointer to a new instance of name.
Parameters:
name the class name

#define BALL_DEFINE_CREATE (  name   ) 
 

Value:

\
  virtual void* create(bool deep = true, bool empty = false) const;\
  static void* createDefault();
Virtual cloning method definition macro.

If the create method has to be implemented by the user, this macro just defines the create method and the createDefault method. The function signatures are: 

			virtual void* create(bool deep = true, bool empty = false) const;
			static void* createDefault();

Typedef Documentation

typedef unsigned char Byte
 

Byte type.

Use this type to represent byte data (8 bit length). A Byte is always unsigned.

Size: 8 bit
persistent

typedef BALL_INDEX_TYPE Distance
 

Distance type.

Use this type to represent distances in indices. Signed.

Size: 32 bit
persistent

typedef double DoubleReal
 

Double-precision real type.

Use this type to represent double precision floating point numbers.

Size: 64 bit
persistent

typedef BALL_INDEX_TYPE ErrorCode
 

Error code property type.

Use this type to represent (signed) error codes.

Size: 32 bit
persistent
See also:
PropertyManager

typedef BALL_SIZE_TYPE Handle
 

Handle type.

Use this type to represent handles. Handles are used for the non-ambiguous identification of objects (e.g. object derived from Object ). Handles are unsigned.

Size: 32 bit
persistent

typedef BALL_SIZE_TYPE HashIndex
 

HashIndex type.

Use this type to access the result of a hash functions. HashIndex is unsigned.

Size: 32 bit
persistent

typedef BALL_INDEX_TYPE Index
 

Index type.

Use this type to represent indices (e.g. in strings or other sequences). Theses indices may be signed, contrary to the Size type.

Size: 32 bit
persistent

typedef BALL_LONG64_TYPE LongIndex
 

Long signed int type.

Use this type to represent very long (64 bit) indices. Theses indices may be signed, contrary to the LongSize type.

Size: 64 bit
persistent

typedef BALL_ULONG64_TYPE LongSize
 

Long unsigned int type.

Use this type to represent sizes of containers, sequences or alike. Variables of type LongSize are unsigned.

Size: 64 bit
persistent

typedef BALL_POINTERSIZEUINT_TYPE PointerSizeUInt
 

Unsigned int with the same size as a pointer.

Used for internal conversion issues mostly. Size: 32/64 bit (platform dependent)

typedef BALL_SIZE_TYPE Position
 

Position type.

Use this type to represent positions (e.g. in a container) that cannot be negative (contrary to Index ).

Size: 32 bit
persistent

typedef BALL_SIZE_TYPE Property
 

Unnamed property type.

Use this type to represent (unnamed) properties.

Size: 32 bit
persistent
See also:
PropertyManager

typedef float Real
 

Real type.

Use this type to represent standard floating point numbers.

Size: 32 bit
persistent

typedef BALL_SIZE_TYPE Size
 

Size type.

Use this type to represent sizes of containers, sequences or alike. Variables of type Size are unsigned.

Size: 32 bit
persistent

typedef time_t Time
 

Time type.

Use this type to represent a point in time (as a replaecement for time_t).

Size: platform dependent
not persistent

Function Documentation

BALL_EXPORT std::ostream& operator<< (  std::ostream &  os,
const Exception::GeneralException &  e
) 
 

Output operator for exceptions.

All BALL exceptions can be printed to an arbitrary output stream. Information written contains the exception class, the error message, and the location (file, line number). The following code block can thus be used to catch any BALL exceptions and convert them to human readable information: 

				try
				{
					.... // some code which potentially throws an exception
				}
				catch (Exception::GeneralException e)
				{
					Log.error() << "caught exception: " << e << std::endl;
				}

BALL_EXPORT string streamClassName (  const std::type_info &  t  ) 
 

Returns a unique name for a class.

This name contains no blanks. It is usually derived by substituting all blanks in the name (as returned by RTTI::getName()) with underscores ("_"). In the case of gcc, however a name demangling decodes the string first. This function is needed for object persistence.

Parameters:
t the type_info structure as returned by typeid
Returns:
string the coverted class name