rpa::thread_stack< Ftor, Thr >::node Class Reference

#include <synchro.h>

Inheritance diagram for rpa::thread_stack< Ftor, Thr >::node:

List of all members.

Public Types
typedef Ftor::result_type	result_type
Public Member Functions
	node (Ftor &aFtor, Thr &aThr)
Private Attributes
node *	_next
Friends
class	iterator
class	thread_stack

template<class Ftor, class Thr>
class rpa::thread_stack< Ftor, Thr >::node

---

Member Typedef Documentation

template<class Ftor, class Thr>

typedef Ftor::result_type rpa::thread_stack< Ftor, Thr >::node::result_type

Reimplemented from rpa::future< Ftor, Thr >.

---

Constructor & Destructor Documentation

template<class Ftor, class Thr>

rpa::thread_stack< Ftor, Thr >::node::node	(	Ftor &	aFtor,
		Thr &	aThr
	)			[inline]

The constructor launches the sub-thread. We MUST cross a memory barrier before launching the sub-threads : Because of '_refFtor' and '_refThr'.

---

Friends And Related Function Documentation

template<class Ftor, class Thr>

friend class iterator [friend]

template<class Ftor, class Thr>

friend class thread_stack [friend]

---

Member Data Documentation

template<class Ftor, class Thr>

node* rpa::thread_stack< Ftor, Thr >::node::_next [private]

Pointer to the next 'node' object, of the same algorithm execution, by another thread. These 'node' objects are organised into a chained list, created by recursive calls, with one element by sub-thread. All 'node' objects are stored on the stack. Once they are all created, the main thread waits for sub-threads termination, and does reduction in original and natural evaluation order. It is done by 'reduction()' called at the deepest recursive level : This ensures that all temporary 'node' objects created on the stack are still available.

---

The documentation for this class was generated from the following file:

• synchro.h

---