spot  2.11.6
Public Member Functions | Protected Member Functions | List of all members
spot::state Class Referenceabstract

Abstract class for states. More...

#include <spot/twa/twa.hh>

Inheritance diagram for spot::state:
Collaboration diagram for spot::state:

Public Member Functions

virtual int compare (const state *other) const =0
 Compares two states (that come from the same automaton). More...
 
virtual size_t hash () const =0
 Hash a state. More...
 
virtual stateclone () const =0
 Duplicate a state. More...
 
virtual void destroy () const
 Release a state. More...
 

Protected Member Functions

virtual ~state ()
 Destructor. More...
 

Detailed Description

Abstract class for states.

Constructor & Destructor Documentation

◆ ~state()

virtual spot::state::~state ( )
inlineprotectedvirtual

Destructor.

Note that client code should call

s->destroy();

instead of

delete s;

.

Member Function Documentation

◆ clone()

virtual state* spot::state::clone ( ) const
pure virtual

◆ compare()

virtual int spot::state::compare ( const state other) const
pure virtual

Compares two states (that come from the same automaton).

This method returns an integer less than, equal to, or greater than zero if this is found, respectively, to be less than, equal to, or greater than other according to some implicit total order.

This method should not be called to compare states from different automata.

See also
spot::state_ptr_less_than

Implemented in spot::state_product, spot::state_ta_product, spot::twa_graph_state, spot::state_ta_explicit, spot::kripke_graph_state, and spot::set_state.

◆ destroy()

virtual void spot::state::destroy ( ) const
inlinevirtual

Release a state.

Methods from the tgba or twa_succ_iterator always return a new state that you should deallocate with this function. Before Spot 0.7, you had to "delete" your state directly. Starting with Spot 0.7, you should update your code to use this function instead. destroy() usually calls delete, except in subclasses that destroy() to allow better memory management (e.g., no memory allocation for explicit automata).

Reimplemented in spot::state_product, spot::twa_graph_state, spot::state_ta_explicit, and spot::kripke_graph_state.

Referenced by spot::state_unicity_table::is_new(), and spot::state_unicity_table::operator()().

◆ hash()

virtual size_t spot::state::hash ( ) const
pure virtual

Hash a state.

This method returns an integer that can be used as a hash value for this state.

Note that the hash value is guaranteed to be unique for all equal states (in compare()'s sense) for only as long as one of these states exists. So it's OK to use a spot::state as a key in a hash_map because the mere use of the state as a key in the hash will ensure the state continues to exist.

However if you create the state, get its hash key, delete the state, recreate the same state, and get its hash key, you may obtain two different hash keys if the same state were not already used elsewhere. In practice this weird situation can occur only when the state is BDD-encoded, because BDD numbers (used to build the hash value) can be reused for other formulas. That probably doesn't matter, since the hash value is meant to be used in a hash_map, but it had to be noted.

Implemented in spot::state_product, spot::twa_graph_state, spot::set_state, spot::state_ta_product, spot::state_ta_explicit, and spot::kripke_graph_state.


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

Please direct any question, comment, or bug report to the Spot mailing list at spot@lrde.epita.fr.
Generated on Fri Feb 27 2015 10:00:07 for spot by doxygen 1.9.1