spot::twa_succ_iterator Class Reference

Iterate over the successors of a state. More...

#include <spot/twa/twa.hh>

Inheritance diagram for spot::twa_succ_iterator:

Collaboration diagram for spot::twa_succ_iterator:

Public Member Functions
Iteration
virtual bool	first ()=0
	Position the iterator on the first successor (if any). More...
virtual bool	next ()=0
	Jump to the next successor (if any). More...
virtual bool	done () const =0
	Check whether the iteration is finished. More...
Inspection
virtual const state *	dst () const =0
	Get the destination state of the current edge. More...
virtual bdd	cond () const =0
	Get the condition on the edge leading to this successor. More...
virtual acc_cond::mark_t	acc () const =0
	Get the acceptance mark of the edge leading to this successor. More...

Detailed Description

Iterate over the successors of a state.

This class provides the basic functionality required to iterate over the set of edges leaving a given state. Instance of twa_succ_iterator should normally not be created directly. Instead, they are created by passing a "source" state to twa::succ_iter(), which will create the instance of twa_succ_iterator to iterate over the successors of that state.

This twa_succ_iterator class offers two types of services, offered by two groups of methods. The methods first(), next(), and done() allow iteration over the set of outgoing edges. The methods cond(), acc(), dst(), allow inspecting the current edge.

The twa_succ_iterator is usually subclassed so that iteration methods and accessor methods can be implemented differently in different automata. In particular, this interface allows computing the set of successors on the fly if needed.

The iterator can be used to iterate over all successors in a loop as follows:

for (i->first(); !i->done(); i->next())
  {
    // use i->cond(), i->acc(), i->dst()
  }

If there are n successors, there will be 1 call to first(), n calls to next() and n+1 calls to done(), so a total of 2(n+1) calls to virtual methods just to handle the iteration. For this reason, we usually favor the following more efficient way of performing the same loop:

if (i->first())
  do
    {
      // use i->cond(), i->acc(), i->dst()
    }
  while(i->next());

This loop uses the return value of first() and next() to save n+1 calls to done().

Member Function Documentation

acc()

virtual acc_cond::mark_t spot::twa_succ_iterator::acc ( ) const

pure virtual

Get the acceptance mark of the edge leading to this successor.

Implemented in spot::fair_kripke_succ_iterator, spot::kripke_succ_iterator, spot::ta_explicit_succ_iterator, spot::ta_succ_iterator_product, spot::tgta_succ_iterator_product, spot::taa_succ_iterator, and spot::twa_graph_succ_iterator< Graph >.

cond()

virtual bdd spot::twa_succ_iterator::cond ( ) const

pure virtual

Get the condition on the edge leading to this successor.

This is a Boolean function of atomic propositions.

Implemented in spot::fair_kripke_succ_iterator, spot::kripke_succ_iterator, spot::ta_explicit_succ_iterator, spot::ta_succ_iterator_product, spot::tgta_succ_iterator_product, spot::taa_succ_iterator, and spot::twa_graph_succ_iterator< Graph >.

done()

virtual bool spot::twa_succ_iterator::done ( ) const

pure virtual

Check whether the iteration is finished.

This function should be called after any call to first() or next() and before any enquiry about the current state.

The typical use case of done() is in a for loop such as:

for (s->first(); !s->done(); s->next())
  ...

Returns

false iff the iterator is pointing to a successor.

It is incorrect to call done() if first() hasn't been called before. If done() returns true, it is invalid to call next(), cond(), acc(), or dst().

Implemented in spot::kripke_graph_succ_iterator< Graph >, spot::ta_explicit_succ_iterator, spot::ta_succ_iterator_product, spot::tgta_succ_iterator_product, spot::taa_succ_iterator, and spot::twa_graph_succ_iterator< Graph >.

dst()

virtual const state * spot::twa_succ_iterator::dst ( ) const

pure virtual

Get the destination state of the current edge.

Each call to dst() (even several times on the same edge) creates a new state that has to be destroyed (see state::destroy()) by the caller after it is no longer used.

Note that the same state may occur at different points in the iteration, as different outgoing edges (usually with different labels or acceptance membership) may go to the same state.

Implemented in spot::kripke_graph_succ_iterator< Graph >, spot::ta_explicit_succ_iterator, spot::ta_succ_iterator_product, spot::tgta_succ_iterator_product, spot::taa_succ_iterator, and spot::twa_graph_succ_iterator< Graph >.

first()

virtual bool spot::twa_succ_iterator::first ( )

pure virtual

Position the iterator on the first successor (if any).

This method can be called several times in order to make multiple passes over successors.

Warning

One should always call done() (or better: check the return value of first()) to ensure there is a successor, even after first(). A common trap is to assume that there is at least one successor: this is wrong.

Returns

true iff there is at least one successor

If first() returns false, it is invalid to call next(), cond(), acc(), or dst().

Implemented in spot::kripke_graph_succ_iterator< Graph >, spot::ta_explicit_succ_iterator, spot::ta_succ_iterator_product, spot::tgta_succ_iterator_product, spot::taa_succ_iterator, and spot::twa_graph_succ_iterator< Graph >.

next()

virtual bool spot::twa_succ_iterator::next ( )

pure virtual

Jump to the next successor (if any).

Warning

Again, one should always call done() (or better: check the return value of next()) to ensure there is a successor.

Returns

true if the iterator moved to a new successor, false if the iterator could not move to a new successor.

If next() returns false, it is invalid to call next() again, or to call cond(), acc() or dst().

Implemented in spot::kripke_graph_succ_iterator< Graph >, spot::ta_explicit_succ_iterator, spot::ta_succ_iterator_product, spot::tgta_succ_iterator_product, spot::taa_succ_iterator, and spot::twa_graph_succ_iterator< Graph >.

---

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

• spot/twa/twa.hh