com.ibm.wala.dataflow.IFDS

Interface TabulationProblem<T,P,F>

Type Parameters:

T - type of node in the supergraph

P - type of a procedure (like a box in an RSM)

F - type of factoids propagated when solving this problem

All Known Subinterfaces:

PartiallyBalancedTabulationProblem<T,P,F>

All Known Implementing Classes:

Slicer.SliceProblem

public interface TabulationProblem<T,P,F>

Representation of a Dyck-language graph reachability problem for the tabulation solver.

Special case: if supportsMerge(), then the problem is not really IFDS anymore. (TODO: rename it?). Instead, we perform a merge operation before propagating at every program point. This way, we can implement standard interprocedural dataflow and ESP-style property simulation, and various other things.

Note that at the moment, the data structures in the TabulationSolver are not set up to do merge efficiently. TODO.

See Reps, Horwitz, Sagiv POPL 95

Method Summary

Modifier and Type	Method and Description
TabulationDomain<F,T>	getDomain()
IFlowFunctionMap<T>	getFunctionMap()
IMergeFunction	getMergeFunction() Special case: if supportsMerge(), then the problem is not really IFDS anymore.
ISupergraph<T,P>	getSupergraph()
java.util.Collection<PathEdge<T>>	initialSeeds() Define the set of path edges to start propagation with.

Method Detail

getSupergraph

ISupergraph<T,P> getSupergraph()

getDomain

TabulationDomain<F,T> getDomain()

getFunctionMap

IFlowFunctionMap<T> getFunctionMap()

initialSeeds

java.util.Collection<PathEdge<T>> initialSeeds()

Define the set of path edges to start propagation with.

getMergeFunction

IMergeFunction getMergeFunction()

Special case: if supportsMerge(), then the problem is not really IFDS anymore. (TODO: rename it?). Instead, we perform a merge operation before propagating at every program point. This way, we can implement standard interprocedural dataflow and ESP-style property simulation, and various other things.

Returns:

the merge function, or null if !supportsMerge()