proguard.analysis

Class CallResolver.Builder

java.lang.Object

proguard.analysis.CallResolver.Builder

Enclosing class:

CallResolver

public static class CallResolver.Builder
extends java.lang.Object

Constructor Summary

Constructors

Constructor and Description
Builder(ClassPool programClassPool, ClassPool libraryClassPool, CallGraph callGraph, CallVisitor... visitors)

Method Summary

Modifier and Type	Method and Description
CallResolver	build()
CallResolver.Builder	setArrayValueFactory(ValueFactory arrayValueFactory)
CallResolver.Builder	setClearCallValuesAfterVisit(boolean clearCallValuesAfterVisit) If true, Call.clearValues() will be called after CallVisitor.visitCall(Call).
CallResolver.Builder	setEvaluateAllCode(boolean evaluateAllCode) See PartialEvaluator.Builder#setEvaluateAllCode(boolean).
CallResolver.Builder	setExecutingInvocationUnitBuilder(ExecutingInvocationUnit.Builder executingInvocationUnitBuilder)
CallResolver.Builder	setIgnoreExceptions(boolean ignoreExceptions) If false, exceptions will be taken into account during control flow analysis.
CallResolver.Builder	setIncludeSubClasses(boolean includeSubClasses) If true, virtual calls on class fields, parameters and return values of other methods will take all possible subclasses into account.
CallResolver.Builder	setMaxPartialEvaluations(int maxPartialEvaluations) See PartialEvaluator.Builder#stopAnalysisAfterNEvaluations(int).
CallResolver.Builder	setShouldAnalyzeNextCodeAttribute(java.util.function.Supplier<java.lang.Boolean> shouldAnalyzeNextCodeAttribute) If returns true, the next code attribute will be analyzed.
CallResolver.Builder	setSkipIncompleteCalls(boolean skipIncompleteCalls) If true, any discovered call that would return true for Call.hasIncompleteTarget() will be discarded and not be forwarded to CallVisitor.visitCall(Call).
CallResolver.Builder	setUseDominatorAnalysis(boolean useDominatorAnalysis) If true, a dominator analysis is carried out using the DominatorCalculator for each method, in order to be able to fill the Call.controlFlowDependent flag.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Builder

public Builder(ClassPool programClassPool,
               ClassPool libraryClassPool,
               CallGraph callGraph,
               CallVisitor... visitors)

Method Detail

setClearCallValuesAfterVisit

public CallResolver.Builder setClearCallValuesAfterVisit(boolean clearCallValuesAfterVisit)

If true, Call.clearValues() will be called after CallVisitor.visitCall(Call). This makes it possible to analyze arguments and the return value of calls while still adding them to a CallGraph afterwards, as call graph analysis itself usually only requires the call locations and their targets, not the arguments or return value.

setUseDominatorAnalysis

public CallResolver.Builder setUseDominatorAnalysis(boolean useDominatorAnalysis)

If true, a dominator analysis is carried out using the DominatorCalculator for each method, in order to be able to fill the Call.controlFlowDependent flag.

setEvaluateAllCode

public CallResolver.Builder setEvaluateAllCode(boolean evaluateAllCode)

See PartialEvaluator.Builder#setEvaluateAllCode(boolean).

setIncludeSubClasses

public CallResolver.Builder setIncludeSubClasses(boolean includeSubClasses)

If true, virtual calls on class fields, parameters and return values of other methods will take all possible subclasses into account.

This is necessary for a more complete call graph, because the runtime type of these objects is not controlled by the current method. E.g. a method that declares its return type to be of type A might also return an object of type B in case B extends A. The same is true for class fields and parameters, so in order to really find all potential calls, this circumstance needs to be modeled. For objects of declared type Object this will be skipped, as the fact that every single Java class is a subclass of object would lead to an immense blow-up of the call graph.

setMaxPartialEvaluations

public CallResolver.Builder setMaxPartialEvaluations(int maxPartialEvaluations)

See PartialEvaluator.Builder#stopAnalysisAfterNEvaluations(int).

setShouldAnalyzeNextCodeAttribute

public CallResolver.Builder setShouldAnalyzeNextCodeAttribute(java.util.function.Supplier<java.lang.Boolean> shouldAnalyzeNextCodeAttribute)

If returns true, the next code attribute will be analyzed. Otherwise, the code attribute will be skipped.

setSkipIncompleteCalls

public CallResolver.Builder setSkipIncompleteCalls(boolean skipIncompleteCalls)

If true, any discovered call that would return true for Call.hasIncompleteTarget() will be discarded and not be forwarded to CallVisitor.visitCall(Call).

setArrayValueFactory

public CallResolver.Builder setArrayValueFactory(ValueFactory arrayValueFactory)

Parameters:

arrayValueFactory - should be set either with DetailedArrayValueFactory or ArrayReferenceValueFactory

Returns:

CallResolver.Builder object

setIgnoreExceptions

public CallResolver.Builder setIgnoreExceptions(boolean ignoreExceptions)

If false, exceptions will be taken into account during control flow analysis.

setExecutingInvocationUnitBuilder

public CallResolver.Builder setExecutingInvocationUnitBuilder(ExecutingInvocationUnit.Builder executingInvocationUnitBuilder)

Parameters:

executingInvocationUnitBuilder - a builder for the invocation unit used for particular value analysis.

Returns:

CallResolver.Builder object

build

public CallResolver build()