All Downloads are FREE. Search and download functionalities are using the official Maven repository.

soot.dava.toolkits.base.AST.structuredAnalysis.StructuredAnalysis Maven / Gradle / Ivy

Go to download 

package soot.dava.toolkits.base.AST.structuredAnalysis;

/*-
 * #%L
 * Soot - a J*va Optimization Framework
 * %%
 * Copyright (C) 2005 Nomair A. Naeem
 * %%
 * This program is free software: you can redistribute it and/or modify
 * it under the terms of the GNU Lesser General Public License as
 * published by the Free Software Foundation, either version 2.1 of the
 * License, or (at your option) any later version.
 * 
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Lesser Public License for more details.
 * 
 * You should have received a copy of the GNU General Lesser Public
 * License along with this program.  If not, see
 * .
 * #L%
 */

import java.util.ArrayList;
import java.util.HashMap;
import java.util.Iterator;
import java.util.List;
import java.util.Map;

import soot.Local;
import soot.Value;
import soot.dava.internal.AST.ASTAggregatedCondition;
import soot.dava.internal.AST.ASTCondition;
import soot.dava.internal.AST.ASTDoWhileNode;
import soot.dava.internal.AST.ASTForLoopNode;
import soot.dava.internal.AST.ASTIfElseNode;
import soot.dava.internal.AST.ASTIfNode;
import soot.dava.internal.AST.ASTLabeledBlockNode;
import soot.dava.internal.AST.ASTLabeledNode;
import soot.dava.internal.AST.ASTMethodNode;
import soot.dava.internal.AST.ASTNode;
import soot.dava.internal.AST.ASTStatementSequenceNode;
import soot.dava.internal.AST.ASTSwitchNode;
import soot.dava.internal.AST.ASTSynchronizedBlockNode;
import soot.dava.internal.AST.ASTTryNode;
import soot.dava.internal.AST.ASTUnaryBinaryCondition;
import soot.dava.internal.AST.ASTUnconditionalLoopNode;
import soot.dava.internal.AST.ASTWhileNode;
import soot.dava.internal.SET.SETNodeLabel;
import soot.dava.internal.asg.AugmentedStmt;
import soot.dava.internal.javaRep.DAbruptStmt;
import soot.jimple.RetStmt;
import soot.jimple.ReturnStmt;
import soot.jimple.ReturnVoidStmt;
import soot.jimple.Stmt;

/*
 * This class is meant to be extended to write structred analyses.
 * The analysis is invoked by invoking the process method sending it
 * the body to be analyzed and the input flowset
 * Currently support is available only for a forward flow analysis.
 * This should soon be refactored to include backwards flow analysis
 * (Nomair 16th November 2005)
 */
public abstract class StructuredAnalysis {

  public static boolean DEBUG = false;
  public static boolean DEBUG_IF = false;
  public static boolean DEBUG_WHILE = false;
  public static boolean DEBUG_STATEMENTS = false;
  public static boolean DEBUG_TRY = false;
  /*
   * public static boolean DEBUG = true; public static boolean DEBUG_IF = true; public static boolean DEBUG_WHILE = true;
   * public static boolean DEBUG_STATEMENTS = true; public static boolean DEBUG_TRY = true; /* /** Whenever an abrupt edge is
   * encountered the flow set is added into a the break or continue list and a NOPATH object is returned
   */
  DavaFlowSet NOPATH = emptyFlowSet();
  public int MERGETYPE; // the confluence operator

  // the three types of operators
  final int UNDEFINED = 0;
  final int UNION = 1;
  final int INTERSECTION = 2;

  // storing before and after sets for each stmt or ASTNode
  HashMap> beforeSets, afterSets;

  public StructuredAnalysis() {
    beforeSets = new HashMap>();
    afterSets = new HashMap>();
    MERGETYPE = UNDEFINED;
    // invoke user defined function which makes sure that you have the merge
    // operator set
    setMergeType();
    // System.out.println("MergeType is"+MERGETYPE);
    if (MERGETYPE == UNDEFINED) {
      throw new RuntimeException("MERGETYPE UNDEFINED");
    }
  }

  /*
   * This method should be used to set the variable MERGETYPE use StructuredAnalysis.UNION for union use
   * StructuredAnalysis.INTERSECTION for intersection
   */
  public abstract void setMergeType();

  /*
   * Returns the flow object corresponding to the initial values for the catch statements
   */
  public abstract DavaFlowSet newInitialFlow();

  /*
   * Returns an empty flow set object Notice this has to be a DavaFlowSet or a set extending DavaFlowSet (hopefully
   * constantpropagationFlowSET??)
   */
  public abstract DavaFlowSet emptyFlowSet();

  /**
   * Make a clone of the flowset The implementor should know when they want a shallow or deep clone
   */
  public abstract DavaFlowSet cloneFlowSet(DavaFlowSet flowSet);

  /**
   * Specific stmts within AST Constructs are processed through this method. It will be invoked everytime a stmt is
   * encountered
   */
  public abstract DavaFlowSet processStatement(Stmt s, DavaFlowSet input);

  /**
   * To have maximum flexibility in analyzing conditions the analysis API breaks down the aggregated conditions to simple
   * unary or binary conditions user defined code can then deal with each condition separately. To be able to deal with
   * entire aggregated conditions the user should wite their own implementation of the method processCondition
   */
  public abstract DavaFlowSet processUnaryBinaryCondition(ASTUnaryBinaryCondition cond, DavaFlowSet input);

  /**
   * To deal with the local used for synch blocks
   */
  public abstract DavaFlowSet processSynchronizedLocal(Local local, DavaFlowSet input);

  /**
   * Deal with the key in the switch construct
   */
  public abstract DavaFlowSet processSwitchKey(Value key, DavaFlowSet input);

  public void print(Object toPrint) {
    System.out.println(toPrint.toString());
  }

  /**
   * This implementation breaks down the aggregated condition to the terminal conditions which all have type
   * ASTUnaryBinaryCondition. Once these are obtained the abstract method processUnaryBinaryCondition is invoked. For
   * aggregated conditions the merging is done in a depth first order of the condition tree.
   */
  public DavaFlowSet processCondition(ASTCondition cond, DavaFlowSet input) {
    if (cond instanceof ASTUnaryBinaryCondition) {
      return processUnaryBinaryCondition((ASTUnaryBinaryCondition) cond, input);
    } else if (cond instanceof ASTAggregatedCondition) {
      ASTCondition left = ((ASTAggregatedCondition) cond).getLeftOp();
      DavaFlowSet output1 = processCondition(left, input);

      ASTCondition right = ((ASTAggregatedCondition) cond).getRightOp();
      DavaFlowSet output2 = processCondition(right, output1);

      return merge(output1, output2);
    } else {
      throw new RuntimeException("Unknown ASTCondition found in structred flow analysis");
    }
  }

  /*
   * The parameter body contains the body to be analysed It can be an ASTNode, a Stmt, an augmentedStmt or a list of ASTNodes
   * The input is any data that is gathered plus any info needed for making decisions during the analysis
   */
  public DavaFlowSet process(Object body, DavaFlowSet input) {
    if (body instanceof ASTNode) {
      beforeSets.put(body, input);
      DavaFlowSet temp = processASTNode((ASTNode) body, input);
      afterSets.put(body, temp);
      return temp;
    } else if (body instanceof Stmt) {
      beforeSets.put(body, input);
      DavaFlowSet result = processAbruptStatements((Stmt) body, input);
      afterSets.put(body, result);
      return result;
    } else if (body instanceof AugmentedStmt) {
      AugmentedStmt as = (AugmentedStmt) body;
      Stmt s = as.get_Stmt();

      beforeSets.put(s, input);
      DavaFlowSet result = processAbruptStatements(s, input);
      afterSets.put(s, result);
      return result;

    } else if (body instanceof List) {
      // this should always be a list of ASTNodes
      Iterator it = ((List) body).iterator();
      DavaFlowSet result = input;
      while (it.hasNext()) {
        Object temp = it.next();
        if (!(temp instanceof ASTNode)) {
          throw new RuntimeException(
              "Body sent to be processed by " + "StructuredAnalysis contains a list which does not have ASTNodes");
        } else {
          /*
           * As we are simply going through a list of ASTNodes The output of the previous becomes the input of the next
           */
          beforeSets.put(temp, result);
          result = processASTNode((ASTNode) temp, result);
          afterSets.put(temp, result);
        }
      } // end of going through list

      // at this point the result var contains the result of processing
      // the List
      return result;
    } else {
      throw new RuntimeException("Body sent to be processed by " + "StructuredAnalysis is not a valid body");
    }
  }

  /*
   * This method internally invoked by the process method decides which ASTNode specialized method to call
   */
  public DavaFlowSet processASTNode(ASTNode node, DavaFlowSet input) {
    if (node instanceof ASTDoWhileNode) {
      return processASTDoWhileNode((ASTDoWhileNode) node, input);
    } else if (node instanceof ASTForLoopNode) {
      return processASTForLoopNode((ASTForLoopNode) node, input);
    } else if (node instanceof ASTIfElseNode) {
      return processASTIfElseNode((ASTIfElseNode) node, input);
    } else if (node instanceof ASTIfNode) {
      return processASTIfNode((ASTIfNode) node, input);
    } else if (node instanceof ASTLabeledBlockNode) {
      return processASTLabeledBlockNode((ASTLabeledBlockNode) node, input);
    } else if (node instanceof ASTMethodNode) {
      return processASTMethodNode((ASTMethodNode) node, input);
    } else if (node instanceof ASTStatementSequenceNode) {
      return processASTStatementSequenceNode((ASTStatementSequenceNode) node, input);
    } else if (node instanceof ASTSwitchNode) {
      return processASTSwitchNode((ASTSwitchNode) node, input);
    } else if (node instanceof ASTSynchronizedBlockNode) {
      return processASTSynchronizedBlockNode((ASTSynchronizedBlockNode) node, input);
    } else if (node instanceof ASTTryNode) {
      return processASTTryNode((ASTTryNode) node, input);
    } else if (node instanceof ASTWhileNode) {
      return processASTWhileNode((ASTWhileNode) node, input);
    } else if (node instanceof ASTUnconditionalLoopNode) {
      return processASTUnconditionalLoopNode((ASTUnconditionalLoopNode) node, input);
    } else {
      throw new RuntimeException("processASTNode called using unknown node type");
    }
  }

  /**
   * This method is called from the specialized ASTNodes. The purpose was to deal with different ASTNodes with similar
   * structure in one go. The method will deal with retrieve the body of the ASTNode which are known to have only one subBody
   */
  public final DavaFlowSet processSingleSubBodyNode(ASTNode node, DavaFlowSet input) {
    // get the subBodies
    List subBodies = node.get_SubBodies();
    if (subBodies.size() != 1) {
      throw new RuntimeException("processSingleSubBodyNode called with a node without one subBody");
    }
    // we know there is only one
    List subBody = (List) subBodies.get(0);
    return process(subBody, input);
  }

  /**
   * returns label on the ASTNode null if the ASTNode cannot hold a label or if the label is null
   */
  public String getLabel(ASTNode node) {
    if (node instanceof ASTLabeledNode) {
      Object temp = ((ASTLabeledNode) node).get_Label();
      if (temp != null) {
        return temp.toString();
      }
    }
    return null;
  }

  /**
   * Whenever a statement has to be processed the first step is to invoke this method. This is to remove the tedious work of
   * adding code to deal with abrupt control flow from the programmer of the analysis. The method invokes the
   * processStatement method for all other statements
   *
   * A programmer can decide to override this method if they want to do something specific
   */
  public DavaFlowSet processAbruptStatements(Stmt s, DavaFlowSet input) {
    if (s instanceof ReturnStmt || s instanceof RetStmt || s instanceof ReturnVoidStmt) {
      // dont need to remember this path
      return NOPATH;
    } else if (s instanceof DAbruptStmt) {
      DAbruptStmt abStmt = (DAbruptStmt) s;

      // see if its a break or continue
      if (!(abStmt.is_Continue() || abStmt.is_Break())) {
        // DAbruptStmt is of only two kinds
        throw new RuntimeException("Found a DAbruptStmt which is neither break nor continue!!");
      }

      DavaFlowSet temp = NOPATH;
      SETNodeLabel nodeLabel = abStmt.getLabel();
      // System.out.println("here");
      if (nodeLabel != null && nodeLabel.toString() != null) {
        // explicit abrupt stmt
        if (abStmt.is_Continue()) {
          temp.addToContinueList(nodeLabel.toString(), input);
        } else if (abStmt.is_Break()) {
          temp.addToBreakList(nodeLabel.toString(), input);
        } else {
          throw new RuntimeException("Found abruptstmt which is neither break nor continue");
        }
      } else {
        // found implicit break/continue
        if (abStmt.is_Continue()) {
          temp.addToImplicitContinues(abStmt, input);
        } else if (abStmt.is_Break()) {
          temp.addToImplicitBreaks(abStmt, input);
        } else {
          throw new RuntimeException("Found abruptstmt which is neither break nor continue");
        }
      }
      return temp;
    } else {
      /**************************************************************/
      /****** ALL OTHER STATEMENTS HANDLED BY PROGRAMMER **************/
      /**************************************************************/
      return processStatement(s, input);
    }
  }

  /*
   * Notice Right now the output of the processing of method bodies is returned as the output. This only works for INTRA
   * procedural Analysis. For accomodating INTER procedural analysis one needs to have a return list of all possible returns
   * (stored in the flowset) And merge Returns with the output of normal execution of the body
   */
  // reasoned about this....seems right!!
  public DavaFlowSet processASTMethodNode(ASTMethodNode node, DavaFlowSet input) {
    DavaFlowSet temp = processSingleSubBodyNode(node, input);
    return temp;
  }

  public DavaFlowSet processASTStatementSequenceNode(ASTStatementSequenceNode node, DavaFlowSet input) {
    DavaFlowSet output = cloneFlowSet(input);// needed if there are no stmts

    for (AugmentedStmt as : node.getStatements()) {
      Stmt s = as.get_Stmt();
      /*
       * Since we are processing a list of statements the output of previous is input of next
       */
      output = process(s, output);
      if (DEBUG_STATEMENTS) {
        System.out.println("After Processing statement " + s + output.toString());
        ;
      }
    }
    return output;
  }

  // reasoned about this....seems right!!
  public DavaFlowSet processASTLabeledBlockNode(ASTLabeledBlockNode node, DavaFlowSet input) {
    DavaFlowSet output1 = processSingleSubBodyNode(node, input);

    // handle break
    String label = getLabel(node);
    return handleBreak(label, output1, node);
  }

  public DavaFlowSet processASTSynchronizedBlockNode(ASTSynchronizedBlockNode node, DavaFlowSet input) {
    input = processSynchronizedLocal(node.getLocal(), input);

    DavaFlowSet output = processSingleSubBodyNode(node, input);
    String label = getLabel(node);
    return handleBreak(label, output, node);
  }

  // reasoned about this....seems right!!
  public DavaFlowSet processASTIfNode(ASTIfNode node, DavaFlowSet input) {
    input = processCondition(node.get_Condition(), input);

    DavaFlowSet output1 = processSingleSubBodyNode(node, input);

    // merge with input which tells if the cond did not evaluate to true
    DavaFlowSet output2 = merge(input, output1);

    // handle break
    String label = getLabel(node);

    DavaFlowSet temp = handleBreak(label, output2, node);

    if (DEBUG_IF) {
      System.out.println("Exiting if node" + temp.toString());
    }
    return temp;
  }

  public DavaFlowSet processASTIfElseNode(ASTIfElseNode node, DavaFlowSet input) {
    // get the subBodies
    List subBodies = node.get_SubBodies();
    if (subBodies.size() != 2) {
      throw new RuntimeException("processASTIfElseNode called with a node without two subBodies");
    }
    // we know there is only two subBodies
    List subBodyOne = (List) subBodies.get(0);
    List subBodyTwo = (List) subBodies.get(1);

    // process Condition
    input = processCondition(node.get_Condition(), input);
    // the current input flowset is sent to both branches
    DavaFlowSet clonedInput = cloneFlowSet(input);
    DavaFlowSet output1 = process(subBodyOne, clonedInput);

    clonedInput = cloneFlowSet(input);
    DavaFlowSet output2 = process(subBodyTwo, clonedInput);

    DavaFlowSet temp = merge(output1, output2);
    // notice we handle breaks only once since these are breaks to the same
    // label or same node
    String label = getLabel(node);
    output1 = handleBreak(label, temp, node);

    return output1;
  }

  public DavaFlowSet processASTWhileNode(ASTWhileNode node, DavaFlowSet input) {
    DavaFlowSet lastin = null;
    DavaFlowSet initialInput = cloneFlowSet(input);

    String label = getLabel(node);
    DavaFlowSet output = null;

    input = processCondition(node.get_Condition(), input);
    if (DEBUG_WHILE) {
      System.out.println("Going int while (condition processed): " + input.toString());
    }

    do {
      lastin = cloneFlowSet(input);
      output = processSingleSubBodyNode(node, input);

      // handle continue
      output = handleContinue(label, output, node);

      // merge with the initial input
      input = merge(initialInput, output);
      input = processCondition(node.get_Condition(), input);
    } while (isDifferent(lastin, input));

    // input contains the result of the fixed point
    DavaFlowSet temp = handleBreak(label, input, node);
    if (DEBUG_WHILE) {
      System.out.println("Going out of while: " + temp.toString());
    }
    return temp;
  }

  public DavaFlowSet processASTDoWhileNode(ASTDoWhileNode node, DavaFlowSet input) {
    DavaFlowSet lastin = null, output = null;
    DavaFlowSet initialInput = cloneFlowSet(input);
    String label = getLabel(node);
    if (DEBUG_WHILE) {
      System.out.println("Going into do-while: " + initialInput.toString());
    }

    do {
      lastin = cloneFlowSet(input);
      output = processSingleSubBodyNode(node, input);

      // handle continue
      output = handleContinue(label, output, node);

      output = processCondition(node.get_Condition(), output);

      // merge with the initial input
      input = merge(initialInput, output);
    } while (isDifferent(lastin, input));

    // output contains the result of the fixed point since do-while breaks
    // of at the processing of cond
    DavaFlowSet temp = handleBreak(label, output, node);
    if (DEBUG_WHILE) {
      System.out.println("Going out of do-while: " + temp.toString());
    }

    return temp;

  }

  public DavaFlowSet processASTUnconditionalLoopNode(ASTUnconditionalLoopNode node, DavaFlowSet input) {
    // an unconditional loop behaves almost like a conditional While loop
    DavaFlowSet initialInput = cloneFlowSet(input);
    DavaFlowSet lastin = null;
    if (DEBUG_WHILE) {
      System.out.println("Going into while(true): " + initialInput.toString());
    }

    String label = getLabel(node);
    DavaFlowSet output = null;
    do {
      lastin = cloneFlowSet(input);
      output = processSingleSubBodyNode(node, input);

      // handle continue
      output = handleContinue(label, output, node);

      // merge this with the initial input
      input = merge(initialInput, output);
    } while (isDifferent(lastin, input));

    // the output is not part of the set returned
    // it is just used to retrieve the set of breaklists stored for this
    // label
    DavaFlowSet temp = getMergedBreakList(label, output, node);
    if (DEBUG_WHILE) {
      System.out.println("Going out of while(true): " + temp.toString());
    }
    return temp;

  }

  public DavaFlowSet processASTForLoopNode(ASTForLoopNode node, DavaFlowSet input) {
    for (AugmentedStmt as : node.getInit()) {
      Stmt s = as.get_Stmt();
      input = process(s, input);
    }

    // finished processing the init part of the for loop
    DavaFlowSet initialInput = cloneFlowSet(input);

    input = processCondition(node.get_Condition(), input);
    DavaFlowSet lastin = null;
    String label = getLabel(node);
    DavaFlowSet output2 = null;
    do {
      lastin = cloneFlowSet(input);

      // process body
      DavaFlowSet output1 = processSingleSubBodyNode(node, input);

      // handle continues (Notice this is done before update!!!)
      output1 = handleContinue(label, output1, node);

      // notice that we dont merge with the initial output1 from
      // processing singleSubBody
      // the handlecontinue function takes care of it

      // handle update
      output2 = cloneFlowSet(output1);// if there is nothing in update

      for (AugmentedStmt as : node.getUpdate()) {
        Stmt s = as.get_Stmt();
        /*
         * Since we are just going over a list of statements the output of each statement is the input of the next
         */
        output2 = process(s, output2);
      }

      // output2 is the final result

      // merge this with the input
      input = merge(initialInput, output2);
      input = processCondition(node.get_Condition(), input);
    } while (isDifferent(lastin, input));

    // handle break
    return handleBreak(label, input, node);
  }

  /*
   * Notice ASTSwitch is horribly conservative....eg. if all cases break properly it will still merge with defaultOut which
   * will be a NOPATH and bound to have empty or full sets
   */
  public DavaFlowSet processASTSwitchNode(ASTSwitchNode node, DavaFlowSet input) {
    if (DEBUG) {
      System.out.println("Going into switch: " + input.toString());
    }

    List indexList = node.getIndexList();
    Map> index2BodyList = node.getIndex2BodyList();

    Iterator it = indexList.iterator();

    input = processSwitchKey(node.get_Key(), input);
    DavaFlowSet initialIn = cloneFlowSet(input);

    DavaFlowSet out = null;
    DavaFlowSet defaultOut = null;

    List> toMergeBreaks = new ArrayList>();

    while (it.hasNext()) {
      // going through all the cases of the switch
      // statement
      Object currentIndex = it.next();
      List body = index2BodyList.get(currentIndex);

      // BUG FIX if body is null (fall through we shouldnt invoke process
      // Reported by Steffen Pingel 14th Jan 2006 on the soot mailing list
      if (body != null) {
        out = process(body, input);

        // System.out.println("Breaklist for this out is"+out.getBreakList());
        toMergeBreaks.add(cloneFlowSet(out));

        if (currentIndex instanceof String) {
          // this is the default
          defaultOut = out;
        }

        // the input to the next can be a fall through or directly input
        input = merge(out, initialIn);
      } // body was non null
    }

    // have to handle the case when no case matches. The input is the output
    DavaFlowSet output = null;
    if (out != null) {
      // just to make sure that there were some cases
      // present

      /*
       * January 30th 2006, FOUND BUG The initialIn should only be merge with the out if there is no default in the list of
       * switch cases If there is a default then there is no way that the initialIn is the actual result. Then its either the
       * default or one of the outs!!!
       */
      if (defaultOut != null) {
        // there was a default
        // System.out.println("DEFAULTSET");
        // System.out.println("defaultOut is"+defaultOut);
        // System.out.println("out is"+out);

        output = merge(defaultOut, out);
      } else {
        // there was no default so no case might match
        output = merge(initialIn, out);
      }

    } else {
      output = initialIn;
    }

    // handle break
    String label = getLabel(node);

    // have to handleBreaks for all the different cases
    List> outList = new ArrayList>();

    // handling breakLists of each of the toMergeBreaks
    for (DavaFlowSet mset : toMergeBreaks) {
      outList.add(handleBreak(label, mset, node));
    }

    // merge all outList elements. since these are the outputs with breaks
    // handled
    DavaFlowSet finalOut = output;
    for (DavaFlowSet outIt : outList) {
      finalOut = merge(finalOut, outIt);
    }

    if (DEBUG) {
      System.out.println("Going out of switch: " + finalOut.toString());
    }

    return finalOut;
  }

  public DavaFlowSet processASTTryNode(ASTTryNode node, DavaFlowSet input) {
    if (DEBUG_TRY) {
      System.out.println("TRY START is:" + input);
    }

    // System.out.println("SET beginning of tryBody is:"+input);
    List tryBody = node.get_TryBody();
    DavaFlowSet tryBodyOutput = process(tryBody, input);
    // System.out.println("SET end of tryBody is:"+tryBodyOutput);

    /*
     * By default take either top or bottom as the input to the catch statements Which goes in depends on the type of
     * analysis.
     */
    DavaFlowSet inputCatch = newInitialFlow();
    if (DEBUG_TRY) {
      System.out.println("TRY initialFLOW is:" + inputCatch);
    }

    List catchList = node.get_CatchList();
    Iterator it = catchList.iterator();
    List> catchOutput = new ArrayList>();

    while (it.hasNext()) {
      ASTTryNode.container catchBody = (ASTTryNode.container) it.next();

      List body = (List) catchBody.o;
      // list of ASTNodes

      // result because of going through the catchBody
      DavaFlowSet tempResult = process(body, cloneFlowSet(inputCatch));
      // System.out.println("TempResult going through body"+tempResult);
      catchOutput.add(tempResult);
    }

    // handle breaks
    String label = getLabel(node);

    /*
     * 30th Jan 2005, Found bug in handling out breaks what was being done was that handleBreak was invoked using just
     * handleBreak(label,tryBodyoutput,node) Now what it does is that it looks for the breakList stored in the tryBodyOutput
     * node What might happen is that there might be some breaks in the catchOutput which would have gotten stored in the
     * breakList of the respective catchoutput
     *
     * The correct way to handle this is create a list of handledBreak objects (in the outList) And then to merge them
     */
    List> outList = new ArrayList>();

    // handle breaks out of tryBodyOutput
    outList.add(handleBreak(label, tryBodyOutput, node));
    // System.out.println("After handling break from tryBodyOutput"+outList.get(0));

    // handling breakLists of each of the catchOutputs
    for (DavaFlowSet co : catchOutput) {
      DavaFlowSet temp = handleBreak(label, co, node);

      if (DEBUG_TRY) {
        System.out.println("TRY handling breaks is:" + temp);
      }

      outList.add(temp);
    }

    // merge all outList elements. since these are the outputs with breaks
    // handled
    DavaFlowSet out = tryBodyOutput;
    for (DavaFlowSet co : outList) {
      out = merge(out, co);
    }

    if (DEBUG_TRY) {
      System.out.println("TRY after merge outList is:" + out);
    }

    // System.out.println("After handling break"+out);

    for (DavaFlowSet co : catchOutput) {
      out = merge(out, co);
    }

    if (DEBUG_TRY) {
      System.out.println("TRY END RESULT is:" + out);
    }

    return out;
  }

  /*
   * MERGETYPE var has to be set 0, means no type set 1, means union 2, means intersection
   */
  public DavaFlowSet merge(DavaFlowSet in1, DavaFlowSet in2) {
    if (MERGETYPE == 0) {
      throw new RuntimeException("Use the setMergeType method to set the type of merge used in the analysis");
    }

    DavaFlowSet out;
    if (in1 == NOPATH && in2 != NOPATH) {
      out = in2.clone();
      out.copyInternalDataFrom(in1);
      return out;
    } else if (in1 != NOPATH && in2 == NOPATH) {
      out = in1.clone();
      out.copyInternalDataFrom(in2);
      return out;
    } else if (in1 == NOPATH && in2 == NOPATH) {
      out = in1.clone();
      out.copyInternalDataFrom(in2);
      return out; // meaning return NOPATH
    } else {
      // both are not NOPATH
      out = emptyFlowSet();
      if (MERGETYPE == 1) {
        in1.union(in2, out);
      } else if (MERGETYPE == 2) {
        in1.intersection(in2, out);
      } else {
        throw new RuntimeException("Merge type value" + MERGETYPE + " not recognized");
      }
      out.copyInternalDataFrom(in1);
      out.copyInternalDataFrom(in2);
      return out;
    }
  }

  public DavaFlowSet mergeExplicitAndImplicit(String label, DavaFlowSet output, List> explicitSet,
      List> implicitSet) {
    DavaFlowSet toReturn = output.clone();

    if (label != null) {
      // use the explicit list
      /*
       * If there is no list associated with this label or the list is empty there no explicit merging to be done
       */
      if (explicitSet != null && explicitSet.size() != 0) {
        // explicitSet is a list of DavaFlowSets
        Iterator> it = explicitSet.iterator();

        // we know there is atleast one element
        toReturn = merge(output, it.next());

        while (it.hasNext()) {
          // merge this with toReturn
          toReturn = merge(toReturn, it.next());
        }
      } // a non empty explicitSet was found
    } // label not null could have explicit sets

    // toReturn contains result of dealing with explicit stmts

    // dealing with implicit set now
    if (implicitSet != null) {
      // implicitSet is a list of DavaFlowSets
      Iterator> it = implicitSet.iterator();
      while (it.hasNext()) {
        // merge this with toReturn
        toReturn = merge(toReturn, it.next());
      }
    }
    return toReturn;
  }

  /**
   * Need to handleBreak stmts There can be explicit breaks (in which case label is non null) There can always be implicit
   * breaks ASTNode is non null
   */
  public DavaFlowSet handleBreak(String label, DavaFlowSet out, ASTNode node) {
    // get the explicit list with this label from the breakList
    List> explicitSet = out.getBreakSet(label);
    // System.out.println("\n\nExplicit set is"+explicitSet);
    // getting the implicit list now
    if (node == null) {
      throw new RuntimeException("ASTNode sent to handleBreak was null");
    }

    List> implicitSet = out.getImplicitlyBrokenSets(node);
    // System.out.println("\n\nImplicit set is"+implicitSet);

    // invoke mergeExplicitAndImplicit
    return mergeExplicitAndImplicit(label, out, explicitSet, implicitSet);
  }

  /**
   * Need to handleContinue stmts There can be explicit continues (in which case label is non null) There can always be
   * implicit continues ASTNode is non null
   */
  public DavaFlowSet handleContinue(String label, DavaFlowSet out, ASTNode node) {
    // get the explicit list with this label from the continueList
    List> explicitSet = out.getContinueSet(label);

    // getting the implicit list now
    if (node == null) {
      throw new RuntimeException("ASTNode sent to handleContinue was null");
    }

    List> implicitSet = out.getImplicitlyContinuedSets(node);

    // invoke mergeExplicitAndImplicit
    return mergeExplicitAndImplicit(label, out, explicitSet, implicitSet);
  }

  /**
   * Invoked from within the UnconditionalWhile processing method Need to handle both explicit and implicit breaks
   */
  private DavaFlowSet getMergedBreakList(String label, DavaFlowSet output, ASTNode node) {
    List> breakSet = output.getBreakSet(label);
    DavaFlowSet toReturn = null;

    if (breakSet == null) {
      // there is no list associated with this label hence no merging to
      // be done
      // since this is a call from unconditional this means there should
      // have been an implicit break
      toReturn = NOPATH;
    } else if (breakSet.size() == 0) {
      // list is empty for this label hence no merging to be done
      // since this is a call from unconditional this means there should
      // have been an implicit break
      toReturn = NOPATH;
    } else {
      // breakSet is a list of DavaFlowSets
      Iterator> it = breakSet.iterator();

      // we know there is atleast one element
      // making sure we dont send NOPATH
      toReturn = it.next();

      while (it.hasNext()) {
        // merge this with toReturn
        toReturn = merge(toReturn, it.next());
      }
    } // a non empty breakSet was found

    // dealing with implicit set now

    List> implicitSet = output.getImplicitlyBrokenSets(node);
    if (implicitSet != null) {
      // implicitSet is a list of DavaFlowSets
      Iterator> it = implicitSet.iterator();

      // making sure that we dont send NOPATH
      if (implicitSet.size() > 0) {
        toReturn = it.next();
      }

      while (it.hasNext()) {
        // merge this with toReturn
        toReturn = merge(toReturn, it.next());
      }
    }
    return toReturn;
  }

  public boolean isDifferent(DavaFlowSet oldObj, DavaFlowSet newObj) {
    if (oldObj.equals(newObj) && oldObj.internalDataMatchesTo(newObj)) {
      // set matches and breaks and continues also match
      // System.out.println("NOT DIFFERENT!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!");
      return false;
    } else {
      // System.out.println(oldObj);
      // System.out.println(newObj);
      // System.out.println("DIFFERENT!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!");
      return true;
    }
  }

  /*
   * The before set contains the before set of an ASTNode , a Stmt , and AugmentedStmt, Notice for instance for a for loop we
   * will get a before set before the loop and an after set after the loop
   *
   * we dont have info about before set before executing a particular stmt that kind of info is available if you know which
   * stmt u want e.g. the update stmt
   */
  public DavaFlowSet getBeforeSet(Object beforeThis) {
    return beforeSets.get(beforeThis);
  }

  public DavaFlowSet getAfterSet(Object afterThis) {
    return afterSets.get(afterThis);
  }

  public void debug(String methodName, String debug) {
    if (DEBUG) {
      System.out.println("Class: StructuredAnalysis MethodName: " + methodName + "    DEBUG: " + debug);
    }
  }

  public void debug(String debug) {
    if (DEBUG) {
      System.out.println("Class: StructuredAnalysis DEBUG: " + debug);
    }
  }

}
 















© 2015 - 2024 Weber Informatics LLC | Privacy Policy