Quark Method Reference

quark.core.quark.py

find_previous_method

The algorithm of find_previous_method

The find_previous_method method uses a DFS algorithm to collect all MethodObjects called by the parent_method and add them to the specified wrapper. The search starts from the base_method and goes on recursively until there are no more levels or all candidates have been processed.

1. Initialize an empty set "visited_methods" if it is not provided.
2. Get a set "method_set" using "self.apkinfo.upperfunc(base_method)".
3. Add "base_method" to the "visited_methods" set.
4. If "method_set" is not None then check if "parent_function" is in "method_set".
   - If yes, append "base_method" to "wrapper".
   - If no, then iterate through each item in "method_set".
        - If the item is in "visited_methods", skip it and continue to the next item.
        - If not, call "find_previous_method" again with the current item, "parent_function", "wrapper", and "visited_methods".

The code of find_previous_method

def find_previous_method(
    self, base_method, parent_function, wrapper, visited_methods=None
):
    """
    Find the method under the parent function, based on base_method before to parent_function.
    This will append the method into wrapper.

    :param base_method: the base function which needs to be searched.
    :param parent_function: the top-level function which calls the basic function.
    :param wrapper: list is used to track each function.
    :param visited_methods: set with tested method.
    :return: None
    """
    if visited_methods is None:
        visited_methods = set()

    method_set = self.apkinfo.upperfunc(base_method)
    visited_methods.add(base_method)

    if method_set is not None:

        if parent_function in method_set:
            wrapper.append(base_method)
        else:
            for item in method_set:
                # prevent to test the tested methods.
                if item in visited_methods:
                    continue
                self.find_previous_method(
                    item, parent_function, wrapper, visited_methods
                )

find_intersection

The algorithm of find_intersection

The find_intersection method takes in two sets, first_method_set and second_method_set, and finds their intersection using a recursive search algorithm.

Here is the process of find_intersection。

1. Check that the input sets are not empty.
    If one of the sets is empty, raise a ValueError.

2. Use the & operator to find the intersection of the two sets.
    If the intersection is not empty, return the resulting set.

3. If the intersection is empty, call the method_recursive_search
    function with the input sets and a specified maximum depth.

4. The method_recursive_search function recursively searches for
    the intersection of the two input sets up to the specified depth
    by splitting the sets into subsets and comparing each subset's elements.
      - If the intersection is found, return the resulting set.
      - Otherwise, return None.

The code of find_intersection

def find_intersection(self, first_method_set, second_method_set, depth=1):
    """
    Find the first_method_list ∩ second_method_list.
    [MethodAnalysis, MethodAnalysis,...]
    :param first_method_set: first list that contains each MethodAnalysis.
    :param second_method_set: second list that contains each MethodAnalysis.
    :param depth: maximum number of recursive search functions.
    :return: a set of first_method_list ∩ second_method_list or None.
    """
    # Check both lists are not null
    if not first_method_set or not second_method_set:
        raise ValueError("Set is Null")
    # find ∩
    result = first_method_set & second_method_set
    if result:
        return result
    else:
        return self.method_recursive_search(
            depth, first_method_set, second_method_set
        )