MP 3 Images and Lists

Direct links to MP 3.1 and MP 3.2

Goals

In this MP (machine problem) you will:

• learn to manipulate linked memory by writing functions to modify linked lists
• practice using templates
• get familiar with iterators

Checking Out the Code

From your CS 225 git directory, run the following on EWS:

git fetch release
git merge release/mp3 -m "Merging initial mp3 files"

If you’re on your own machine, you may need to run:

git fetch release
git merge --allow-unrelated-histories release/mp3 -m "Merging initial mp3 files"

Upon a successful merge, your MP2 files are now in your mp3 directory.

Background Information: Template Classes

Identical to what you saw in lecture, template classes provide the ability to create generic container classes. In this MP, you will be writing a List class.

template <typename T>
class List {
    // implementation
};

This simply says that our class List has a parametrized type that we will call T. Similarly, the constructor will look like this:

template <typename T>
List<T>::List() {
    // implementation
}

We need the template <typename T> or template <class T> above all of our functions—it becomes part of the function signature. The keywords class and typename can be interchanged.

Template classes need access to the implementation for compilation. Every time a different class is used as the template, the code must be compiled to support containing it. For example, if you want to make a List<int>, the compiler must take the generic List<T> implementation code and replace all the Ts with ints inside it, and compile the result (this process is called template instantiation). Our solution to this is to #include "list-inl.h" at the bottom of our list.h file. This ensures that whenever a client includes our header file, he/she also gets the implementation as well for compilation purposes (there are other solutions, but this is how we will solve it in this course).

Background Information: Linked Lists

The interface of this List class is slightly different from what you have seen in lecture. This List has no sentinel nodes; the first node’s prev pointer, and the last node’s next pointer, are both NULL. In lieu of these sentinels, we keep a pointer head to the first node, and a pointer tail to the last node in the List. (In an empty list, both head and tail are NULL.) The List class also has an integer member variable, length, which represents the number of nodes in the List; you will need to maintain this variable.

Background Information: Iterators

We use iterators to figure out where we currently are in the list, what is the next/previous node, and to access the data. Iterator class has one member variable, namely a pointer to the node in the list. Some of the core functionality includes moving the pointer, getting current location, and checking the location of the iterator.

MP 3.1: A Linked List Implementation

In your mp3 folder, you will find that the List class is split into four .h files:

• list.h
• list-inl.h
• list-iter.h
• list-given-inl.h

You should leave ‘list-given-inl.h’ as it is. We have provided you with a skeleton for the rest of the functions needed for this part of the MP, but you will need to write the implementations. They are designed to force you to write pointer manipulation code. You will write code for the functions, which are declared in list.h but not defined in list-inl.h. You must add your implementation to the file list-inl.h. Also, you will need to implement three functions in list-iter.h

See the Doxygen for MP 3 for details of the List class.

valgrind ./mp3
valgrind ./test
valgrind ./test "List::reverse"

MP 3.1.1: Iterator

In order to provide the client code with the ability to read the data from the list in a uniform way, we need to have an iterator. We have provided a list iterator class list-iter.h which has some functionality implemented. You will need to add the following implementations:

• ListIterator& operator++()
• ListIterator operator--(int)
• bool operator!=(const ListIterator& rhs)

MP 3.1.2: ~List() and clear()

Since the List class has dynamic memory associated with it, we need to define all of the Rule of Three. We have provided you with the Copy Constructor and overloaded operator=.

• You will need to implement the List destructor (~List()) and the clear() helper function called by operator= (the assignment operator)
• Both the List destructor and clear() function should free all memory allocated for ListNode objects.

MP 3.1.3: Insertion

MP 3.1.3.a: The insertFront Function

(See the Doxygen for insertFront.)

• This function takes a data element and prepends it to the beginning of the list.
• If the list is empty before insertFront is called, the list should have one element with the same value as the parameter.
• You may allocate new ListNodes.

< 5 4 7 >

< 6 5 4 7 >

MP 3.1.3.b: The insertBack Function

(See the Doxygen for insertBack.)

• This function takes a data element and appends it to the end of the list.
• If the list is empty before insertBack is called, the list should have one element with the same value as the parameter.
• You may allocate new ListNodes.

< 5 4 7 >

< 5 4 7 6 >

Testing Your insert Functions

Once you have completed insertFront and insertBack, you should compile and test them:

make test
./test "List::insertFront"
./test "List::insertBack"

MP 3.1.4: Pointer Manipulation

MP 3.1.4.a: The reverse Helper Function

(See the Doxygen for reverse.)

In list-inl.h you will see that a public reverse method is already defined and given to you. You are to write the helper function that the method calls.

• This function will reverse a chain of linked memory beginning at startPoint and ending at endPoint.
• The startPoint and endPoint pointers should point at the new start and end of the chain of linked memory.
• The next member of the ListNode before the sequence should point at the new start, and the prev member of the ListNode after the sequence should point to the new end.
• You may NOT allocate new ListNodes.

< 1 2 3 4 5 6 7 >

< 7 6 5 4 3 2 1 >

MP 3.1.4.b: The reverseNth Function

(See the Doxygen for reverseNth.)

• This function accepts as a parameter an integer, $$n$$, and reverses blocks of $$n$$ elements in the list.
• The order of the blocks should not be changed.
• If the final block (that is, the one containing the tail) is not long enough to have $$n$$ elements, then just reverse what remains in the list. In particular, if $$n$$ is larger than the length of the list, this will do the same thing as reverse.
• You may NOT allocate new ListNodes.

< 1 2 3 4 5 6 7 8 9 >

< 3 2 1 6 5 4 9 8 7 >

< 1 2 3 4 5 6 >

< 4 3 2 1 6 5 >

Testing Your reverse Functions

Once you have completed reverse and reverseNth, you should compile and test them.

make test
./test "List::reverse"
./test "List::reverseNth #1"
./test "List::reverseNth #2"

MP 3.1.4.c: The waterfall Function

(See the Doxygen for waterfall.)

• This function modifies the list in a cascading manner as follows.
• Every other node (starting from the second one) is removed from the list, but appended at the back, becoming the new tail.
• This continues until the next thing to be removed is either the tail (not necessarily the original tail!) or NULL.
• You may NOT allocate new ListNodes.
• Note that since the tail should be continuously updated, some nodes will be moved more than once.

< 1 2 3 4 5 6 7 8 >

< 1 3 5 7 2 6 4 8 >

< 1 2 3 4 5 6 7 8 >

< 1 2 3 4 5 6 7 8 > - Skip the 1
  ^             ^
curr           tail

< 1 3 4 5 6 7 8 2 > - Remove the 2 and move it at the end
    ^           ^
  curr         tail

< 1 3 5 6 7 8 2 4 > - Skip the 3, and move the 4 to the end
      ^         ^
    curr       tail

< 1 3 5 7 8 2 4 6 > - Skip the 5 and move the 6 to the end
        ^       ^
      curr     tail

< 1 3 5 7 2 4 6 8 > - Skip the 7 and move the 8 to the end
          ^     ^
        curr   tail

< 1 3 5 7 2 6 8 4 > - We have moved past the original tail of the list.
            ^   ^       This is okay! Skip the 2 and move the 4 to the end,
          curr tail     now for the second time!

< 1 3 5 7 2 6 4 8 > Skip the 6 and move the 8 to the end, now for the second time!
              ^ ^
           curr tail

Testing Your waterfall Function

Once you have completed waterfall, you should compile and test it.

make test
./test "List::waterfall"

MP 3.1.5: Testing

Compile your code using the following command:

make test

After compiling, you can run all of the MP 3.1 tests at once with the following command:

./test [part=1]

MP 3.1: Extra Credit Submission

For extra credit, you can submit the code you have implemented and tested for part one of MP 3. Follow the instructions in the MP 3 Submission section for handing in your code.

MP 3.2: Sorting

You will be implementing the helper functions for one more member function of the List template class: sort. This is designed to help you practice pointer manipulation and solve an interesting algorithm problem. In the process of solving this problem, you will implement several helper functions along the way—we have provided public interfaces for these helper functions to help you test your code.

MP 3.2.1: The split Helper Function

(See the Doxygen for split.)

• This function takes in a pointer start and an integer splitPoint and splits the chain of ListNodes into two completely distinct chains of ListNodes after splitPoint many nodes.
• The split happens after splitPoint number of nodes, making that the head of the new sublist, which should be returned. In effect, there will be splitPoint number of nodes remaining in the current list.
• You may NOT allocate new ListNodes

list1 = < 1 2 3 4 5 >

list1 == < 1 2 >
list2 == < 3 4 5 >

Testing Your split Function

Once you have completed split, you should compile and test it.

make test
./test "List::split"

You should see images actual-split_*.png created in the working directory (these are generated by repeatedly splitting split.png). Compare them against expected-split_*.png.

MP 3.2.2: The merge Helper Function

(See the Doxygen for merge.)

• This function takes in two pointers to heads of sublists and merges the two lists into one in sorted order (increasing).
• You can assume both lists are sorted, and the final list should remain sorted.
• You should use operator< on the data fields of ListNode objects. This allows you to perform the comparisons necessary for maintaining the sorted order.
• You may NOT allocate new ListNodes!

list1 = < 1 3 4 6 >
list2 = < 2 5 7 >

list1 == < 1 2 3 4 5 6 7 >
list2 == < >

Testing Your merge Function

Once you have completed merge, you should compile and test it.

make test
./test "List::merge"

You should see the image actual-merge.png created in the working directory if your program terminates properly. This is generated by merging the images tests/merge1.png and tests/merge2.png. Compare this against expected-merge.png.

MP 3.2: The mergesort Helper Function

(See the Doxygen for mergesort.)

• This function sorts the list using the merge sort algorithm, explained below.
• You should use operator< on the data fields of ListNode objects. This allows you to perform the comparisons necessary for sorting.
• You should use the private helper functions you wrote above to help you solve this problem.
• You may NOT allocate new ListNodes
• This function’s runtime will be graded for efficiency (correct Big-Oh runtime)

< 6 1 5 8 4 3 7 2 9 >

< 1 2 3 4 5 6 7 8 9 >

Merge Sort — Algorithm Details

Merge Sort is a recursive sorting algorithm that behaves as follows:

• Base Case: A list of size 1 is sorted. Return.
• Recursive Case:
  • Split the current list into two smaller, more manageable parts
  • Sort the two halves (this should be a recursive call)
  • Merge the two sorted halves back together into a single list

In other words, Merge Sort operates on the principle of breaking the problem into smaller and smaller pieces, and merging the sorted, smaller lists together to finally end up at a completely sorted list.

MP 3.2: Testing

Compile your code using the following command:

make test

After compiling, you can run the MP 3.2 tests at once with the following command:

./test [part=2]

compare out.png out_01.png out_difference.png

MP 3: Submission

Our grading system will checkout your most recent (pre-deadline) commit for grading. Therefore, to hand in your code, all you have to do is commit it to your Subversion repository.

Be sure your working directory is the mp3 folder that was created when you checked out the code. To hand in your code, you first need to add the new files you created to the working copy of your repository by typing:

To commit your changes to the repository type:

git add -u
git commit -m "<your message>"
git push origin master

Grading Information

The following files are used to grade MP 3:

• list.h
• list-inl.h
• list-iter.h

All other files including any testing files you have added will not be used for grading.

Good Luck!