Project 0: C Programming warm-up -------------------------------- INFORMATION ----------- Out: Monday March 26 Due: Wednesday April 4 electronically @ 11:59 p.m. Start early. BACKGROUND ---------- Despite incredible advances in programming languages over the last 30 years, most serious systems programming is still done in C. Why is this? Because C gives the programmer more control and power over the code's execution than do other, higher-level languages like Java or even C++. Also, C typically has less runtime overhead than higher-level languages, which can translate into increased performance. Suppose you have a function that takes an integer and returns a double. In a strongly typed language, all you can do with this function is call it while passing an integer and treat the result as a double. Of course, you can do this in C. But you can also call it with no parameters, call it with 5 parameters, take the result and store it in an integer. Even better, you could treat the function as an array and read each instruction as an integer if you like. Or, you could call not the first instruction in the function, but maybe the second, or the third, or ... there is a reason why C is sometimes referred to as a "high-level assembly language". What's bad about this freedom? Bugs. Forgot a parameter? Maybe you did it on purpose. Or maybe (and probably) not. In Java, the compiler shows you your mistake. In C, the compiler is very easy to please, but when you run the program, it fails, generally in a very cryptic way - "segmentation violation," and "core dumped" are the principle error messages. Both mean you made a mistake. All of this means that you need to program carefully and deliberately in C. If you do this, you can write programs that are as well structured and clear as you can in languages like Java. But, if you don't, you'll quickly have a big mess on your hands. Hard to debug. Hard to read. Hard to modify. There are lots of good references for programming in C. The primary one is "C Programming Language (2nd Edition)" by Kernighan and Ritchie. ASSIGNMENT ---------- You should do this assignment on a Linux machine, which will provide you not only with the compiler (cc, or gcc, depending on the installation) but also a debugger (gdb). You can use any editor you like, but we recommend that you check out emacs, which has support for C programming and debugging. We recommend the use of the CSE home Linux virtual machine (http://www.cs.washington.edu/lab/homeVMs/homeVMs.shtml) for this assignment. It's handy to set up, and comes pre-configured with everything you need. (It's sufficient for all projects in this course, except for Project 1, which is done on a CSE Linux box devoted to this course and configured in an odd way.) The one gotcha is that the VM runs a 64-bit guest operating system, which requires that the host OS (the one running directly on your hardware) be 64-bits as well. Note on 64-bits: The department has converted from 32-bit operating systems to 64-bit as the default. The most noticeable change you're likely to run into in C programming is that pointers are 8 bytes long, while ints are 4 bytes long. Casting between pointers and ints will result in a warning. If you're doing that sort of thing, you should consider using "long int" instead. Part 1. The Basic Queue ------------------------ The queue is one of the most important data structures you'll be dealing with in this class. Consequently, it's a good one to start working with early. For this part of the assignment, we will be providing you with a complete interface and mostly complete implementation for a queue. Starting with this, you will: 1. Find and fix two critical bugs in the implementation (queue.c). (We've put these bugs in after getting the code working). You will need to extend the test infrastructure (queuetest.c) significantly and ensure that it functions correctly. You should include comments that make it clear any problems you fixed (either in queue.c or in queuetest.c). 2. Implement the two declared, but not implemented, methods: queue_reverse() and queue_sort(). Both methods must work in-place: they can't create a new queue and move or copy elements from the original queue to build the result. The time efficiency of the sorting algorithm is not important, as long as it's something reasonable (not worse than O(n^2), not better than O(nlog(n)). queue_reverse() should execute in O(n) time. You should follow the coding style (indentation, naming, etc) that you find inside queue.c and queue.h. You should not change the interface that is already defined in queue.h. Part 2. The Hash Table ----------------------- Write the code in hash.c that implements the hash table interface defined in hash.h. The abstract hash table allows you to store and retrieve pointers to values of any kind based on keys of any type. You should not change the interface that is already defined in hash.h in the skeleton code. Internally, you're free to use any hash table implementation you like. You learned several variations, such as linear probing or separate chaining, in your data structures course. Testing ------- We have provided two small test files, queuetest.c and hashtest.c, that test the functionality of the queue and the hash table. These are very basic, limited test files that do not test the complete functionality of the data structures. You should extend these test files (or write your own) with additional test cases that test the complete data structure interfaces, as well as any corner cases you can think of. We will grade your data structures with test files that are much more comprehensive than the skeleton test files. A good practice is to write your test code BEFORE implementing your data structures, to make sure that you understand the expected behavior of the data structure's interface; you can then write the internal data structure code and see if your test cases pass or fail. Additionally, you should run the queuetest and hashtest programs under the valgrind tool to check for memory leaks and other memory bugs. More information about valgrind will be given in section or in a separate tutorial on the class website, but it is very easy to use; just run: valgrind ./queuetest valgrind ./hashtest 10 Submission ---------- The turnin instructions page describes the mechanics of the turnin tool: http://www.cs.washington.edu/education/courses/cse451/12sp/projects/turnin.txt In addition to following those instructions, you must ensure that: - Your Makefile is located in the top-level directory of your submission, and running the "make all" command produces the object files queue.o and hash.o in your top-level directory. We may deduct points for submissions that do not meet these criteria.