ICS 142: Compilers and Interpreters

Summer Session II, August 6 – September 12, 2001

Homework Assignments

As part of your homework assignments, you will construct an interpreter and compiler for Simple, the simple imperative language example. While the assignments are available above, we will provide some notes and hints on the project parts here.

General Instructions

Short Version: We need all your sources, excluding standard libraries. We need a file called "sc" that accepts one command line parameter, the file to be processed. We need either a Makefile or a script "mksc" that performs the compilation from your sources to an executable program we can run using "sc". We need all this printed, as well as sent to the email address listed on the handout in the form of a tar archive compressed with gzip; the archive should be named "sc-studentid.tar.gz" and create a directory "sc-studentid" with everything in it (studentid should be your eight digit UCI student ID number).

Long Version:

Your programs have to run on the ICS department's SunOS 5 workstations. To make your life easy, you should develop on these workstations as well. If you don't know how to work on a Unix system, you should start learning about it right away; many of the following terms refer specifically to Unix. You can work on these machines remotely using telnet or ssh.
If you can not work on these machines for some reason, your best bet is to use Java on your personal computer. Avoid libraries (and language extensions) not available in the standard JDK (i.e. don't use Apple, Microsoft, or other vendor-specific libraries or extensions in your code).
You may use Java, C, or C++ to implement your programming assignments, nothing else. The only exception is Python, and assignments using that language will be graded by the instructor, not the assistants.
Your program should always be named "sc" so that we can run it within our automated grading system. If your program needs a special environment, please write a shell script "sc" that performs the required setup and then runs the "real" program. A simple script for Java is included as part of the solutions for assignment 1, available on the Downloads page.
You have to include a Makefile (if you know how to write one) or a script that executes the compilation steps you did to get from your sources to your executable. Writing such a script is easy, since you just put the same commands you use all the time during development (e.g. javac, cc, ...) in one file. If you include a script and not a Makefile, call that script "mksc".
Your program should accept one command line parameter which is the name of the file to be processed; the required output (described on the particular assignment) should be made on the standard output. The exception is the final assignment in which error messages go to the standard output but you create an object file "sc.out" in the current directory.
Starting with assignment 2, you have to hand in a printed copy of your source code as well. Please make sure that you include all relevant information (name, student-id, who you collaborated with, etc.) in the header of each source file.
Send the source code for your programs to the email address indicated on the particular assignment in the form of a tar archive compressed with gzip. Make sure that extracting this archive creates a directory "sc-studentid" that contains everything needed to run your program, especially the "sc" file referred to above and a Makefile or "mksc" file for building your program. Of course, studentid should be your eight-digit student id number! Use the subject "ICS 142: Assignment x" where x is the number of the assignment.
Note that we will take off points for everything we have to do in addition to extracting, running, and comparing the results produced by your program.

Assignment 1: A Scanner for Simple

Peter: The "@" sign in the example output just indicates that the following is the position of the token. Similarly, the brackets "<" and ">" indicate that everything between them is the value associated with a token of that kind. This is really just a formatting issue.
Peter: To find out the keywords you have to recognize, look at the grammar. If it's in there, and if it's in quotation marks, then it probably is a keyword (or a symbol like ":=" for that matter). If you can't find something in the grammar, it's not a keyword (or symbol)!
Peter: The position should be the position from the beginning of the file (or string), not from the beginning of the line. Remember that we ignore linefeeds as whitespace, just like blanks or tabs. BTW, ignoring one TAB character just advances the position by one!
Peter: Test your scanner with the example given on the assignment. Once that works, test it with this file which should produce this output. Use "Save as..." to get these files, otherwise something might get lost!
Peter: For exercise 1, when I say "no abbreviations" I mean that you can not use "|" for alternatives, "[]" for optional, etc. If you believe the solution requires recursion, you are on the right track.
Peter: We will use the command sc our-test-file to run your programs. If you hand in a C/C++ solution, just call your program "sc" and you are fine. If you hand in a Java program, just include a short script like this with suitable modifications for your main class.
Peter: If you generate tokens individually in your scanner, you don't have to assemble them into a list in the driver program. Just print them! Try to come close to the output used on the (updated) homework assignment and in the test file from above. You don't have to match it 100% though.
Peter: You can download solutions for homework 1 either in Java or in Python. These are provided with no guarantee for correctness, so make sure they work before you use them! If you find a bug, please contact the author listed in the sources. Thanks!

Assignment 2: A Parser for Simple

Peter: Here are a couple of test cases for the parser. As before, please get these with "Save as.." just to be sure you have the same files: Program 1 Output 1 Program 2 Output 2 Program 3 Output 3 Please note that I did not prove my parser correct, so there might be a bug in one of these...

Assignment 3: The Simple Intermediate Representation

Peter: Note that you are not required to handle semantic errors "elegantly" in any way, you can still just abort with an error message when you find one.
Peter: Here is a first bunch of test cases, mostly for the symbol table but also for constant expressions. As before, please get these with "Save as.." just to be sure you have the same files: Program 1 Output 1 Program 2 Output 2 Program 3 Output 3 Program 4 Output 4 Again, please note that I did not prove my program correct, so there might be a bug in one of these. Also, please don't try to match my output perfectly since it depends on my particular data structures. As long as you show "essentially" the same thing, details really don't matter. Only if we can't tell at all that you are doing it right, that might be troublesome...
Peter: More test cases, also for instructions and all that. You know the drill, get these with "Save as.." just to be sure you have the same files: Program 5 Output 5 Program 6 Output 6 Program 7 Output 7 Program 8 Output 8 All notes from a above and all disclaimers apply, as usual. Enjoy!

Assignment 4: An Interpreter for Simple

No notes available, check the mailing list archive.

Assignment 5: A Code Generator for Simple

Here is a bunch of example programs written in Simple, together with the object files produced by Joe's compiler. If there is time I will also post the output by my compiler, but let's see. The code you generate does not have to look exactly like this, but it has to do the same things. As always, there is no guarantee for correctness. (Note that the file extension of these is ".sel" instead of ".sim" for historical reasons...)