Please follow the instructions given in the lecture to complete the assignment.

* Download/install Xtext, preferably using a complete Eclipse distribution from http://xtext.itemis.com/xtext/language=en/23947/downloads
* Create a new workspace and a new Xtext project
* Write a grammar corresponding to the simplified EngineeringC model shown in the lecture
* Run the generated workflow and test the parser with some EngineeringC diagrams

To hand in your homework, ZIP the Eclipse workspace which contains the Xtext project and upload the ZIP file to your CVS repository. Deadline is Friday, Feb. 5th.

This homework is available as PDF, so you can easily print it. It’s written in German, so if you have problems understanding it, please contact me for an English version.

Update: Updated PDF!

Implement code generation for functions and function calls. You should test and implement the following features:

• Generate a function from a funcDecl, using a separate InstructionList and adrTab.
• Generate function calls, pushing actual parameters onto the stack and invoke a previously generated function by name.

You do not need to implement parameter overloading or access to global variables from functions.

Deadline is before the next lecture, which will be held Jan. 22nd. Until then, happy holidays everyone!

Implement code generation for if, else and while statements. You should start by reenabling your interpreter unit tests and try to fix one by one by extending your code generator. The compiler should handle boolean operators <, >, =, <=, etc. The expression compiler should generate a branch instruction if it detects a conditional expression. Afterwards, the surrounding code handling the if statement is responsible for jump target fixup to implement branching logic. Please make sure to write tests for all if/else combinations with true and false conditions. while should be simpler, but make sure to test it, too.

We’re now ready to actually generate working code. Your task, should you accept it, is to change your interpreter to a compiler for arithmetic expressions. Generate code for the following “features”:

• Numbers (PUSH)
• Addition / subtraction (DADD, DSUB)
• Multiplication / division (DMUL, DDIV)
• Variable assignment (DSTORE)
• Variable access (DLOAD)
• Edit: variable assignments do NOT have to work like expressions, i.e. you don’t have to support “x = y = 5”!

A common problem with code generation is that you have to distinguish between compile time and run time. These are two separate phases, which is different from the way most interpreters work. Compile time is the first phase, in which you parse statements and generate code. You can use the JVM debugger to check this phase for errors, have a look at the symbol table and the results of the parser. When running the generated code, you cannot simply attach a debugger to it. This means that in the case of a wrong result, you have to manually check the bytecode to see what’s wrong. You do not see that stack, you have no variable names and there are no explicit control structures. You can use the Navigator view in Eclipse to open the .class files in your bin folder and see the bytecode. To make sure the bytecode is working, manually interpret it in your head. Keep track of the values on the stack by writing them on a sheet of paper. Check out the BCEL manual for help on the bytecode!

And if anybody happens to know a good bytecode debugger, please let everyone know on the mailing list…

We’ve made the jump from interpretation to actual compiling of expressions. To start your code-generating compiler, download the BCEL library from http://jakarta.apache.org/bcel/.

Add the library to your project build path and write an example Java program as shown in the lecture. Run the BCELifier tool that comes with the library and take a good look at the output. Refer to the BCEL manual to get a better understanding of the code. Now, modify your interpreter to initialize the bytecode generation facilities. Remove the call to your recursive interpreter. This will cause your tests to fail, so delete or ignore them for now. Write a simple test which evaluates a string which only contains a number. Include the code produced by BCELifier in your parser method. You can leave the hardcoded numerical value in the code. Check that a .class file is generated and load it using the Java ClassLoader, invoke the generated method using java.lang.reflect, and return the result produced by the generated code.

We will start to generate code from the input text in the next homework.

Homework #4 (due Nov. 27th, see last entry!) requires you to implement the following:

• extend the parser to handle if, else, and while statements. The new grammar rules are:

  ifStat ::= 'if' condExpr '{' block '}' ('else' '{' block '}')?
  whileStat ::= 'while' condExpr '{' block '}'
  condExpr ::= expr (bop expr)?
  bop ::= '<' | '>' | '==' | ...

  A conditional expression (condExpr) is true if its value is > 0, and false otherwise.

• parse function declarations:

  funcDecl ::= name '(' paramDeclList ')' '{' block '}'
  paramDeclList ::= name*

  Extend your symbol table to be able to store a function declaration.

• interpret function calls:

  factor ::= number | name | name '(' expr* ')'

  Make sure to handle parameter symbols correctly when calling functions. It should be possible to access globals from a function, but you may choose to make function execution free of side effects. Please make a comment or two about how your implementation works in the code!

Please remember to write simple test cases first (if without else, functions without parameters…). Develop iteratively, adding features as you go. Do not attempt to get everything right on first attempt…

Please implement the following extensions to your interpreter:

• identifiers longer than one character
• floating point numbers

To do this, you have to implement a simple scanner, as demonstrated in the lecture. Create a Token class which contains information about a word. Tokens are created by the scanner using a finite automaton and then interpreted by the parser. As usual, create tests first!

I was asked what the criteria for evaluation of the homeworks are. The tests are very important, as is a consistent and readable code style. If the functionality is complete and working, and the tests make sense and cover most of the implementation, you will get full score. Comments are welcome, but not required, and I prefer meaningful method and variable names. In extreme cases, I may deduct points for code I’m not able to comprehend, so keep it simple 🙂

Requirements for the second homework:

• parse variables
• interpret assignments and keep a symbol table
• treat assignments as expressions, i.e. handle statements such as “x = y = 5”

As usual, follow a test-driven programming style, and write unit test for each feature of your interpreter.

Hint: use the grammar given in the lecture to check your interpreter logic.

block ::= stat*
stat ::= expr | assign
assign ::= name '=' expr
expr ::= term ('+'|'-' term)*
term ::= factor ('*'|'/' factor)*
factor ::= number | '(' expr ')' | name
number ::= '0'...'9'
name ::= 'a'...'z' | 'A'...'Z'

When handling assignments like expressions, you have to extend this grammar in some way.

A student has asked me how repository folders should be named to avoid confusion. I suggest the following:

• Call the Eclipse project exactly as the CVS module name, e.g. “compilerbau” or something similarly short, no blanks please
• Create a normal Java project with a src and a bin folder!
• Inside the source folder, put your homework in a Java package, e.g. “de.uniks.compilerbau.homework01” for your first homework

Remember, this is only a suggestion, but you should really use Java packages. If you haven’t, you can use Eclipse’s refactoring tools to move your code into a package.