Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: THEend8_
Computer Science Department
CSC 413
Assignment 5 - Debugger
Due Date
Wednesday, May 18, BEFORE MIDNIGHT
No late submissions can be accepted for this assignment!
Note that the due date applies to the last commit timestamp into the main branch of your
repository.
Overview
The purpose of this assignment is to continue our work in the large compiler codebase, and implement
a Debugger.
You will be using your code for the Interpreter class, which must be copied into your github repository
when you begin the assignment via this assignment link.
Submission
Your assignment will be submitted using github. Only the “main” branch of your repository will be
graded. Late submission is determined by the last commit time on the “main” branch. You are required
to submit a documentation PDF named “documentation.pdf” in a “documentation” folder at the root of
your project.
Implementation Requirements
You may create additional classes as needed to implement the requirements defined here, but must
implement at least those classes listed here. If you decide to add additional classes, they must follow
object oriented design principles - proper encapsulation and data hiding, and implementing a single
responsibility.
Project Setup
1. You must be able to execute your program by typing:
javac interpreter/debugger/commands/*.java
javac interpreter/bytecode/*.java
javac interpreter/bytecode/debuggercodes/*.java
javac interpreter/Interpreter.java
java interpreter.Interpreter
Note that the Interpreter may now also be run in debugger mode by providing a switch (-d) and the
base file name (instead of the fully qualified bytecode file name):
java interpreter.Interpreter -d
You may assume that the bytecode programs that are used for testing are generated correctly, and
therefore should not contain any errors.
Note that the assignment 5 skeleton code provides you with an updated main method in the
Interpreter class to handle this switch!
2. Begin by replacing the placeholder classes in the project skeleton with your assignment 4
implementation. The following files should be replaced:
a. VirtualMachine.java
b. ByteCodeLoader.java
c. CodeTable.java
d. Program.java
e. Any other supporting files you created or that you need(i.e. RuntimeStack).
3. Copy your byte code classes from your assignment 4 implementation into the bytecode package.
New Implementation
1. You must add three new byte codes - LINE, FUNCTION, and FORMAL - to the set of byte codes we
are implementing. All existing byte codes should continue to function, and may need to add
behavior to support debugging.
a. The FUNCTION and FORMAL byte codes will be generated as a header for each function
declaration. The byte codes generated for each function will begin with:
LABEL name1 — branch label for function call
LINE n — start of function definition
FUNCTION name start end — name of function with source line
— number boundaries given by start
— and end
FORMAL f1 0 — f1 is first formal with offset 0
FORMAL f2 1 — f2 is second formal with offset 1
b. When debugging, we will not need to dump(), so no dump behavior is required for these byte
codes.
2. You must implement the FunctionEnvironmentRecord that will be used to track the current
function’s state in the debugger. The FunctionEnvironmentRecord consists of a symbol table,
a function name, the start and end lines of a function in the original source code file, and the current
line number. Note that the current line number must be reset when branching instructions are
processed.
The symbol table works a lot like the symbol table implementation we saw in constraining, though it
is slightly simpler - you may reuse the code from the Table class in the Constrainer, but you will
need to make some modifications. The symbol table is modified as follows when the given byte
codes are encountered:
FORMAL xyz n — enter(“xyz”, n)
LIT 0 i — enter(“i”, currentOffset)
POP n — delete the last n items entered into the symbol table
Note the that class provided in the assignment 5 skeleton code includes a main method to help test
your implementation. Compare the output from that main method to the correct output provided in
main’s function header.
A simulation of the environment stack changes during a debugger execution for the factorial
Bytecode Example Description
LINE LINE n LINE 5 n is the current source line number;
the generated byte codes for line n
will follow this code.
FUNCTION FUNCTION name start end FUNCTION g 1 20 name is the name of the function,
start is the source code line that
this function starts on, and end is
the source code line that this
function ends on.
FORMAL FORMAL name offset FORMAL f1 0 name is the name of the formal
parameter, offset is the stack offset
for the variable.
program is provided, with the environment stack shown, in Appendix A.
A simulation of symbol table changes during a debugger execution for a simple program is provided
in Appendix B.
3. You must provide debugger implementations for each of the byte codes discussed in class
(implementations from assignment 4), in addition to three new byte codes that will be introduced to
facilitate debugging. Note that byte codes for the debugger do everything they would do for the
interpreter, with the addition of some behavior for debugging.
a. Debugger byte codes must be placed in the sub package of interpreter.debugger created for this
purpose in your assignment 5 skeleton: interpreter.bytecode.debuggercodes.
4. To accommodate these new byte codes, we need a DebuggerCodeTable. The interface for the
DebuggerCodeTable is the same as the CodeTable, and only needs to include the new byte
codes that are needed by the debugger (i.e. any byte code that requires additional behavior for
debugging, and the three new byte codes described below). In the event that a byte code is not
found in the DebuggerCodeTable, you will return the result of getting that code from the
CodeTable. An implementation is provided for you, but you need to add your debugger byte codes
to its initialization method.
5. In the debugger package, you must update the implementation for the Debugger that extends the
Interpreter.
a. Just after the Interpreter (Debugger) starts, and before the VM executes the first
instruction, you should print the source program and then prompt the user for a command. This
should be handled by the DebuggerShell object, described below.
b. The Debugger will require some additional data structures in order to allow the user to debug a
program:
i. A Vector of entries where each Entry contains
int lineNumber: The line number for this Entry
String sourceLine: The source program line i for Vector slot i
Boolean isBreakpointLine: Is a breakpoint set for this line?
Accessors for these fields.
ii. A Stack of FunctionEnvironmentRecords. A new record will be pushed when a
function is entered, and popped when returning from a function.
c. The following commands must be implemented by the debugger:
i. ?
Display available commands. Do not display the command list at any other time!
Type ? for help
> ?
set
list
locals
source
step
continue
exit
Type ? for help
>
ii. set (set breakpoint)
This command records that a breakpoint has been set for a specific line
Type ? for help
> set
Enter line number:
> 2
Type ? for help
>
iii. list (list breakpoints)
This command lists all breakpoints currently set with the debugger
Type ? for help
> list
* 2: int factorial(int n) {
Type ? for help
>
iv. source (display source)
This command displays the source code of the current function, with an indication of where
the execution has stopped (if executing), and any breakpoints that have been set:
Type ? for help
> source
1: program {boolean j int i
-> * 2: int factorial(int n) {
3: if (n < 2) then
4: { return 1 }
5: else
6: {return n*factorial(n-1) }
7: }
8: while (1==1) {
9: i = write(factorial(read()))
10: }
11: }
Type ? for help