due: Monday, September 19
In this assignment you will implement a parser for the Xi programming language (this will include devising a grammar to describe the language's syntax). The end result will be a program that reads a Xi source file and produces a pretty-printed version of the AST representing the program.
Your parser must implement the interface
edu.cornell.cs.cs4120.xi.parser.Parser. We also ask that you
provide an implementation of the interface
edu.cornell.cs.cs4120.testing.ParserFactory, which constructs an
instance of your parser. Your factory class must be public and
have a no-argument constructor.
Your parser must be implemented using an LALR(1) parser generator, such as CUP for Java. If you are using some language other than Java, consult the course staff for the appropriate parser generator to use.
You must also supply a driver—a class with a
main method—that we can run. Your program should accept as
the sole argument the filename of a Xi source file. For
example:
$ java netid.Driver source.xiYou are free to add other options and arguments, but the command above must work.
Your program should behave as follows:
System.out)
an error message that includes the position of the error.As with PA1, we ask that you submit an overview document describing your
work. So we can use your code, include in the metadata section the
fully-qualified class names of your ParserFactory
implementation and your driver class.
The requirements for source control and package nomenclature are the same as in PA1. You are expected to use SVN or a similar source control system with the permission of course staff.
You will be using your lexer from PA1. Part of your task for this assignment is to fix any problems that you had in PA1.
If we discovered a problem with your lexer, for this assignment you must devise one or more test cases that clearly expose the bug. After you have done this (and confirmed that your PA1 implementation indeed fails these tests), fix the bug. Discuss these tests in your overview document, and explain briefly what the problem was.
The authors of JFlex have provided good support
for CUP compatible lexers. As such, if you used JFlex for PA1, we don't
require that your parser communicate with your JFlex-generated lexer through
the Lexer interface. Instead, you can modify your JFlex
specification to generate a lexer that your CUP-generated parser is able to
understand without an adapter. This likely requires some minor changes, but
the heart of your lexer will be the same. If you do exercise this option,
make sure to delete any extraneous source files.
Much like last time, all the code that you need for this assignment is
packaged in cs4120-pa2.jar, with source
in cs4120-pa2.zip. We are
including a few tools that you may find useful:
edu.cornell.cs.cs4120.xi.lexer.cup.LexerAdapter class is a
tool that allows Lexer implementations to interface with
CUP-generated parsers. If you properly implemented the Lexer
interface in PA1, you can create an instance of this adapter with
Lexer instance and your CUP-generated sym class.
You aren't required to use this if you modify your JFlex-generated lexer to
communicate directly with CUP.
In addition, we are providing you with a stub CUP specification from which to begin.
We have provided an interface AbstractSyntaxNode that your AST
classes must implement. It requires each node to provide a label for itself
and the references to its children. This interface gives us a way to inspect
the structure of your AST without constraining your creativity. You need to
develop your own accessor methods for individual AST node types.
Submit the following: