Compiler Construction (D7011E) – Programming project

Step 2: Parsing MiniJava

Instructions

• Implement a parser for the MiniJava language according to the grammar in Appel, Appendix
  A.1, using the ANTLR tool .
  
  
• Apart from simply checking whether the input conforms to the MiniJava definition, the parser
  should also construct an abstract syntax representation of the parsed program.
  
  
• Input to the parser should be the token stream produced by the lexer implemented in step 1.
  
  
• Check that your parser accepts all files in testsuite (same files as in step 1). Also check that it
  is able to detect and report the errors embedded in the badsuite files.
  
  
• The main class of your program must be named Main.
  
  
• A short main method shall call the parser when the program is executed with the "-p" flag,
  
  
• To make it possible to observe the result of parsing, your program should print a textual
  representation of the abstract syntax tree once it is built. and then print ”Parsing succeeded/failed”
  depending on whether parsing errors were found in the input or not.
  
  That is, you should be able to run the parser with java Main -p BinarySearch.java to test the
  parser on BinarySearch.java.
  
  
• Note that you should retain the ability to run the lexer with -l, as in the previous lab!
  
  
• Note: the grammar specification in Appel, A.2 cannot be used as is in any of the parser
  implementations listed above. The challenge is thus to rewrite the grammar so that it suits the
  predictive parsing technique of ANTLR. This particularly involves sorting out how nested
  expressions like !a.b(c)+d[e]*f.length<g.h().i() should be read, and how to
  express that grammatically.


• Note: A good start is to respect the Java operator precedence as given in the JLS or for instance in the following quick reference.