Code from March 6 CIS 211 lecture

As we all discovered last week, parsing of even the simplest grammars is pretty tedious and hard. For this reason, we rarely write parsers by hand. Instead we write "specifications" and use tools to automatically generate parsers. I have generated the parser for our machine's object files automatically. (I started this program early this morning. I omitted the check for line numbers to save time. Otherwise I believe it works fine.) Here is what I had to do:

Write a specification of the lexical structure of object files. The bottom of the file contains the meat. The top is mumbo-jumbo that you shouldn't care too much about at this point.
Using a tool called "JLex", I produced a Java program that performs the lexical analysis.
Next I wrote a specification of the grammatical structure of object files. Again the bottom of the files contains the meat and you can ignore the top.
Using a tool called "CUP", I produced two Java programs that define all the symbols used, and perform the actual parsing.
Finally to put it all together, I wrote three files:
- a file to handle errors during lexical analysis or parsing,
- a file that specifies the interface of the lexical analyzer,
- a file to test this whole thing in the context of the duckMachine.

Even without access to the JLex and CUP tools, you can copy all the .java files to your duckMachine/operatingSystem directory, compile everything as usual, and run the TestObjectFile class giving it an object file as argument. Enjoy!

Visited

times since December 15, 1997 (or the last crash).

sabry@cs.uoregon.edu