- Davy Chuon
- Varun Peesapati
The main objective of this project is to enable translation of E, an Object Oriented Language (OOPS) that is similar to C++ into code that can be compiled by the Java Virtual Machine (JVM). In other words, this project translates code written in E to code written in Java based on the grammar specified by the BNF included below.
Note: The BNF above only partially describes the E language with the non-terminals in lowercase and terminals enclosed in single quotes (' ').
In this part of the project, we read every character from our input (which is the code written in E) and classify them into different tokens. In the long run,this classification makes it easier to parse this input in parts 2 and 3. The code for this is present in Scan.java, Token.java, and TK.java.
This part of the project was written in Parser.java, where after tokenizing all the characters read from the input, we needed to parse it to check if the E program is syntactically correct. We managed to accomplish this by implementing a parse tree for all the different possibilities of code that could be present.
This part of the project made us keep track of all the variables that were declared in the E program with the help of a Symbol Table that was implemented by using an ArrayList of SymbolTableItem objects defined in SymbolTableItem.java. Since Java didn't support redeclaration of variables in lower scopes, we had to use the Symbol Table to ensure that the variables were declared in the right scope. The Symbol Table also kept track of the scope of each variable as well as the current scope of the program.
This part of the project involved us to generate code in Java by simply keeping print statements in several parts of the parse tree and is implemented in Parser.java.
$ make
$ java e2jOptional: Can also redirect input from an E file as follows:
$ make
$ java e2j < test.e$ tree .
.
├── e2j.java
├── EBNF.png
├── Makefile
├── Parser.java
├── README.md
├── Scan.java
├── SymbolTableItem.java
├── TK.java
└── Token.java