dnsmkl/dlex

Learning (lexers/regexps) project

DLEX - personal attempt on creating a simple lexer (or lexer generator) in dlang

Why?

Personal challenge

What?

Lexer (or lexer generator). Configurable during runtime. (so userdefined lexer could be usable without recompilation)

Usage will be something along the lines:

    auto lexer = makeLexer();
    lexer.add(regexstr1, name1, action1);
    lexer.add(regexstr2, name2, action2);
    lexer.setInput(input);

    foreach(auto tokenObject; lexer)
    {
        writeln(tokekenObject.name);
        writeln(tokekenObject.regexText);
        writeln(tokekenObject.matchedText);
        tokekenObject.associatedAction.execute();
    }

How?

To implement lexer, regular expression matcher will be needed.

Plan:

• implement regular expression
• make adjustments to regular expressions till it becomes lexer (many regular expressions will have to be melted into same NFA/DFA. Similar, but not exactly as normal union)

Examples of internal representations (just a reminder)

For regular expression dataflow will look like this: RegexString -> RegexAST -> NFA -> DFA -> match (instead of arrows there will be parsers/converters)

Example of RegexString "ab*b"

Example of RegexAST Sequence(Letter("a"), Repeat(Letter("b")), Letter("b"))

Example of NFA start:state1 transitions: state1 | 'a' -> state2 state2 | 'b' -> state2 state2 | 'b' -> state3 final:state3

Example of DFA start:state1 transitions: state1 | 'a' -> state2 state2 | 'b' -> state23 state23 | 'b' -> state23 final:state23