LinddM / Cursed-compiler

Compiler for decaf

C# compiler for decaf

Run the program

dotnet run

Flags and params

-o

-target
    scan
    parse
    semantic
    irt
    codegen

-opt
    constant
    algebraic

-debug
    scan
    parse
    semantic
    irt
    codegen

Design

Files and directories

────class
│   ├───ast
│   │       Ast.cs
│   │
│   ├───codegen
│   │       Codegen.cs
│   │
│   ├───compiler
│   │       Compiler.cs
│   │
│   ├───irt
│   │       Irt.cs
│   │
│   ├───opt
│   │       Algebraic.cs
│   │       ConstantF.cs
│   │
│   ├───parser
│   │       Parser.cs
│   │
│   ├───scanner
│   │       Scanner.cs
│   │
│   └───semantic
│           Semantic.cs

Scanner

Tokenizes and classifies

Steps

1. Split incoming file by lines and separate by spaces

List<List<String>> myCleanTokens = cleanTokens(text);

2. Separate by operators

Dictionary<string, List<string>> tokensAndTypes = new Dictionary<string, List<string>>(classify(myCleanTokens));

3. Classify reserved words, exluding variables, objects and numbers

if(classifyTypes.ContainsKey(tokens[i][j])){
    // lineType positions: line, type, value
    lineType = new List<string>(){(i+1).ToString(), classifyTypes[tokens[i][j]].ToString(), tokens[i][j]};
                        
}

4. Define states for variables, objects and numbers. Evaluate

if(isVariable(tokens[i][j])){
    hasType=true;
    lineType = new List<string>(){(i+1).ToString(),"<variable>", tokens[i][j]};
}
if(isObject(tokens[i][j])){
    hasType=true;
    lineType = new List<string>(){(i+1).ToString(),"<object>", tokens[i][j]};
}
if(isNumber(tokens[i][j])){
    hasType=true;
    lineType = new List<string>(){(i+1).ToString(),"<number>", tokens[i][j]};
}

5. Throw errors in console if tokens doesn't belong to any type

if(entry[1]=="error"){
    Console.WriteLine("Error en la linea "+entry[0]+": "+entry[2]);
}