Hưng Đặng’s Reviews > Writing An Interpreter In Go > Status Update

Start to read "The lost chapter". It is about macro. Macro treats code as data, not text. When lexing, the input we receive is text but when parsing, we convert tokens to AST nodes. These AST nodes are different from text, are treated as data to be passed around so they are the macro system

This part is about Hash data structure and it is quite interesting. This is the first time I know that a hash key is converted into a number and its modulo will offset its position in an array. Under the hood, the hash is an array.

Adding builtin functions and some common data structure to the language. This is good to remind me how to work on things from the start to the end. Lexing/tokenizing -> Parsing to AST -> Turn things to object for evaluation

Environment is exciting! I have some guesses that variables in env will be saved in the memory and soon enough, Ball will have to talk about GC. Can't wait for the part about the GC

This is an artwork because Ball made me feel that things were so simple

Coming to conditionals. it seems like this one is harder than evaluating boolean or integer

Note: Ball has a great habit of writing tests before writing the features. I feel like we have even more test code than feature code

Each section was encapsulated well in their part of the codebase so that they don't affect one another. I feel it very easy to extend

In term of performance, executing machine code > executing bytecode > executing AST