https://github.com/mushrom/nscheme

High-performance scheme interpreter
https://github.com/mushrom/nscheme

Last synced: about 2 months ago
JSON representation

High-performance scheme interpreter

• Host: GitHub
• URL: https://github.com/mushrom/nscheme
• Owner: mushrom
• Created: 2016-07-19T05:51:35.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2023-06-19T00:55:11.000Z (12 months ago)
• Last Synced: 2024-01-29T06:14:53.349Z (5 months ago)
• Language: C
• Size: 127 KB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Lists

• AwesomeInterpreter - nscheme

README

        
nscheme readme

==============

A fast, stack-based, JIT-compiling, concurrency-friendly scheme interpreter.

Currently a work in progress, it doesn't do much yet.

To build, type `make`, and that's it for now. A ./configure script and testing

framework will be coming soon.

## TODO

- [ ] parser

    - [x] basic lexer+parser

    - [x] global symbol table

        - [ ] add locking once multithreading is implemented

    - [x] handle basic types (ints, characters, null, ...)

    - [ ] handle unicode

        - [ ] generic utf* handling or just utf8?

    - [x] handle quoted forms

    - [x] handle dotted form pairs, (a . b)

    - [ ] handle vectors

    - [ ] handle strings

    - [ ] handle bytevectors

	- [ ] properly handle parser errors

- [ ] I/O

    - [/] `write` primitives

	    - [x] `display`

		- [ ] `write`

    - [x] `read` primitive 

    - [ ] implement port handles

        - [ ] implement a generic port structure to handle I/O for many different

              sources eg. sockets, files, memory buffers

        - [ ] handle port argument for I/O primitives display, write, read, ...

        - [ ] Implement open/close procedures for different sources

- [ ] virtual machine

    - [x] basic stack-based structure for vm

    - [ ] tree walker interpreter

        - [x] some basic forms (lambda, if, define, top-level begin)

        - [ ] expression-like begin

        - [ ] tail recursion (note: may not be needed if all lambdas are JIT'd on first call)

        - [ ] ! syntax expansion

    - [ ] JIT compiler

        - [x] Tail recursion

        - [ ] Scope analysis pass

            - [ ] handle local (define ...) forms

                - [x] single local variable definitions

                - [ ] function definition form

            - [ ] handle expression-like begin forms that generate new scope

        - [ ] ! handle nested lambdas

        - [ ] ! syntax expansion

        - [ ] optimizations

            - [ ] common subexpression elimination

            - [ ] dead code elimination

            - [ ] optimizing code which is guaranteed to have no side effects

            - [ ] inline small functions

                - will be especially useful for syntax extensions

                  which generate lambdas, eg. named lets)

    - [ ] garbage collection

        - [ ] ! implement basic pointer bumping allocation

        - [ ] ! implement mark-and-compact collector

        - [ ] figure out a way to make it work with multiple threads (arenas?)

    - [ ] exception handling

        - [ ] implement exception stack in vm struct

        - [ ] add `with-exception-handler` and `error` primitives

    - [ ] module system

        - [ ] figure out what to put here

        - [ ] implement r7rs libraries and associated procedures

    - [ ] FFI

        - [ ] maybe look at how python does it, or lua or chicken or something

- [ ] command line interface

    - [x] add help output

    - [x] add file argument

    - [ ] allow specifying options like environment, modules, output location as arguments

- [ ] testing

    - [ ] I/O tests

        - [ ] for basic std{in,out}

        - [ ] for network sockets

        - [ ] for memory buffers

    - [ ] stress testing for compiler

    - [ ] unit tests for scheme procedures