1
0
mirror of https://github.com/janet-lang/janet synced 2024-12-23 15:00:27 +00:00
A dynamic language and bytecode VM.
Go to file
2018-01-20 17:21:59 -05:00
contrib/vim Work on interpreter. adding more opcodes and syscalls. 2017-11-24 23:17:04 -05:00
src Add bytecode verification 2018-01-20 17:21:59 -05:00
test Add bytecode verification 2018-01-20 17:21:59 -05:00
.gitignore Gitignore was ignore files. 2018-01-20 11:04:30 -05:00
.travis.yml Add OSX to travis. 2017-06-29 22:01:50 -04:00
CMakeLists.txt Add bytecode verification 2018-01-20 17:21:59 -05:00
LICENSE License 2017-04-17 22:36:44 -04:00
Makefile Oh travis 2018-01-20 10:57:01 -05:00
README.md Add self reference in closures without having to use vars. 2018-01-19 12:37:37 -05:00

dst

Build Status

dst is a functional programming language and vm. It is a variant of Lisp with several native useful datatypes. Some of the more interesting and useful features are first class functions and closures, immutable and mutable hashtables, arrays, and bytebuffers, macros (NYI), tail-call optimization, and continuations (coroutines, error handling). The runtime and compiler are written in C99, but should eventually be completely compatible with C89 compilers.

There is a repl for trying out the language, as well as the ability to run script files. This client program is separate from the core runtime, so dst could be embedded into other programs.

Features

First class closures Garbage collection lexical scoping First class green threads (continuations) Mutable and immutable arrays (array/tuple) Mutable and immutable hashtables (table/struct) Mutable and immutable strings (buffer/string) Byte code interpreter with an assembly interface Proper tail calls for functional code Direct interop with C

Compiling and Running

To build the runtime and run test, run

make test

A repl can also be run with

make repl

Todo

  • Documentation, with the caveat that things may change.
  • Use serialization to allow creation of bytecode files that can be loaded. This includes defining a file format for the bytecode files. This mostly done. The byte code serialization could also be useful for the module system for loading artifacts.
  • Pattern matching/regex library, as well as string formatting functions. This would also be useful for implementing better error messages.
  • Better error messages, expecially for compilation. This probably means string formating functions.
  • Macro/specials system that happens before compilation
  • Module system. Something similar to node's require.
  • Change name (dst is the name of many projects, including GNU Smalltalk). Maybe make logo :)?
  • Change C API to be stack based for fewer potential memory management problems. This could mean making current C API internal and use separate API externally.
  • Store source information in parallel data structure after parsing
  • Use source information during compilation
  • Use Lua style memory alocator backend C API (one single function for allocating/reallocating/freeing memory).
  • More builtin libraires.
  • Fuzzing