https://github.com/Zelex/jo_clojure

Fast Embeddable Clojure in C/C++
https://github.com/Zelex/jo_clojure

c clojure cpp language

Last synced: 2 months ago
JSON representation

Fast Embeddable Clojure in C/C++

• Host: GitHub
• URL: https://github.com/Zelex/jo_clojure
• Owner: Zelex
• License: mit
• Created: 2022-03-03T19:59:46.000Z (almost 3 years ago)
• Default Branch: main
• Last Pushed: 2024-05-20T21:47:24.000Z (8 months ago)
• Last Synced: 2024-08-03T21:02:50.022Z (5 months ago)
• Topics: c, clojure, cpp, language
• Language: C++
• Homepage:
• Size: 3.32 MB
• Stars: 144
• Watchers: 7
• Forks: 7
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • Funding: .github/FUNDING.yml
  • License: LICENSE.md
  • Code of conduct: CODE_OF_CONDUCT.md

Awesome Lists containing this project

• awesome-lisp-languages - jo_clojure - C] Fast Embeddable Clojure in C/C++, including persistent datastructures and STM (Languages / C/C++)

README

        
# JO Clojure

Author Jon Olick. @JonOlick on twitter. 

A Fast, (soon to be) Embeddable Clojure-like language implementation in C/C++. 

The near-term goal is to re-make clojure in an embedable form. 

The long-term goal is to add types/functions for machine learning, tensors, etc... 

# Usage:

* For the REPL: jclj

* For running a script: jclj file.clj

# Currently:

* Native implementation of almost entire core lib. See TODO.md

* Parses code into native structures (AST), then executes. Essentially interpreted - though does some clever things here and there. 

* Lazy sequences

* Startup time is ridiculously fast by comparison to clojure

* Implementations of persistent lists, vectors, hash-map, hash-set, matrix

* Software Transactional Memory (STM)

* Atoms

* IO

# Differences:

* IO is very different than clojure's IO cause JVM Clojure's IO functions are practically non-existant. 

* Software Transactional Memory (STM) now works seamlessly with Atoms - thus refs are only useful if you need the history feature?

* Not entirely complete implementation (yet). There is a LOT to do.

* See TODO.md for information on what is left to do.

# Future:

* Compile to C/C++ code so you can compile a native executable. 

* Save/Restore program state (continue later, saved in binary format). 

* Native types and operations for Matrices and Tensors. 

* Automatic Parallelization.

* sorted-set

* sorted-map

* refs

* agents

* other etc....

# Compile/install:

Unix:

Run `make` to build and `make install` to install to `/usr/local/bin` (might require `sudo`)

Windows:

Run `c.bat` for Microsoft C/C++ compiler. Run `make` for mingw gcc.