Skip to content
Oeiuwq Faith Blog OpenSource Porfolio

txus/lambra

Lambra is an experiment to implement a functional, distributed Lisp on the Rubinius VM.

txus/lambra.json
{
"createdAt": "2012-05-11T18:49:41Z",
"defaultBranch": "master",
"description": "Lambra is an experiment to implement a functional, distributed Lisp on the Rubinius VM.",
"fullName": "txus/lambra",
"homepage": "",
"language": "Ruby",
"name": "lambra",
"pushedAt": "2014-03-07T15:29:42Z",
"stargazersCount": 36,
"topics": [],
"updatedAt": "2023-10-17T06:36:33Z",
"url": "https://github.com/txus/lambra"
}

Lambra is an experiment to implement a functional, distributed Lisp on the Rubinius Virtual Machine, much à la Erlang.

Lambra processes are much like Erlang’s, but implemented as native OS threads. They share absolutely no memory, and behave like actors: they communicate via their mailboxes.

(defn echo []
(receive
[pid msg] (send pid self msg)))
(let [echo-pid (spawn echo)]
(send echo-pid self "Hello world!")
(receive
[pid msg] (print msg)))

The actor primitives in Lambra are receive and send.

receive blocks until a message in the mailbox can be processed, then pattern matches ([pid msg] is a catch-all clause) and executes the form that matched.

send sends a message to the mailbox of an actor. It just needs its pid, and the rest of the arguments will be sent as a message. It is a convention for the first argument of the message to be the sender’s pid, but it is in no way necessary.

self is the current process’ pid.

The structure for this first version is taken from Brian Shirai’s Poetics.

Joe Armstrong for his Programming Erlang book, and well, for Erlang.

This was made by Josep M. Bach (Txus) under the MIT license. I’m @txustice on twitter (where you should probably follow me!).