Requested features:

• Support dynamically setting the prompt
• Fully expose xrepl configuration to end users / developers

Tasks:

• TBD

Since almost (totally?) its very beginnings, LFE has made continuous references to HHGG, 42, etc. I think if would be fantastic if we did our own version of what Clojure does, but along our own HHGG theme ;-)

Clojure:

user=> (quit)
Bye for now!

xrepl:

lfe> (quit)
So long, and thanks for all the fish!

😆

This will depend upon #10 (so one can easily switch between REPL and adventure!)

Core capabilities:

• Manage configuration state
• Manage session $ENV

Might want to just start right off with ETS ...

Features:

• Handler support
• Middleware support

The task here is to review the current LFE shell code (in lfe_shell.erl) and identify all the functions that we would like to use when creating the xrepl. It looks like most of these are currently private, so we'll chat with Robert and open a PR to see if he's open to this, once we have the complete list together.

See https://vlaaad.github.io/reveal/

Subasks:

• Add README, complete with vision statement for lfe/xrepl
  • aims/goals
  • how it will take advantage of LFE's REPL/shell
  • how innovations in xrepl will be fed back into LFE proper
• Add OTP/gen_server skeleton project
• Add CI/CD pipeline

Features:

• Create abstraction for xrepl transport
• Support ETF communications between nodes (secure environment)
• Support SSH access
• Support SSH commands
• Support HTTPS access via REST API w/session support
• Support TTY access
• Support the nREPL protocol

Tasks:

• Review the old LFE ticket here: lfe/lfe#153
• Review the associated first draft here: oubiwann/lfe@2557cac
• Review the closed PR here: lfe/lfe#154
• Review related LFE tickets:
  • lfe/lfe#103
  • lfe/lfe#425
  • lfe/lfe#481
• Collect the latest requests from users/developers
• #6
• Use the above to update/add to the list of features below

Feature work:

• #8
• #9
• Manage multi-line support, complete with continuation lines
• #10

Use-case illuminating comments from John Goff on the LFE Discord:

Yeah, to be honest it feels limiting to be able to define functions in the repl but not overwrite functions defined in a file. The syntax of Elixir's defmodule makes this comparatively easy, since there's no "switching namespaces", you just type in a block and when that block's end is reached, we're back in the shell. But since LFE follows Erlang in that you're limited to one module per file, the syntax in LFE wasn't really made with this in mind.

The idea that I had was if this is possible, then you can connect your editor to your running system, and you can just send the entire buffer to the repl when you've made changes in the buffer that you want to appear in the live system.

I definitely think that LFE should provide some way of reloading a module via the repl, whether that's a part of xrepl or lfe itself I'm not sure.