I’m building an editor integration to evaluate Elixir code in an IEx session. While Code.eval_string/3 allows tracking variable bindings, i think there’s currently no way to persist the environment (aliases, imports, etc.) between evaluations using this approach.
Use Case:
User is using an editor to send code to IEx (through a named pipe, e.g.), which is then read by a custom module defined in .iex.exs.
User defines alias MyApp.Module in one evaluation.
In subsequent evaluations, they expect Module.function() to work (as if it would in a regular IEx shell).
Proposal:
Expose a public API to evaluate code directly in the IEx instance, e.g.:
# it could be a helper function
result = IEx.Helpers.eval_string("IO.puts(123)")
# or maybe work by directly sending a message to IEx’s evaluator process (e.g., using eval_pid() if such a function were available)
send(eval_pid(), {:eval, "IO.puts(123)"})
I think this would make integrating IEx with external tools such as code editors a much easier task. Helix allows piping selected code directly to a Bash command, so i was using that to send code to a named pipe and read that in IEx. IEx is very useful, but i always experience some friction between it and my editor because i have to manually copy and paste code to it.
I found this post from a while ago that suggests something that i think would work for my case, but i tried it on Elixir 1.18.1 with OTP 26 and it didn’t work, maybe because it’s not a public API.
I’d love to hear if others have faced similar challenges or if there’s an existing approach i might have missed. I’d also be happy to contribute to this implementation if there’s interest, though I’m still learning Elixir and don’t have experience with Erlang.
