conjure 2026-02-08

Hi @olical - does conjure have versions of nrepl and cider/cider-nrepl that it supports ? I ask because when I use nrepl >= 1.5 the 1st evaluation of a ns is not recognised. Cheers

✅ 1

Yes, I get the same error. The first command to evaluate (either an ns form or the whole buffer) fails, unable to find the namespace. A second evaluation command in the same repl session works successfully. I started a new repl session a few times to see if it was a specific eval command (not the case it seems). Each session started a minimal repl with nrepl 1.5.2 and cider-nrepl 0.58.0, using Clojure 0.12.0

hmmm i vaguely remember running into this

If we can improve the backwards compatibility for this we totally should. The best way to solve this is use conjure#debug set to true, capture the log output and the results / error and get that in an issue. Happy to take a look.

I'm not aware of any prescribed minimum version, but there should be or we should fix backwards compatibility.

Sorry for the error!

Very simple example.

{:paths ["src"]
 :deps {org.clojure/clojure {:mvn/version "1.12.3"}}}
Evaluate this buffer:
(ns example.main)

(defn -main []
  (println "Hello, world"))
jack-in with: cider/cider-nrepl 0.58.0 With nrepl 1.5.0 ; Namespace not found: example.main With nrepl 1.4.0 nil #'example.main/-main Running it directly in the nrepl 1.5.0 works ok. Can anyone else reproduce this?

Its a weird one and only noticed it because of this post 🤔 I should have some time today to experiment. I'll share what I find (as a github issue). This doesn't detract from how awesome conjure is gratitude

I've created https://github.com/Olical/conjure/issues/759 to share what I did to replicate this issue. I tested with nrepl 1.4 (works correctly) and nrepl 1.5 & 1.5.2 (both have the issue). I notice that nrepl 1.5 release has a breaking change, which may be related (as the issue seems to be with 1.5.x versions of nrepl) As I havent debugged conjure before I https://practical.li/journal/debug-conjure/#conjure. I am doing some diffs of the output to see what has changed (all debug output has been committed to https://github.com/practicalli/clojure-through-code where I was testing the issue. Let me know if there is anything I missed. Thanks.

@jr0cket great investigation - thank you.

1