Clojurians Log v2
Clojure programming
Channels
Apps
nrepl 2026-11-02
oyakushev 2026-11-02T09:04:45.694099Z
When multiple forms are sent to nREPL in a single message, and an earlier form throws an exception, the next ones continue being evaluated. Example:
$ clojure -Sdeps "{:deps {nrepl/nrepl {:mvn/version \"1.8.0\"}}}" -m nrepl.cmdline -i
user=> (/ 1 0) (println "Hello!")
Execution error (ArithmeticException) at user/eval2499 (REPL:1).
Divide by zero
Hello!load-file op (which powers e.g. C-c C-k) disables this behavior.
Does anybody rely on this? Do you ever happen to evaluate multiple forms in a single message and want all of them to be executed regardless of errors?
Harold 2026-11-02T13:28:17.668969Z
I have noticed this behavior, here's a related example:
user> (+ 1 2))
3Syntax error reading source at (REPL:4:15).
Unmatched delimiter: )3 in there, somewhere. 🙂
I think I understand this behavior, and it makes sense to me, are you considering changing to a different behavior? What would it be, and why would it be better?oyakushev 2026-11-02T13:54:16.348859Z
Your example is about a different thing. Forms are read from the message and evaluated one by one. What I consider changing is when exception is thrown in one form, the rest of the input is discarded without being read and evaluated.
;; now
user=> (/ 1 0))
Execution error (ArithmeticException) at user/eval9623 (REPL:1).
Divide by zero
Syntax error reading source at (REPL:1:9).
Unmatched delimiter: )
;; after change
user=> (/ 1 0))
Execution error (ArithmeticException) at user/eval9623 (REPL:1).
Divide by zeroload-file anyway
seancorfield 2026-11-02T14:00:20.672919Z
I've run across that behavior occasionally and always found it annoying, so I'd be happy to see it "fixed".
❤️ 1
bozhidar 2026-11-02T15:45:21.125739Z
I also think we should probably fix it, and I'm reasonably sure this wouldn't have any negative impact.
Harold 2026-11-03T13:19:28.758699Z
I see, thanks for the explanation. I think it would be fine either way. 👍
❤️ 1