2022-07-21 sci | Clojure Slack Archive

sci

zane 2022-07-21T22:39:17.943249Z

This isn’t a bug report (I can’t reproduce with the latest version of sci) but I’m dying to know why sci.core/eval-string hangs on trivial inputs (e.g. "0") on sci 0.2.7 if reitit is also included as a dependency. 😆 🤔

zane 2022-07-22T15:32:33.084619Z

Ah!

zane 2022-07-21T22:41:46.041839Z

Weirdly, I don’t even need to require reitit!

zane 2022-07-21T22:42:24.198589Z

❯ nix-shell --pure  -p clojure

[nix-shell:~]$ timeout -v 60 clojure -Srepro -Sdeps '{:deps {metosin/reitit {:mvn/version "0.5.18"} org.babashka/sci {:mvn/version "0.2.7"}}}' -M -e "(require '[sci.core :as sci]) (sci/eval-string \"0\")"

timeout: sending signal TERM to command 'clojure'

zane 2022-07-21T22:44:02.321559Z

Works fine if reitit is excluded:

[nix-shell:~]$ time clojure -Srepro -Sdeps '{:deps {org.babashka/sci {:mvn/version "0.2.7"}}}' -M -e "(require '[sci.core :as sci]) (sci/eval-string \"0\")"
0

real	0m4.011s
user	0m14.914s
sys	0m0.576s

zane 2022-07-21T22:52:36.993179Z

Also works with sci 0.2.8:

[nix-shell:~]$ time clojure -Srepro -Sdeps '{:deps {metosin/reitit {:mvn/version "0.5.18"} org.babashka/sci {:mvn/version "0.2.8"}}}' -M -e "(require '[sci.core :as sci]) (sci/eval-string \"0\")"
0

real	0m4.263s
user	0m15.740s
sys	0m0.657s

zane 2022-07-21T23:01:14.268989Z

Simply requiring sci is fine, even with reitit:

[nix-shell:~]$ time clojure -Srepro -Sdeps '{:deps {metosin/reitit {:mvn/version "0.5.18"} org.babashka/sci {:mvn/version "0.2.7"}}}' -M -e "(require '[sci.core :as sci])"

real	0m4.255s
user	0m20.633s
sys	0m0.658s

zane 2022-07-21T23:08:54.480549Z

I’ve been staring at the 0.2.8 changelog and nothing stands out. 😆 https://github.com/babashka/sci/blob/master/CHANGELOG.md#v028

borkdude 2022-07-22T05:38:59.193029Z

Probably because of Malli which included another version of sci under my personal GitHub account. Inspect the classpath, there's likely two versions of SCI on there

Clojurians Log v2

sci