Clojurians Log v2

It's easier to do this via timbre which is the default logger and plugged into clojure.tools.logging. Let me find some snippets for you

ah it's from a thread you asked before :) https://clojurians.slack.com/archives/CLX41ASCS/p1747678325976339?thread_ts=1747671299.406549&cid=CLX41ASCS

dang, I knew I was researching this before and lost track of it facepalm

thank you, as always

well, let's thank @highpressurecarsalesm :)

indeed :-)

Yes, something has changed, load-string now binds *data-readers* similar to Compiler/load but this might have introduced a bug. Can you verify if the behavior is different in Clojure and file an issue?

I'll take a look at this first thing in the morning

> And, later, in a namespace that sends requests: Is that in the same namespace or a different one?

Also, this is now failing in BB but not Clojure:

(->> "nucli/commands/managed_deps/pulls-response.edn"
                   io/resource
                   slurp
                   (edn/read-string {:readers readers/nu-readers}))

ERROR in nucli.commands.managed-deps.repo-test/repo-activity (parser.cljc:555)
Uncaught exception, not in assertion.
Exception: java.lang.Exception: No reader function for tag nu/instant
 at edamame.impl.parser$parse_sharp.invokeStatic (parser.cljc:555)
    edamame.impl.parser$dispatch.invokeStatic (parser.cljc:638)
    edamame.impl.parser$parse_next.invokeStatic (parser.cljc:769)
    edamame.impl.parser$parse_next.invoke (parser.cljc:757)
    edamame.impl.parser$parse_next.invokeStatic (parser.cljc:758)
    edamame.impl.parser$parse_next.invoke (parser.cljc:757)
    edamame.impl.parser$parse_to_delimiter.invokeStatic (parser.cljc:248)
    edamame.impl.parser$parse_to_delimiter.invoke (parser.cljc:235)
    edamame.impl.parser$parse_to_delimiter.invokeStatic (parser.cljc:237)
    edamame.impl.parser$parse_map.invokeStatic (parser.cljc:574)
    edamame.impl.parser$dispatch.invokeStatic (parser.cljc:672)
    edamame.impl.parser$parse_next.invokeStatic (parser.cljc:769)
    edamame.impl.parser$parse_next.invoke (parser.cljc:757)
    edamame.impl.parser$parse_next.invokeStatic (parser.cljc:758)
    edamame.impl.parser$parse_next.invoke (parser.cljc:757)
    edamame.impl.parser$parse_to_delimiter.invokeStatic (parser.cljc:248)
    edamame.impl.parser$parse_to_delimiter.invoke (parser.cljc:235)
    edamame.impl.parser$parse_to_delimiter.invokeStatic (parser.cljc:237)
    edamame.impl.parser$parse_map.invokeStatic (parser.cljc:574)
    edamame.impl.parser$dispatch.invokeStatic (parser.cljc:672)
    edamame.impl.parser$parse_next.invokeStatic (parser.cljc:769)
    edamame.impl.parser$parse_next.invoke (parser.cljc:757)
    edamame.impl.parser$parse_next.invokeStatic (parser.cljc:758)
    edamame.impl.parser$parse_next.invoke (parser.cljc:757)
    edamame.impl.parser$parse_to_delimiter.invokeStatic (parser.cljc:248)
    edamame.impl.parser$parse_to_delimiter.invoke (parser.cljc:235)
    edamame.impl.parser$parse_to_delimiter.invokeStatic (parser.cljc:237)
    edamame.impl.parser$parse_map.invokeStatic (parser.cljc:574)
    edamame.impl.parser$dispatch.invokeStatic (parser.cljc:672)
    edamame.impl.parser$parse_next.invokeStatic (parser.cljc:769)
    edamame.impl.parser$parse_next.invoke (parser.cljc:757)
    edamame.impl.parser$parse_next.invokeStatic (parser.cljc:758)
    edamame.core$parse_next.invokeStatic (core.cljc:126)
    sci.impl.parser$parse_next$fn__3133.invoke (parser.cljc:166)
    sci.impl.parser$parse_next.invokeStatic (parser.cljc:166)

So problem #1 is related to *data-readers* value reverting after the call to set!. Problem #2 seems to be clojure.edn/read-string ignoring the :readers option.

is that a new problem or was it already in bb?

In .207 all I needed was the set! of *data-readers* and that code was using clojure.core/read-string not clojure.edn/read-string.

can you make an isolated repro to avoid confusion?

Yes, let me put something together.

I'm not sure if the example earlier intended to be read-string or edn/read-string. edn/read-string should not print anything related to edamame

Here's a little test case:

borkdude@MBP25-2 ~/dev/scratch/data_reader $ cat src/foo.clj
(ns foo)

(set! *data-readers* (assoc *data-readers* 'foo/bar identity))

(require 'bar)

(defn exec [_]
  (prn bar/x))
borkdude@MBP25-2 ~/dev/scratch/data_reader $ cat src/bar.clj
(ns bar)

(def x #foo/bar [1 2 3])
borkdude@MBP25-2 ~/dev/scratch/data_reader $ clj -X foo/exec
[1 2 3]
borkdude@MBP25-2 ~/dev/scratch/data_reader $ bb -cp src -x foo/exec
[1 2 3]

and this fails in both clj + bb:

(ns foo
  (:require [clojure.edn :as edn]))

(set! *data-readers* (assoc *data-readers* 'foo/bar identity))

(require 'bar)

(defn exec [_]
  (prn bar/x)
  (prn *data-readers*)
  (prn (edn/read-string {:readers *data-readers*} "#foo/bar [1 2 3]")))

This does work in both clj + bb:

(prn (edn/read-string {:readers (assoc *data-readers* 'foo/bar identity)} "#foo/bar [1 2 3]"))

So, possibly my code was invalid but still worked in .207. Just about have a test case for you, but nothing new compared to your notes above.

Yes, this was "fixed" in bb to align more with JVM Clojure

Ok. I'm only about 15 years into Clojure, I'm sure I'll figure out how Vars get bound in a REPL soon. 😞

the issue that was previously there was that in threads, *data-readers* would not be bound correctly in a future:

(prn @(future (load-string "*data-readers*")))

If you still have something that isn't conform clojure with respect to Vars, happy to hear it.

Ok, I was totally panicing on the edn/read-string part. Turns out I had a second test using read-string that also needed to switch to edn/read-string.

Glad we sorted it out 😅

and sorry for the inconvenience

Nah, this was all me. Thanks for helping me sort it out.