If you do :out :inherit , then the output is sent to the OS stdout

so you have to omit that, if you want to read the output yourself

I've tried that already, ommiting :out in a process map, changes nothing, it acts the same. Output to console, line is nil

with :out :inherit it wil surely not work

Okay, I've removed it, but any idea why it still is not working?

Try this:

(require '[babashka.process :as p] '[ :as io])

(let [proc (p/process "bb -e '(loop [] (println \"Hello\") (Thread/sleep 3000) (recur))'"
                      {:shutdown p/destroy-tree
                       :err :inheri})]
  (with-open [rdr (io/reader (:out proc))]
    (binding [*in* rdr]
      (loop []
        (when-let [line (read-line)]
          (prn :line line))
        (when (or (.ready rdr) (.isAlive (:proc proc)))
          (recur))))))

This example works for me

Yes, this example also works for me

But when it's replaced with docker command it's not working

The logical explanation could be that docker is writing to stderr and not stdout?

no, that's not it, I'm testing locally now

When you do:

docker build . > /tmp/log.txt

so maybe first investigate that

This does work:

docker build . 2> /tmp/log.txt

Ah got it:

(require '[babashka.process :as p] '[ :as io])

(let [proc (p/process "docker build .'"
                      {:shutdown p/destroy-tree})]
  (future (with-open [rdr (io/reader (:out proc))]
            (binding [*in* rdr]
              (loop []
                (when-let [line (read-line)]
                  (prn :stdout line))
                (when (or (.ready rdr) (.isAlive (:proc proc)))
                  (recur))))))
  (with-open [rdr (io/reader (:err proc))]
    (binding [*in* rdr]
      (loop []
        (when-let [line (read-line)]
          (prn :stderr line))
        (when (or (.ready rdr) (.isAlive (:proc proc)))
          (recur))))))

When you read from stdout it will block, but when you also want to read from stderr, you'll have to do the first in a future or so, to not block that

So tl;dr, docker logs to stderr

Thanks borkdude! You are a hero!

anyway I'll start playing with this stuff tomorrow, but probably the discussion has at least some value

It does

Afk now brb

$ bb -cp 'foo' -e '(System/getProperty "java.class.path")'
"foo"

Thanks! Will investigate tomorrow why multiple things in the cider <-> bb-nrepl integration aren't doing the right thing

this requires a relatively new version of bb, the sync between System/getProperty of the classpath is a recent thing I believe

other way of getting the bb classpath: (babashka.classpath/get-classpath) , this has worked for years

@U45T93RA6 I made this commit on a flight to strangeloop yesterday: https://github.com/babashka/babashka.nrepl/tree/classpath-op does that look alright? perhaps you have a suggestion on how to test this?

(except for the failing test, this is just a minor problem ;))

If you can include a unit test, I'll be able to confirm that its 'schema' matches cider-nrepl's

basically it would look like {"classpath" ["src" "test" "/Users/borkdude/.m2/foobar/foobar.jar"]}

Thanks, I've set a reminder for tomorrow

I'm using cljfmt from clojure-lsp whenever I hit a keystroke in emacs that does formatting. No complaints here.

wrong 🧵

ooh sorry

jetlag ;)

(Have a great one!)

thank you :)

something like this: https://github.com/babashka/babashka/commit/0cb0ccd9d448fa764a56849775c31a23423d1b32 we'll talk tomorrow or later.

(I had to move this to babashka proper since babashka.nrepl can be re-used for other non-bb projects)

You can exercise nrepl's classpath op with:

(cider-sync-request:classpath)

(<--
  id         "7"
  session    "5c300c9d-19d2-43d4-aa68-fc5db6a61328"
  time-stamp "2023-09-23 15:00:28.293704000"
  classpath  ("/Users/vemv/orchard/test-cljs" "/Users/vemv/orchard/test" "/Users/vemv/orchard/test-newer-jdks" "/Users/vemv/.lein/scripts" "/Users/vemv/formatting-stack/src" "/Users/vemv/formatting-stack/worker" "/Users/vemv/formatting-stack.alias-rewriter/src" "/Users/vemv/parallel-reload/src" "/Users/vemv/eastwood/src" "/Users/vemv/eastwood/copied-deps" "/Users/vemv/tufte.auto/src" "/Users/vemv/orchard/src-spec-alpha-2/src/main/clojure" "/Users/vemv/orchard/dev" "/Users/vemv/orchard/src" "/Users/vemv/orchard/src-jdk8" "/Users/vemv/orchard/src-newer-jdks" "/Users/vemv/formatting-stack/resources" "/Users/vemv/eastwood/resource" "/Users/vemv/eastwood/resources" "/Users/vemv/.lein/resources/rebl.jar" "/Users/vemv/orchard/test-resources" "/Users/vemv/orchard/not-a.jar" "/Users/vemv/orchard/does-not-exist.jar" "/Users/vemv/orchard/dev-resources" "/Users/vemv/orchard/resources" "/Users/vemv/orchard/target/classes" "/Users/vemv/.m2/repository/com/cognitect/aws/api/0.8.612/api-0.8.612.jar" "/Users/vemv/.m2/repository/com/stuartsierra/component.repl/0.2.0/component.repl-0.2.0.jar" "/Users/vemv/.m2/repository/joda-time/joda-time/2.10/joda-time-2.10.jar" "/Users/vemv/.m2/repository/org/apache/maven/shared/maven-shared-utils/3.3.4/maven-shared-utils-3.3.4.jar" "/Users/vemv/.m2/repository/com/gfredericks/test.chuck/0.2.13/test.chuck-0.2.13.jar" "/Users/vemv/.m2/repository/org/apache/maven/resolver/maven-resolver-util/1.8.2/maven-resolver-util-1.8.2.jar" "/Users/vemv/.cache/mx.cider/enrich-classpath/1601/2147402929/1797031119.jar" "/Users/vemv/orchard/java" "/Users/vemv/orchard/test-java" "/Users/vemv/.cache/mx.cider/unzipped-jdk-sources/1601" "/Library/Java/JavaVirtualMachines/adoptopenjdk-16.jdk/Contents/Home/lib/src.zip")
  status     ("done")
)

So, yes, the format has little room for error, it's a vector of strings You may want to verify it matches a few nuances here https://github.com/clojure-emacs/cider-nrepl/blob/master/src/cider/nrepl/middleware/classpath.clj and that it passes https://github.com/clojure-emacs/cider-nrepl/blob/master/test/clj/cider/nrepl/middleware/classpath_test.clj

Hey Michiel 👋 just catching up Is there an issue I can subscribe for this one? I'll want to try out this in CIDER again once it's fixed

Hi. Just got back from Strange Loop and test positive for COVID so taking it easy for a bit ... Just created the issue, also have a branch, no PR yet: https://github.com/babashka/babashka.nrepl/issues/66

Will get back to this soon

Thanks, subscribed! Please take it indeed