fix: remove the possible duplicate output introduced in a052dd8

[hugoduncan]

Output to the original output outside of the with-out-binding loop.

[vemv]

Is it possible to add test coverage that would have caught this bug?

[vemv]

For satisfying kondo, you can teach it about this macro similarly to what is done here:

https://github.com/clojure-emacs/refactor-nrepl/blob/da810357676803d43ab32680b4a244833868c792/.clj-kondo/config.edn#L41-L42

[vemv]

Btw, from what I've seen it's fine to use a branch within this repo, instead of a fork. Makes it easier to review locally

[vemv]

I verified locally that if reverting the fix, the included tests would fail with duplicates as intended:

FAIL in (forking-printer-test) (out_test.clj:68)
forking-printer prints to all message streams and original stream with String argument
expected: (= "Hello" (.toString out-writer))
  actual: (not (= "Hello" "HelloHello"))

lein test :only cider.nrepl.middleware.out-test/forking-printer-test

FAIL in (forking-printer-test) (out_test.clj:73)
forking-printer prints to all message streams and original stream with String argument
expected: (= "HelloHello" (.toString out-writer))
  actual: (not (= "HelloHello" "HelloHelloHelloHello"))

lein test :only cider.nrepl.middleware.out-test/forking-printer-test

FAIL in (forking-printer-test) (out_test.clj:84)
forking-printer prints to all message streams and original stream with int
expected: (= " " (.toString out-writer))
  actual: (not (= " " "  "))

lein test :only cider.nrepl.middleware.out-test/forking-printer-test

FAIL in (forking-printer-test) (out_test.clj:89)
forking-printer prints to all message streams and original stream with int
expected: (= "  " (.toString out-writer))
  actual: (not (= "  " "    "))

lein test :only cider.nrepl.middleware.out-test/forking-printer-test

FAIL in (forking-printer-test) (out_test.clj:99)
forking-printer prints to all message streams and original stream with char array
expected: (= "and" (.toString out-writer))
  actual: (not (= "and" "andand"))

lein test :only cider.nrepl.middleware.out-test/forking-printer-test

FAIL in (forking-printer-test) (out_test.clj:104)
forking-printer prints to all message streams and original stream with char array
expected: (= "andand" (.toString out-writer))
  actual: (not (= "andand" "andandandand"))

lein test :only cider.nrepl.middleware.out-test/forking-printer-test

FAIL in (forking-printer-test) (out_test.clj:114)
forking-printer prints to all message streams and original stream with String with offsets
expected: (= "good" (.toString out-writer))
  actual: (not (= "good" "goodgood"))

lein test :only cider.nrepl.middleware.out-test/forking-printer-test

FAIL in (forking-printer-test) (out_test.clj:119)
forking-printer prints to all message streams and original stream with String with offsets
expected: (= "goodgood" (.toString out-writer))
  actual: (not (= "goodgood" "goodgoodgoodgood"))

lein test :only cider.nrepl.middleware.out-test/forking-printer-test

FAIL in (forking-printer-test) (out_test.clj:129)
forking-printer prints to all message streams and original stream with char array with offsets
expected: (= "bye" (.toString out-writer))
  actual: (not (= "bye" "byebye"))

lein test :only cider.nrepl.middleware.out-test/forking-printer-test

FAIL in (forking-printer-test) (out_test.clj:134)
forking-printer prints to all message streams and original stream with char array with offsets
expected: (= "byebye" (.toString out-writer))
  actual: (not (= "byebye" "byebyebyebye"))

[bbatsov]

Thanks!

[vemv]

@bbatsov : the bug being fixed seemed pretty bad, so it might be a good idea to cut a bugfix release now?

[bbatsov]

It seems to me that only a tiny number of people have been affected by this, so it can wait until the scheduled release on Sunday. I haven't seen any reports of the problem at all - after all this only affects output that would normally end up in the server buffer, not regular output.