Clojurians Log v2

CIDER is not doing anything wrong here. You're converting the map using seq. CIDER prints exactly the data structure you asked for 😅

Nope, expected output is ([:a 1] [:b 2])

that's not what seq on a map does

Ah

Sorry 😅

weird

the inspector does this too - try cider-inspect on

(list (vec {:a 1 :b 2}))

this smells like it could be a MapEntry thing

Yeah definitely, tracking it down now

This must be related to the recent changes in internal CIDER printing logic. I will take a look

Yep, the orchard.print/print multifn on :map-entry

https://github.com/clojure-emacs/orchard/blob/ed119f0075eacb951012763fccd8afa1a9fb1975/src/orchard/print.clj#L95

@alexyakushev I see you were the author of those changes, alright to leave the fixes to you without opening an issue?

I believe inspector was always like that. Not sure why.

An issue would still be appreciated for posterity. I'll get to the fix

that's strange, surely someone else would've noticed it before

The inspector rendering code was not used in the debugger before. And in the inspector, I guess you don't often inspect sequences of mapentries.

I see, this behavior was used to print a map like a sequence of mapentries so that it looks like a regular printed map.

Here's my workaround for now, until I regain my trust in the debugger actually helping me to debug things :P

(require 'orchard.print)
(remove-all-methods orchard.print/print)
(defmethod orchard.print/print :default
  [x w] (print-method x w))

This will change the look of the inspector too, but sure, whatever works

Now I'm getting all sorts of TruncatingStringWriter$TotalLimitExceeded errors trying to debug a large nested value :/ I think these hardcoded values in cider-nrepl are the cause: https://github.com/clojure-emacs/cider-nrepl/blob/af0d1e0e2351293019fc95ec8b9c8822c51ac6c9/src/cider/nrepl/middleware/debug.clj#L185 I'm not so familiar with Java writers, but it appears to be throwing to signal a non-local exit instead of an unexpected error? Seems like there's a missing try-catch wrapper somewhere

(binding [orchard.print/*max-total-length* 10]
  (orchard.print/print-str :longer-than-ten-chars))
;; => Execution error (TruncatingStringWriter$TotalLimitExceeded) at mx.cider.orchard.TruncatingStringWriter/write (TruncatingStringWriter.java:87).
;;    null

Ah, my bad - I'm on an older version of cider-nrepl / orchard, I see it's been fixed on the current release

Fix is live in cider-nrepl 0.50.1

https://github.com/clojure-emacs/cider-nrepl/blob/master/src/cider/nrepl/middleware/out.clj performs bindings and alter-var-roots. Say that the 'old' code was running

(while true (println 1))

Probably reusing 'dead repls' (there's a defcustom for that) would result in old output and new output gathered under the same Emacs buffer, but it doesn't fix the behavior I described. I'd look instead at using file-backed logging, for example.

We also have cider-log-mode which certainly survives reconections :)

It's worth checking if this is related to the recent nrepl changes, just in case. Can you achieve a small reproducible that demonstrates this behavior?