Cider debugger hangs on infinite lazy sequences

[yuhan0]

Expected behavior

Using the debugger to step through intermediate forms which produce infinite lazy sequences should not cause Cider to hang indefinitely.

Actual behavior

The debugger attempts to realize the entire lazy sequence when displaying it, the user cannot continue past that point and has to interrupt the evaluation with C-c C-b.

Steps to reproduce the problem

• Evaluate the form

(let [n 5]
  #dbg
  (take (inc n) (repeat n)))

• Press n a couple of times when the debugger is activated to step past the intermediate forms
• Upon reaching the form (repeat n) which produces an infinite lazy sequence, it hangs indefinitely with no way to step past it to the finite result (via take)

Environment & Version information

CIDER version information

;; CIDER 0.22.0snapshot (package: 20190331.2014), nREPL 0.6.0
;; Clojure 1.10.0, Java 12

Lein/Boot version

-NA-

Emacs version

26.1

Operating system

OS X 10.14.4 "Mojave"

[Malabarba]

Odd. Cider generally avoids printing sequences indefinitely. It could be something in your config, or it could be some pretty-printing change on our side.

What do you see printed if you evaluate (repeat 1) with C-x C-e?

[yuhan0]

It displays a finite sequence of (1 1 1 1 ...) truncated at *print-length*.
All the cider-eval and cider-inspect commands work fine with infinite seqs, it's only the debugger which behaves this way.

[bbatsov]

Likely we forgot to pass some print params to the eval commands used by the debugger. This should be a relatively simple fix once someone gets to investigate the problem.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contribution and understanding!

[stale]

This issues been automatically closed due to lack of activity. Feel free to re-open it if you ever come back to it.

[yuhan0]

Can this issue be reopened? It's still an existing pain point I regularly face, not only with lazy sequences but also large data structures.
The debugger tries to print the entire structure every step of the way, which leads to very poor performance:

(let [db (apply hash-map (range 100000))]
  #dbg
  (-> db
    (assoc 0 1.0)
    (dissoc 2)
    (update 4 inc)))

[bbatsov]

Of course. We just need to also find someone to tackle it.