Open vemv opened 4 months ago
It might be similar to https://github.com/clojure-emacs/logjam/issues/16 (easy fix)
I'll see if I can produce a minimal repro case.
👋 👋 👋 @vemv. This is decently minimal. Hope it helps. :)
https://github.com/timvisher/cider-middleware-out-repro-issue-878
Thanks, Tim!
Hope we can jump to it one of these days.
I was able to reproduce this using logback as well which is documented as working in the manual. It's up in https://github.com/timvisher/cider-middleware-out-repro-issue-878 (commit). Cheers! :)
Thanks!
Just as additional context, why do you need to reconnect with a certain frequency?
It seems fairly unusual to me to have a repl connection dropped.
I'm connecting from my laptop which sleeps and wakes with some regularity. The process I'm running takes days to complete (currently) and I want to be monitoring its progress while doing other things.
Make sense? It's a bit of strange use case even in my career but doesn't seem wildly undesirable.
I've observed that any logs that are created after reconnect do get captured in the REPL buffer. And if I use a future
to manage the long running process then I can cancel it and recreate it and new logs are captured. This is less than ideal obviously but an interesting data point. This is all in the repro repo.
Thanks!
The process I'm running takes days to complete (currently) and I want to be monitoring its progress while doing other things.
In case it helps, for anything that is production-like, we don't particularly recommend having the cider-nrepl middleware stack activated.
(It's perfectly possible to have a bare nREPL server running, without cider-nrepl, and have a CIDER connected to it. It will have a more bare-bones functionality)
The reason is that cider-nrepl is generally aimed for dev-time activities and its side-effects are not guaranteed to not intefere with production matters.
With all this said, we should certainly aim to fix these. Personally I can't exactly promise when - PRs welcome as well.
The process I'm running takes days to complete (currently) and I want to be monitoring its progress while doing other things.
In case it helps, for anything that is production-like, we don't particularly recommend having the cider-nrepl middleware stack activated.
(It's perfectly possible to have a bare nREPL server running, without cider-nrepl, and have a CIDER connected to it. It will have a more bare-bones functionality)
The reason is that cider-nrepl is generally aimed for dev-time activities and its side-effects are not guaranteed to not intefere with production matters.
With all this said, we should certainly aim to fix these. Personally I can't exactly promise when - PRs welcome as well.
Very much understood. :)
https://clojurians.slack.com/archives/C0617A8PQ/p1716294437574019
This happens after a reconnection via
cider-connect
.