Bozhidar Batsov

It's hard for the server to tell apart the different output messages. I was thinking back in the day that maybe they can be processed differently by clients based on...

@arrdem Did you see what Chas had written? > Most users want to put an nREPL port on their app, and that should always be on a typical port. Auto-selection...

> My 2c: I actually was thinking that a mechanism like .nrepl-port would be sufficient, most of the tooling now understands that - for local nrepl that is. Yeah, that...

For reference - here's the existing documentation https://nrepl.org/nrepl/0.8/usage/misc.html#symbol-lookup Normally, this would be in the https://nrepl.org/nrepl/0.8/ops.html section, but it's a bit hard to have a lot of info there, as this...

There's more on the subject here https://github.com/nrepl/nrepl/discussions/275#discussioncomment-2752278

The real issue is deeper and it has to do the with the stderr and stdout bindings of nREPL. Basically output from background threads ended up in the console when...

After digging around I found a few references in the history here https://github.com/clojure-emacs/cider/pull/1376 Long story short - I think we need something similar to the out middleware to be bundled...

Great analysis of the current situation! Thanks! > The proposed solution by altering the root vars will break the out middleware for sure because the latter writes its own forked...

@vspinu Btw, if you take a closer look at the PR you'll see it also address the problem with the missing request ids, which is just as painful as the...

Frankly, I don't recall why we implemented the subscription model. I think @malabarba would know best. Maybe you can dig into the the history of the middleware in git and...