Shen Tian

Ah ha. History makes sense. Thoughts on having middleware depend on `nrepl.config`? any foreseeable problems with that?

👋🏽 Just catching up a bit on what's happening, and this is very exciting. I'll need to get myself up to speed a bit first, but should then be able...

I hope #254 improves things with `tty-server` and `basic-fs-socket-behaviour`

Sorry I've been awol. I just had a look through changelog, and nothing jumps out at me. The exactly 1 minute part is also curious. The only thing close I...

Might be missing something... but don't you just recombine the streamed results, and check for the first/second/nth complete form? The difference between the expected and actual is a client side...

Oh right. To pick a corner case, if we do streamed printing with buffer size 1, then both `1 2` and `12` yields two messages, with values of `1` and...

I just experimented a bit around this, by setting `nrepl.middleware.print/buffer-size` to 1, as per my previous comment, in Cider. Turns out streamed printing won't break `12` into `1` and `2`....

@tpope : think this requires some parsing, but naively, it looks like it's mostly paren matching? Can you clarify a few things: 1. Is this a "can't tell apart" or...

Thanks @tpope. That makes perfect sense. Having had a quick look, the last line here sends that last message https://github.com/nrepl/nrepl/blob/f189b4c07579f8b6752f2e385e52c3680c2e1d60/src/clojure/nrepl/middleware/print.clj#L122-L136 I'll need to test a bit, but looks like a...

I've had a proper look at this, and I think having a key indicate the end of a stream is a good, non-intrusive solution. How the streamed printer works, right...