@malabarba: reopening the file did help. Thanks!

Don’t forget to mention that in your future blog post;)

could anyone help me with specifying an indent spec?

the docs are a bit confusing

When I use cider with boot, cider-find-var jumps to a path in my boot tempdir. Is there a way around that?

looks like this should be handled by https://github.com/clojure-emacs/cider-nrepl/pull/270

hm... I'm using boot 2.5.5, cider/cider-nrepl "0.12.0" and cider 20160629.946

still jumps to the tempdir

@malabarba: Did you have any ideas on getting *nrepl-server* to mirror its messages to *cider-repl*?

I’m getting this error now when starting cider:

[nREPL] Establishing direct connection to localhost:49716 ...
error in process filter: [nREPL] Direct connection failed [2 times]

it just started yesterday

I upgraded cider this morning and still have the same issue

i have no idea what might have caused it. I haven’t modified my emacs configuration in months

hmm, nevermind, I figured it out. Something screwed up my hosts file.

@radon what emacs version are you on?

24.5.1

@pesterhazy: It works fine for me. Are you sure these are boot tempdir and not clojurescript tempdirs? Also, maybe try 0.13.0-snapshot

@radon: Try this: `

(add-hook 'cider-connected-hook
          (lambda ()
            (save-excursion
              (goto-char (point-min))
              (insert
               (with-current-buffer nrepl-server-buffer
                 (buffer-string))))))

It'll just insert on the repl buffer everything that's on the nrepl buffer when you connect

What if it never connects? When something goes terribly wrong, e.g. with the dependencies, sometimes the REPL fails to even start, and I just end up waiting five minutes before I check *nrepl-server* and realize that the *cider-repl* buffer is never going to open. (This is the primary reason that the default behavior annoys me.)

But I will try your code, for sure.

ah

that should never happen

if the server fails to start Emacs should tell you

I think I’ve gotten that a few times. But maybe I just needed to wait longer?

I guess that’s possible.

In that case cider prints the stacktrace on the *messages* buffer

Oh, well, yes, but *messages* isn’t open by default.

C-h e will open it. But you should still get a message on the echo-are

That’s probably true as well, but since the REPL for a large project can take several minutes to start up, I’ll be in the middle of other editing by that point.

In fact, I think the message is the entire stacktrace, so it usually resizes my whole window to make the echo-area fit the stack trace. It's quite hard to miss ^^

ah

that's possible

We could popup the buffer in that case anyway

The user is probably interested in the exception

Yes, I think that would be good default behavior. I would still prefer the REPL buffer to open immediately, so it can show log messages even if the REPL server hasn’t properly started, but I can see why other people wouldn’t want that.

It's not that people wouldn't want that. I think it's more that it would complicate code.

The repl buffer is created when the repl connection process is started. Separating the two would not be worth the complications.

And we can't make the repl connection any sooner than we already do, because that's up to lein/boot, not us.

I am really very knowledge-less in the field of elisp, but it seems to me that it wouldn’t be too hard to make cider-jack-in create a buffer named *cider-repl*, and then replace the code that creates the buffer with code that uses the existing buffer (or creates a new one, if the first one has been killed by the user in the interim).

But, you know more than I do about how this works.

Except we already have code that uses pre-existing buffers (which applies to cases where a REPL has previously existed), and it would conflict with this situation. 🙂

Ohhh…

Of course we can workaround it.

But that's how complexity slowly increases. 🙂