I'm getting a strange error when trying to run cljr-add-project-dependency:

Retrieving list of available libraries...
cljr--get-error-value: Error in nrepl-refactor: clojure.lang.Compiler$CompilerException: java.lang.ClassNotFoundException: javax.xml.bind.DatatypeConverter, compiling:(mranderson048/http_kit/v2v2v0/org/httpkit/client.clj:1:1)
 at clojure.lang.Compiler.load (Compiler.java:7391)
...

followed by a hundred lines of stacktraces involving refactor_nrepl, cider.nrepl, and nrepl.middleware

I'm running the latest versions of cider 0.21.0 and lein 2.90, the same error also happens in projects using deps.edn

this is a java9+ related bug as far as i remember

you are better off adding deps to your repl the way it is recommended in the cider docs

there is a pr in clj-refactor to make this work again but afaik stalled atm

Is it possible to use Cider for ClojureCLR?

how do you run ClojureCLR? I'm not sure CIDER speaks any of the build tools for it?

I don’t know how you run it

I was wondering what editor I should use for it

I hoped to be able to use Spacemacs

But if Cider doesn’t work I don’t know what to use

https://github.com/clojure-emacs/cider/blob/c698df1c905127f190efee3403822b2b09a811c3/doc/caveats.md#clojureclr-support

>>>CIDER currently doesn’t support ClojureCLR. The reasons for this are the following: nREPL itself runs only on the JVM (because it leverages Java APIs internally). There’s an nREPL port for ClojureCLR, but it’s not actively maintained and it doesn’t behave like the Clojure nREPL. cider-nrepl uses a lot of Java code internally itself. Those issues are not insurmountable, but are beyond the scope of our current roadmap. If someone would like to tackle them, we’d be happy to provide assistance.

Hmm that’s a bummer

ah, well there we go

🙂 thanks

no problem. perhaps inf-clojure mode can work?

it basically just forms commands that you could write yourself so in theory it doesn't care about which Clojure you talk to. (i think?)

but that would keep you in spacemacs

clj-refactor does work I guess

Except for adding dependencies

so I forgot how to add my custom aliases to a jack in command 😞

oh right cider-clojure-cli-global-options

grab this patch: https://github.com/clojure-emacs/cider/pull/2586

and use prefix argument to edit the jack in command

way easier. and way easier to randomly set it to whatever you want.

I saw that, do I understand right that you basically need to modify the whole string?

you don't have to. it plops the whole command in the minibuffer so you can see exactly where your modifications need to go

so do that, go to beginning of line, and add -A:fig or whatever you need to do

its so that every readme that shows which command line form to use is very easy to follow

by the by, i'm running make autoloads and it doesn't seem to be working for me

@dpsutton I think I do make and it does it all now

what is the error?

i've tried both. i get the classic command-execute: Wrong type argument: commandp, cider-find-var

ah

let me try here, I am on master

uhm I do not see it either

weird I see the section in cider-autoloads.el:

;;;### (autoloads nil "cider-find" "cider-find.el" (0 0 0 0))
;;; Generated autoloads from cider-find.el
(autoload 'cider-find-var "cider-find" "\
(autoload 'cider-find-dwim-at-mouse "cider-find" "\
(autoload 'cider-find-dwim "cider-find" "\
(autoload 'cider-find-resource "cider-find" "\
(autoload 'cider-find-ns "cider-find" "\
(autoload 'cider-find-keyword "cider-find" "\

so its supposed to build up a file .depend and that never has any data written to it

question for you @dpsutton I still don't really understand if use-package is supposed to load cider-autoloads.el for me (kind a new to it)

i've never used it either

ok cool tnx 😉

> If you're not using one of tho mechanisms that registers autoloads, such as :bind or :hook, and your package manager does not provide autoloads, it's possible that without adding :demand t to those declarations, your package will never be loaded.

it looks like it gets a little manual. can you put that (load "cider-autoloads" t t) form right after use-package?

I was going to try exactly that

😄

this is also interesting: https://emacs.stackexchange.com/questions/19695/use-package-defer-t-and-add-hook

as I understand that you could just declare the function you are interested in and use-package created the autoloads basically

this one is also a good one to know https://emacs.stackexchange.com/questions/19694/use-package-defer-t-and-autoloads

this works for me:

(load "cider-autoloads" t t)

(use-package cider
  ;; This seems enough for cider, see also:
  ;; 
  :defer t

  :init
  (setq cider-popup-stacktraces t
        cider-popup-stacktraces-in-repl nil
        cider-repl-use-clojure-font-lock t
        cider-overlays-use-font-lock t
        cider-repl-wrap-history t
        cider-repl-history-size 1000
        cider-show-error-buffer t)

  :config
  (add-hook 'cider-mode-hook 'eldoc-mode)
  (add-hook 'cider-mode-hook 'paredit-mode)
  (add-hook 'cider-repl-mode-hook 'paredit-mode)
  (add-hook 'cider-repl-mode-hook 'eldoc-mode))

only thing I am missing is how to bind cider-insert-last-sexp-in-repl to C-c C-c

also to scroll the REPL when cider-insert-last-sexp-in-repl is doing its thing, @dpsutton do I have to set something for the scroll part?

Not that I know of. Is it not scrolling with the input?

Might just need to be fixed.

Nope, I think that feature has been changed/tweaked - I will read more

I put (setq scroll-conservatively 101) like I read somewhere but I don't see it scrolling unfortunately

from the CHANGELOG: > * Option cider-repl-scroll-on-output is now obsolete, and the default REPL behaviour has changed to not recenter the window. The built-in variable scroll-conservatively can be set to 101 (either globally or locally in the REPL buffer) to restore the old behaviour. This change has a dramatic positive effect on REPL performance.

with (setq scroll-conservatively 0) it does not do it either

Yeah, scroll-conservatively only has any affect if the point is moved – but it seems the insert-*-in-repl commands leave the point at the start of the input

yes @cichli that's what happens, I see it as well now - maybe @dpsutton I can open an issue? no rush on this, just gathering some thoughts

I think the right thing to do is probably move the point to the input end… but I haven’t really used these commands so not sure if there’s a good reason they don’t do that 😄

@richiardiandrea I’ve pushed another attempt at fixing https://github.com/clojure-emacs/cider/issues/2589 if you want to test again

let me try live 😉

I’m curious as to why we end up calling buffer-substring-no-properties on the *cider-result* buffer rather than the source buffer… we explicitly don’t select it when we display it, but possibly something in your config is altering that

it can be, I have a possible bloated Emacs Live custom branch

I am trying to trim it down as much as possible

@cichli it seems like it is working now

All the possible config around display-buffer is quite confusing so it’s possible we’re missing an edge case in the popup code

Great news! 🎉 Thanks for testing

😄 no problem, thank you for debugging!

this is the first time I am trying this feature actually, my workflow is quite different, usually I just send to the REPL buffer, having a dedicated result window has its advantages though

good stuff

In my cider repl buffer, *e used to print stacktraces, now it just prints

*e
#object[java.lang.AssertionError
        "0x5d79b28d"
        "java.lang.AssertionError: Assert failed:

(setq cider-print-fn 'pprint)?

Or do you mean printed like by clojure.repl/pst?

I’m not entirely sure who is responsible for printing the exception, but at the moment, it’s behaving as if it’s doing the clojure.core/print-method default, rather than pst

Is that a cider behavior, or my repl?

hrm, appears to be cider-related, because the lein repl prints the whole stacktrace, as expected, when I *e