The latest version of cider broke cljs repl type customization for me, no matter the value of cider-cljs-lein-repl cider always starts a rhino repl with cider-jack-in-clojurescript

I used to just set (cider-cljs-lein-repl . "(do (load-file \”script/dev.cljc\"))") in my .dir-locals.el, I also tried using setq but I get the same behavior

@jfntn latest Snapshot? Can you open an issue, if it is the latest Snapshot I changed some stuff in there

richiardi, didn't you rename that var?

@richiardiandrea yes I saw that, was about to mention you 😉 Any ideas how I might go about trying to fix this in the meantime?

No I did not rename the var, I added to that boot and Gradle var and a cider-cljs-repl-form that queries depending on the project

Maybe the logic is a bit off, I tried to reproduce the old behavior (of course), I will have a look today

Thanks for reporting and sorry about that @jfntn

ah. were you going to rename it at one point and decided not to?

just wondering where I got that wrong impression

Ok I think I saw what is going on

and now i remember yall deciding not to and keeping the gradle stuff around

Yep but it will be another pr, Gradle is all over cider 😄

haha

@dpsutton do you know what I need to add to my init.el in order to format elisp code correctly (for instance describe in cider tests) ?

https://github.com/clojure-emacs/clojure-mode#indentation-of-macro-forms

i've never done it before though

yeah that's for clojure, what about elisp?

oh

sorry

oh ok 😄

hmm

i've seen it before

but only in the definition of macros

but i've seen what you're talking about

what did i do

i went to the definition of the macro

and eval-ed it

because it was only brought in after the autoloads

so, navigate to source, C-M-x that badboy, pop back out to your describe form and reformat it

it should work

oh wow ok

its because you haven't loaded all of buttercup yet

just the autoloads

yes I don't think I have

thanks I'll play with it

let me know if that works

i remember that working for me

mmm if I require it I don't even have it in my emacs packages...ok

its brought in transitively i think

I will format manually, I want to try that problem above

oh you mean in your elpa directory?

yeah

I don't use elpa

but that's another story 😄

uhm, the test succeeds

not a very big elisp user here but:

(describe "cider-cljs-repl-form"
          (before-each
           (setq-local cider-cljs-lein-repl "(do (load-file \"script/dev.cljc\"))"))
          (describe "when there overriding cider-cljs-lein-repl"
                    (it "returns the overridden form not the default"
                        (expect (cider-cljs-repl-form "lein") :to-equal "(do (load-file \"script/dev.cljc\"))"))))

uhm scratch that there is indeed a problem

@richiardiandrea ok cider-cljs-repl-form works as expected with a cl-cases instead of a pcase

uhm

it is weird, I have a successful test

what is the difference if I may ask? not a big elisp user

not sure to be honest, I don’t understand pcase

Hmm actually it might need more debugging because eventhough it says “starting a custom repl” nrepl doesn’t actually run my script

what’s the testing library?

@jfntn did you try to ielm on your buffer

buttercup

in cider you just make clean test

it seemed to work once, but now I get a make test error running the test suite

will need to investigate later

ok, whenever you have time, I will open a PR with the tests, we can talk there as well

Ran 254 specs, 0 failed, in 2.3 seconds.

oh I do get 1 failure though for something unrelated

Could be an emacs version thing, I built from master recently

ah, then it could be

@jfntn PR opened, thanks for reporting, you pushed me to add tests 😄

@jfntn is the failure related to the new clj/cljs reader font lock thing?

the test failure I mean

@dpsutton the test was reader conditional font-lock when cider is connected uses cider-reader-conditional-face

yup

that one fails on my machine but not the CIDER build tooling

its confusing to me

difference in emacs versions maybe?

i'm building off of master

is that what you are doing as well?

yes

well there we go

i'm on probably a 2month old commit or something

I have a more recent build `GNU Emacs 26.0.50.2 (x86_64-apple-darwin15.6.0, NS appkit-1404.47 Version 10.11.6 (Build 15G1004)) of 2017-01-15 `

wow that's good that you guys try already against that

is cider-jack-in got broken, or I am missing something?

what issues are you seeing?

@ag i just jacked in from the latest 0.15 snapshot on windows

what version are you on?

CIDER 0.15.0snapshot emacs 25.1.1

I have rebooted, trying again

please let me know

hope it works for ya

ok, it seems cider-jack-in works, cider-jack-in-clojurescript is broken though

ah, I think @richiardiandrea is working on that

@ag what do you see?

is there an issue where I can follow up?

All the tests pass here, do you have a dir-locals.el?

@richiardiandrea null pointer exception and long stacktrace related to weasel

yes, and my dir-locals.el

((nil . ((cider-boot-parameters . "repl -s wait")
         (cider-cljs-lein-repl . "(boot (start))"))))

it was working yesterday

Ok @ag for your case you need to change to boot-cljs-boot-repl

trying that

same thing 😞

are you using boot-repl-cljs?

ah yes it is written there

I wonder what (boot (start)) is doing

another thing is, you could try to remove the dir local

because the default is to launch boot-cljs-repl/start-repl

(deftask start []
  (future (boot (dev)))
  identity)

oh no ok

so basically you are launching two repls it seems

right

so it was working, nothing has changed except cider

I would actually try this in your case:

((nil . ((cider-boot-parameters . "start")))))

and you start will become

(deftask start []
  (future (boot (dev)))
  (comp (repl :server true) (wait))

let me try that, because removing dir-locals didn’t help

when you jack in, at the bottom you should see the task called that is matching your start

it is basically same as my setup

can you paste exact content of .dir-locals.el please

((nil . (cider-boot-parameters . "dev-web"))))

which you will replace with start

yes there was one parent more up there

problem not in parens, I spotted that, something else not right

something wrong with deftask I think

it’s not working in terminal either

uhm, I did not try, let me see

oh, hold on… I think I missed paren there 🙂

missing paren yeah

"try Lisps" they say, "parentheses not a problem” they say.

ah ah, I should have add highlight maybe

test

clojure
(deftask start []
  (future (boot (dev)))
  (comp (repl :server true) (wait))

no it does not work like this 😄

is it working now?

of course 😄

@richiardiandrea Holy shit!!! that worked… this is so cool. I can’t thank you enough… this been killing me for the past couple of hours

no problem, that's good that it is solved!

you didn’t just fixed it, I couldn’t find a way to automatically start the cljs-repl, it would still have 2 separate clojure repls until I type in one of them (start-repl)

yooohooo!

yeah your setup was a bit weird 😄

by the way, I think you could get rid of repl :server wait even

because cider will always launch two repls

the second one is reduntant (unless you use it with another classpath)

well, to paraphrase Tolstoy: “All Sublime Text configs are alike; each Emacs config is weird and awesome in its own way"

oh yeah Tolstoy for sure was using emacs 😄

you bet he did. that’s why he rejected orthodox church. he found better one

ah ah ha

that might actually be true, we should check the dates 😄

@ag do you like having two repls open?

i was always confused why emacs would launch two clojure repls and then make one a cljs repl

and i wanted to do away with that

(in the future)

just getting ideas/opionions

@dpsutton absolutely, two separate clojure and clojurescript repls are amazing. Something I believe Cursive can’t do. Well it can’t do a lot more, to be honest

i wouldn't preclude that

but when you jack in clojurescript, you get a single clojure script repl

and if you like, you could jack in to get a clojure repl

i just don't get why it tries to do both automatically

@dpsutton I don’t get what you’re saying. They are both indeed clojure repls (until one of them starts cljs-repl), and even after one becomes cljs-repl there’s always way to do :cljs.quit and go back to clj repl

but why would you need 2 clojure repls? is that what you’re talking about

that's my point, ag

when you jack-in-clojurescript you get two repls

but its actually a clojure repl, and a clojure repl that just invoked whatever you invoke to get a cljs repl

so why not just make one?

i don't understand why there are wto

or why there are always two

when you do clojure-jack-in, you get one repl

correct

so why when you do clojurescript-jack-in do you get two?

because then you can evaluate clojure code in one and clojurescript code in another, without having to run :cljs.quit and (start-repl) back and forth

is that a common use case in clojurescript projects?

i'm not familiar with clojurescript, but it seems like they don't require a jvm at al

in my case it is

ah ok

then i'm glad i talked to you then 🙂

this is really cool, so you can go to clojure file, evaluate piece of code and see the results, and then you can go to clojurescript file and evaluate piece of clojurescript code and see the results.

really nice