Fork me on GitHub
#cider
<
2018-06-06
>
bhauman00:06:04

does nREPL middleware have a concept of finally or teardown?

bhauman00:06:27

hmmm I guess the :close message

bhauman00:06:20

yeah that's not guaranteed

richiardiandrea00:06:57

I am not sure they support that yeah...never heard of it

manuel07:06:32

hi, with latest CIDER from MELPA I get this every time I try to do cider-load-buffer:

Caused by clojure.lang.ExceptionInfo
   No such namespace: cheshire.core, could not locate cheshire/core.cljs,
   cheshire/core.cljc, or JavaScript source providing "cheshire.core"
cheshire is in my project.clj and I can run my project (and the tests) without any issues. Do you want me to report it on GitHub?

andre.stylianos07:06:24

@manuel seems like Cheshire is a clj only lib and you're trying to use it from Clojurescript?

manuel07:06:41

nope, I am doing it from a clj file

4
andre.stylianos07:06:28

Odd. But it's trying to load it in a cljs environment from what I can see. It's looking for a cljs/cljc source file

manuel07:06:19

well, I did cider-jack-in-clojurescript because it is clj/cljs project, but still I did cider-load-buffer from a clj file

😕 4
bozhidar09:06:40

> does nREPL middleware have a concept of finally or teardown?

bozhidar09:06:48

@bhauman What do you need this for?

bozhidar09:06:23

@manuel Maybe something’s wrong with the connection mapping? Can’t tell you more without being able to play with some reproducible example.

bozhidar09:06:12

I don’t think anything important has changed recently, though - except auto-injection of cljs deps. You can try disabling it to see if that won’t fix your problem, although I can’t imagine how it could cause what you’re describing.

manuel09:06:49

thanks. It'd be hard to give an example, the code is under non-disclosure agreement.

dpsutton12:06:55

I'm not sure I've seen an easy way to inhibit invection of cljs deps

bozhidar15:06:35

You just set them to nil. 🙂 It doesn’t get much easier than this.

manuel15:06:45

@bozhidar FWIW I switched to another project and everything's working fine.

manuel15:06:59

must be something messy in the other one's config

bozhidar15:06:28

Good to know!

dpsutton15:06:35

@manuel does the other one still rely on cemerick/piggieback?

manuel15:06:53

yep just noticed it and updating its project.clj

manuel15:06:01

thanks 🙂

dpsutton15:06:54

i think there's still some incompatability with figwheel piggieback and cemerick. i'm still trying to track it down

richiardiandrea15:06:01

Yesterday Bruce fixed a problem FYI in Piggieback: it was not threading in some compiler options

bhauman15:06:16

@bozhidar the figwheel repl launches a node process, and it would be nice if there was a finally block in nrepl that sends a tear-down message through the middleware stack, because if folks don't :cljs/quit from the cljs repl it seems to hang. This may be an none issue, as it may just be Reply's handling of shutdown, still investigating. But ... the nature of the system doesn't really allow for clean up.

bhauman16:06:56

hmmm it does look like its just a problem in reply

bhauman16:06:41

yeah its only a problem when I do control-D out of reply it leaves a node process running on my machine

bhauman16:06:55

but its fine when I quit out of cider

richiardiandrea17:06:55

@bhauman is the new piggieback in a SNAPSHOT? Cannot use gitlibs because of https://github.com/clojure-emacs/cider/issues/2308

richiardiandrea17:06:39

(I mean I can clone it of course 😉 just asking)

bhauman17:06:48

its not released

bhauman17:06:05

supposedly I have permission to do that, I just haven't figured it out yet

richiardiandrea17:06:40

lein deploy clojars

bhauman17:06:45

I also fixed this in figwheel.main proper to work with the current piggieback

richiardiandrea17:06:52

ok I am trying thanks!

bhauman17:06:13

@richiardiandrea but I haven't release the figwheel.main yet

richiardiandrea17:06:22

figwheel-main can now become a proper repl type 😄

bhauman17:06:48

I'd really like this to be more configurable, need to be able to supply args to figwheel.main start

bhauman17:06:17

also I think figwheel main start should probably take the command line args

bhauman17:06:30

or there should be another fn that does that

richiardiandrea17:06:30

cider usually has configurable command line strings so probably we can take the same approach

richiardiandrea17:06:47

the problem is how to make the REPL forms configurable enough

richiardiandrea17:06:30

say you want to launch different builds .... this is a more general cider improvement though

bhauman17:06:57

thats what I was refering to

👍 4
richiardiandrea17:06:12

at some point I wanted to have something like a "plugin" in inf-clojure where you declare your maven deps and a function that will be called for configuration (that sends forms to the repl)...didn't have time to expand it and dropped it 😄

richiardiandrea17:06:20

this is probably an unnecessary complication, as we could just add functions specific to figwheel, for instance, that produce the form to send...with defcustoms ad-hoc for its options

bhauman17:06:08

much easier

richiardiandrea17:06:27

and that's how a cider-figwheel.el was born 😄

bhauman18:06:22

isn't is easy to make that interactive to ask which build id you want?

richiardiandrea18:06:20

yes I just opened it for discussing more about it 😄