Fork me on GitHub
#cider
<
2018-07-05
>
bozhidar05:07:27

Just a quick FYI - the nasty problem with cider-test.el is finally fixed. This time - for real. 😉

👏 28
orestis10:07:47

Do I have to do something special to get expound to work with a CIDER-started REPL? I haven’t been able to get it to work.

orestis10:07:20

Also, I’m using io.aviso/pretty and it works on the actual REPL output, but not on the *cider-error* buffer.

bozhidar11:07:30

> Do I have to do something special to get expound to work with a CIDER-started REPL? I haven’t been able to get it to work.

bozhidar11:07:51

No idea how this works. Explaining the problems you’ve got with it would be helpful.

bozhidar11:07:59

> Also, I’m using io.aviso/pretty and it works on the actual REPL output, but not on the *cider-error* buffer.

bozhidar11:07:24

Because this buffer is now a raw rendering of the stracktrace, but it’s custom built by CIDER.

bozhidar11:07:16

I think we had some setting to just display the raw stacktrace, but I don’t remember it off the top of my head.

orestis11:07:10

Sorry, I should have been more clear.

orestis11:07:27

It could be I’m misunderstanding how expound works, but I think it taps into the clojure.spec *explain-out* function. You are supposed to set it up either with a binding, or by using alter-var-root to bind *explain-out* for all the threads. I’ve never seen this in action, so I don’t know if this is a lein issue or a CIDER issue.

orestis11:07:37

(or an expound issue, for that matter)

orestis11:07:18

I’m only asking here just in case it’s a known thing with CIDER — if it’s not common then please don’t waste time, I’ll debug it on my own, as it could be just some weird project setup I have here.

orestis11:07:00

Thanks for the navigating stacktraces link — I can see it’s very powerful, I had no idea.

dpsutton12:07:51

@orestis I have used expound and calls to 'expound/expound' have worked without a hitch

orestis12:07:58

@dpsutton Thanks for the vote of confidence. What about Compiler Exceptions? I was expecting that those would be formatted using *explain-out* as well, but perhaps I’m misunderstanding.

andre.stylianos14:07:57

I'd appreciate it if you let me know what you discover

andre.stylianos14:07:21

That also happened to me, but I didn't have time yet to understand what was going on

andre.stylianos14:07:00

using expound/expound directly works fine, but binding *explain-out* doesn't :thinking_face:

dpsutton14:07:37

i've never tried hooking it into the spec apparatus like you are trying to do unfortunately. I've only used it for explicit calls to expound.alpha/expound

dominicm14:07:41

When using clojure tools cider-jack-in doesn't inject piggieback. Should/Could it?

dominicm14:07:56

(This is 0.17.0, might be master fixed)

andre.stylianos14:07:06

shouldn't it be injected only when using cider-jack-in-clojurescript?

dpsutton14:07:07

it looks like the autoinject piggieback commit comes after 0.17 @dominicm

dpsutton14:07:44

i believe he is right

dominicm14:07:56

is there a different between -cljs and -clojurescript?

andre.stylianos14:07:25

I just listed the command from the top of my head, don't know which one it is exactly 😛

dpsutton14:07:45

0.17 predates the huge change of the connection api. i believe cljs migrated to clojurescript

dpsutton14:07:39

@orestis i see it alters a var root. this may be similar to the map printing argument *print-namespace-maps* not being honored. I think these bindings are per thread and nrepl uses 4 threads from a thread pool (i think). I think there is currently a bug in the way some of these bindings are propagated

tianshu15:07:06

anyone run into this issue? the new feature eval toplevel in comment breaks the auto completion in comment.

dpsutton16:07:19

no i've never seen this.

dpsutton16:07:21

let me look

dpsutton16:07:59

@doglooksgood i can get autocomplete here. are you unable to do this?

dpsutton16:07:58

ok i'm able to reproduce with a function from the same namespace

dpsutton16:07:39

immediate workaround: (setq cider-eval-toplevel-inside-comment-form nil). I put this behind a flag to disable it if it ever caused problems

dpsutton16:07:04

@doglooksgood if you want to file an issue you can include this workaround on the issue so if people google for it

tianshu17:07:47

cider-eval-toplevel-inside-comment-form is by default nil

tianshu17:07:30

I'm not sure which case it not work.

tianshu17:07:50

also, will it affect the ns generation?

dpsutton17:07:23

I have that set to true. I see the bug with it true and don't see the bug with it set to nil. I put that enabling flag precisely because it's so important

dpsutton17:07:42

Do you see this bug regardless of that custom var?

tianshu18:07:53

bug only exists when flag set to true.

dpsutton19:07:39

excellent. at least there's an easy workaround right now. thanks for the bug report

bozhidar20:07:17

> shouldn’t it be injected only when using cider-jack-in-clojurescript? (edited) Yes.

bozhidar20:07:28

> is there a different between -cljs and -clojurescript?

bozhidar20:07:23

@dominicm No. We just decided to add a shorter alias to the original name. In general we’ll favour cljs over clojurescript in command names for the sake of brevity.

bozhidar20:07:10

We’re also adding clj suffix to cider-jack-in and cider-connect, so it’s clearer to newcomers how they differ from the cljs commands.

bozhidar20:07:05

0.18 (`master`) features a ton of cleanup and unifications of various parts of CIDER - a big step in the direction of the dream 1.0 release.