https://dev.clojure.org/jira/browse/CLJS-2975

I would imagine that’s working as intended or somehting

There might just be some trickery with clojure.core being aliased to cljs.core

or some macro trickery involving + which sometimes is a macro in Clojurescript.

there never was an fdef to instrument, so unstrument should always return empty

is there a macro fdef for + in cljs? hmm, that could explain it

I think I’ve found a bug with instrumenting args, at least on my machine

(defn defx [key & [doc]])

(s/fdef defx
        :args (s/cat :key keyword?
                     :doc (s/? string?)))

(st/instrument)

(defx :a "")
;; => not thrown - expected

(defx 1)
; => throws - expected

(defx 1 1)
;; => not thrown - bug

(defx 1 1 1 1)
;; => not thrown - bug

(defx :a "" 1)
;; => not thrown - bug

it seems to validate the length of the sequence in some cases but not the members

that spec is wrong btw

or your def is wrong, hard to tell

well, I guess maybe it’s not wrong, just hard to tell what you’re expecting

but there is a known issue with conformed values of trailing optional elements and you may be running into that

I’ve tried something like: (s/fdef defx :args (s/cat :key keyword? :opts (s/spec (s/? string?)))), that works.

btw, (defx 1 1 1 1) does throw on clj with the old spec. I’ll try on cljs next

you don’t need the s/spec - & is automatically creating a sequence and there is no nested collection here

right

it’s weird in that the def allows N args but the spec constrains to 1 or 2

The spec I intend is that it should accept 1 optional string arg

I can see that the problem reporting is failing though which I think is the bug I mentioned above

I did call instrument, and calling with 1 Arg throws

what version of cljs are you on?

It was throwing in my clj repl as expected

Latest cljs

Earlier versions have a different bug with varargs which was recently patched

https://dev.clojure.org/jira/browse/CLJS-2793

Ach I've got to run, be back in an hour or so

can you make a repro? this repro does work:

$ clj -Srepro -Sdeps '{:deps {org.clojure/clojurescript {:mvn/version "1.10.439"} org.clojure/test.check {:mvn/version "RELEASE"}}}' -m cljs.main -re node -r

ClojureScript 1.10.439
cljs.user=> (require '[clojure.spec.test.alpha :as stest])
nil
cljs.user=> (require '[clojure.spec.alpha :as s])
nil
(s/fdef defx
       #_=>         :args (s/cat :key keyword?
       #_=>                      :doc (s/? string?)))
cljs.user/defx
cljs.user=> (defn defx [key & [doc]])
#'cljs.user/defx
cljs.user=>  (stest/instrument)
[cljs.user/defx]
cljs.user=>  (defx 1 1 1 1)
Error: Call to #'cljs.user/defx did not conform to spec.

(btw Alex, making repros like this is fantastic with tools.deps)

ok, am working on a repro

might be something with my test.check version

Can repro - the issue occurs when :static-fns compiler option is true

well. I haven’t been able to repro with the cljs.main repl

but in my build I can toggle the behaviour by toggling :static-fns

still trying to figure out if i am passing the options correctly to clj

ok, got it. start repl with :static-fns true:

clj -Srepro -Sdeps '{:deps {org.clojure/clojurescript {:mvn/version "1.10.439"} org.clojure/test.check {:mvn/version "RELEASE"}}}' -m cljs.main -co '{:static-fns true}' -re node -r 

(require '[clojure.spec.alpha :as s])
   (require '[clojure.spec.test.alpha :as st])
   
   (defn defk [key & [doc]]
     key)

   (s/fdef defk
           :args (s/cat :key keyword?
                        :doc (s/? string?)))

   (st/instrument)

   (defk 1 1)

the last line should throw, but doesn’t with :static-fns enabled

what happens if you make defk a multi-arity function? just curious

i wouldn’t be surprised if it works, the other bug linked above only affected varargs. but let me check

yeah, it works as expected with a multi-arity function

at this point I’d say JIRA, but maybe @mfikes or @dnolen has something sensible to say

i’ve added a comment to the related issue

though I apparently do not know how to get jira to format code :man-shrugging::skin-tone-2:

This unfortunately appears to also be a regression relative to 1.10.339

@mhuebert replace backquotes by {{ ... }} and triple-backquotes by {code} ... {code}