There is a known bug with conforming trailing ? parts inside a regex spec - sounds like you might be seeing that.
https://dev.clojure.org/jira/browse/CLJ-2105 On Tuesday, January 16, 2018 at 8:07:19 AM UTC-6, Khalid Jebbari wrote: > > Thank you Alex. > > It happens that I forgot to name the kwargs part in the s/cat call, so the > working version should be : > > (s/fdef adder > :args (s/cat :a integer? :b integer? :kwargs (s/keys* :opt-un > [::c])) > :ret string?) > > I'm asking because I've just discovered something about s/cat + s/or + > s/keys* (after 1 day and a half of trials). > > Before I added the keyword args spec, the specs of my function were like > this (simplified) : > > (s/fdef adder > :args (s/or :version1 (s/cat :a integer? :b integer?) > :version2 (s/cat :a number? :b number?) > :ret integer?)) > > This version is working just fine, with stest/check etc. > > When I added keyword args spec yesterday, I added it like this : > > (s/fdef adder > :args (s/cat :normal-args (s/or :version1 (s/cat :a integer? :b > integer?) > :version2 (s/cat :a float? :b > float?)) > :kwargs (s/keys* :opt-un [::c])) > :ret number?) > > It produced some strange behavior. Under `stest/check`, the arguments > passed to the function were in a messed up form. The smallest input found > by stest/check is `:smallest [((1.0 1.0))]` which looks like a sequence > inside a sequence. When I finally realized that I couldn't "un-factorize" > the spec of the keyword args and the put it into each s/or branch, it > worked : > > (s/fdef adder > :args (s/or :version1 (s/cat :a integer? :b integer? :kwargs > (s/keys* :opt-un [::c])) > :version2 (s/cat :a float? :b float? :kwargs (s/keys* > :opt-un [::c]))) > :ret number?) > > I realize that nesting s/cat calls can produce nested sequences, but in my > case it seemed natural at first to factorize the spec of the keyword args > and express the specificity of the "normal args" separately. It's like > check couldn't somehow "unroll" the nested s/cat... > > My question is: does it make sense ? Should it work and it's a bug ? Is > there no way to factorize like I tried (I would be fine with this really). > > I hope I made my question clear, it took me quite some time to nail it and > create a reproduction case. > > On Tuesday, January 16, 2018 at 1:36:47 PM UTC+1, Alex Miller wrote: >> >> You don’t need the final s/? but otherwise that’s right! > > -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
