Yeah, I read your expound code, really nice btw. I'm currently trying something very similar to expound but with a GUI approach to explore errors.
I think encoding map paths as [key pos-in-map-entry] should work in all cases as is currently doing in s/map-of. I can't see any issues if s/keys explains in the same way, which it currently doesn't because it just reports [key]. Juan El jueves, 13 de julio de 2017, 11:27:44 (UTC-3), Ben Brinckerhoff escribió: > > No, you're not missing anything. I've run into the same issue. I see > you've found the related JIRA issue, but for those who haven't it's > https://dev.clojure.org/jira/browse/CLJ-2192 > > > So the thing is that if you see a problem in [::k 1] you don't know if > its the ::k value, or the value is a sequence and the problem is its > element at pos 1. > > Yes, you're correct, this is ambiguous right now. In Expound ( > https://github.com/bhb/expound) I have some code that tries to determine > which case it is, based upon some heuristics, but I'm sure there are cases > in which I'll get it wrong. > > There are two other interesting related cases I've run into related to the > `:in` path. > > First, there is the case where the key is wrong, not the value: > > (s/def ::spec-1 (s/map-of keyword? int?)) > > (-> (s/explain-data ::spec-1 {"foobar" 5}) > :clojure.spec.alpha/problems > first > :in) > ;; ["foobar" 0] > > This is also ambiguous in the way you've noted above, but it's not clear > how to construct this path - how do we "point to" a key of a map? > > Second, there is the case where we want to use 'coll-of' with a map. In > this case, the `in` path is just [0] > > (s/def ::coll-of (s/coll-of int? :kind map?)) > (-> (s/explain-data ::coll-of {"foobar" 5}) > :clojure.spec.alpha/problems > first > :in) > ;; [0] > > AFAICT, we don't know from looking at that path whether it means: "look up > the key `0` in the map" (incorrect) or "convert the map to a sequence, then > get the 0th element" (correct). Again, Expound has some heuristics to try > to figure this out. > > Right now, this logic is buried in the (messy) internals of Expound, but I > will soon be providing a cleaner set of functions that allows clients to do > more with the :clojure.spec.alpha/problems, including one that will provide > a `in` path that works around these issues (in most common cases). > > Ben > > On Wednesday, July 12, 2017 at 8:50:19 AM UTC-6, Juan Monetta wrote: >> >> Hi everyone, >> >> I started exploring the clojure.spec "better user errors" thing and I'm >> having some trouble matching clojure.spec.alpha/problems inside >> :clojure.spec.alpha/value in explain-data. >> >> I'm trying to understand if there is currently a way of using problem to >> directly find the part of the value that is causing the issue independent of >> the spec that failed. >> >> For example given a spec : >> >> (s/def ::spec-1 (s/map-of keyword? int?)) >> >> (-> (s/explain-data ::spec-1 {::k 8.8}) >> :clojure.spec.alpha/problems >> first >> :in) >> => [::k 1] ;; which means the problem was with the value, not the key >> >> but given a different spec for the same data : >> >> (s/def ::k int?) >> (s/def ::spec-2 (s/keys :req [::k])) >> >> (-> (s/explain-data ::spec-2 {::k 8.8}) >> :clojure.spec.alpha/problems >> first >> :in) >> => [::k] >> >> So the thing is that if you see a problem in [::k 1] you don't know if >> its the ::k value, or the value is a sequence and >> the problem is its element at pos 1. >> >> Does it make sense or I'm missing something? >> >> I'm trying with [org.clojure/clojure "1.9.0-alpha17"] >> >> -- 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.
