This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2016-07-31
Channels
- # admin-announcements (1)
- # beginners (1)
- # boot (91)
- # cljs-dev (11)
- # cljsjs (8)
- # cljsrn (38)
- # clojure (89)
- # clojure-dusseldorf (1)
- # clojure-russia (6)
- # clojure-spec (14)
- # clojure-uk (4)
- # clojurescript (25)
- # data-science (1)
- # datascript (3)
- # datavis (4)
- # emacs (25)
- # events (1)
- # jobs-rus (3)
- # off-topic (1)
- # proton (3)
- # re-frame (19)
- # reagent (5)
- # specter (10)
- # yada (1)
that's a nice explanation
(I’m just repeating what Stu said in the latest Cognitect clojure.spec 'cast! 🙂 )
@alexmiller: the call to generate
here undermines test.check determinism (similar to CLJ-1949): https://github.com/clojure/clojure/blob/f374423053b75b7b484ffbc8b49b2ede9d92e406/src/clj/clojure/spec/test.clj#L155
in particular if you're running check
while instrumenting with stubs
I don't think there's as straightforward a fix as in CLJ-1949 though
probably the cleanest thing would be a private dynamic var somewhere for registering quickcheck-driven generation fns
it could use gen/generate
as the default when you're not actually running a test.check test
shall I make a ticket?
@gfredericks: FYI, I made a bunch of progress on the JSON Schema thing, knocking a few remaining ones down; the really really hard one is references (idk if you’re familiar with JSON schema, but basically you have a base URL and URL-y internal/external references, so you need a way to resolve things like #../../xyzzy/0 or whatever). Dunno if it’s going to be of a size where it’s interesting for spec, but we’ll see