Fork me on GitHub
#clojure-spec
<
2016-06-20
>
Alex Miller (Clojure team)02:06:06

there will likely eventually be a deep-describe, which will require something like that

alqvist13:06:51

I feel that the usefulness of s/inst-in is hampered by cljs incompability. Any words of wisdom?

alqvist14:06:03

Or I could just use the latest clojurescript - Never mind

angelini15:06:32

is there anyway to reproduce the behaviour of alpha-5’s s/instrument without using spec.test? The rational of only checking :args makes sense, but it was really useful when testing code at the REPL to verify that :ret and :fn we’re correct.

adamfrey15:06:18

when you have a test.check generator that is failing a such-that predicate 100 times, is it possible to print out the failing values, just to make the generator less opaque? Nothing stood out to me in the source code for doing that.

akiel17:06:05

Are there plans to support some kind of validation context? I think about a spec for JSON Web Tokens which are signed. Such a validation would need the public key as context. Or do I stretch the intended usage patterns here?

Alex Miller (Clojure team)19:06:30

@adamfrey nothing explicit but you can gen/sample the s/gen of some simpler part to get an idea. It's usually the first pred in an s/and that is creating things that won't pass the later preds in the s/and

Alex Miller (Clojure team)19:06:17

Prob a sign you are taking things too far :)

Alex Miller (Clojure team)20:06:54

@cigitia: this channel is for spec, please take to #off-topic

cigitia20:06:57

My apologies; accidentally sent to wrong Slack team; meant to sent to personal friends