This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2019-07-10
Channels
- # announcements (3)
- # beginners (67)
- # calva (4)
- # cider (3)
- # clj-kondo (58)
- # cljs-dev (4)
- # clojure (172)
- # clojure-berlin (4)
- # clojure-chicago (8)
- # clojure-europe (4)
- # clojure-greece (8)
- # clojure-italy (12)
- # clojure-nl (4)
- # clojure-spec (7)
- # clojure-uk (77)
- # clojurescript (13)
- # clojutre (16)
- # core-async (10)
- # cursive (3)
- # datomic (29)
- # figwheel-main (27)
- # fulcro (22)
- # garden (3)
- # jobs (2)
- # jobs-discuss (13)
- # juxt (5)
- # leiningen (14)
- # mount (4)
- # off-topic (28)
- # pathom (6)
- # pedestal (8)
- # portland-or (2)
- # re-frame (20)
- # remote-jobs (6)
- # shadow-cljs (13)
- # sql (74)
- # testing (17)
- # tools-deps (1)
- # vim (1)
- # xtdb (1)
The second two args are new
@arohner the uuid generator is fully random with a uniform distribution, so you shouldn't be any more likely to get collisions than you are in Real Life, i.e. it shouldn't ever happen
My guess is you're seeing the effects of shrinking, where your test gets run repeatedly with similar data
So the UUIDs won't shrink, but your test is run again, so if you have state in a database that you're not resetting between trials, that'd be a problem
I agree it's annoying, because UUIDs otherwise seem like such an effective way of implicitly partitioning a stateful external system by trial
I'm curious whether, in practice, blasting lots of data into a test DB w/o deleting it is faster than truncating between trials