spec loves data structures

on YouTube ↗43m

" summary

Shows that Clojure spec plus test.check are used to build white‑box invariant tests and custom generators that find subtle persistent data structure bugs (hashmap transient errors from 2013, a split bug reported 2016, AVL rebalancing omissions). Converts trees to a map representation, generates reachable and unreachable instances, runs generative checks (about 50 property runs, 14 invariant checks), fixes defects and raises confidence that the rebuilt implementations actually obey structural and performance contracts.

AI-written summary · about

" chapters (13)

1. 0:00Introduction
2. 9:10data.avl Aug 2016 (DAVL-8)
3. 11:05data.avi Oct 2016 (new invariant tests)
4. 16:12The final black box
5. 20:36Enter spec cont'd
6. 26:24data.avl leaf specs
7. 27:22data.avl :: node spec
8. 28:49data.avl :: avl-map spec
9. 29:52data.avl invariant test
10. 31:35data.avl spec'd cont'd
11. 36:51s/fdef as test specification cont'd
12. 39:41No parameterized specs
13. 41:31Next steps

" related talks

• 38mLiving with SpecSimon Belak · EuroClojure 2016
• 24mSpeculative DevelopmentAndrew Mcveigh · :clojureD 2017
• 32mGenerative Testing: Properties, State and BeyondJan Stępień · EuroClojure 2015
• 25mPowerful Testing with test.checkReid Draper · Clojure/West 2014
• 38mAbove and beyond type systems with clojure.specCarlo Sciolla · Codemotion Milan 2017
• 31mTesting Made SimpleTony Kay · Clojure/west 2017
• 48mContracts & Clojure: The Best-Yet Compromise of Types and TestsJessica Kerr · Philly ETE 2015
• 47mTesting the Hard Stuff and Staying SaneJohn Hughes · Clojure/West 2014