Dealing with Test Results
summary
Argues to put all test results in CouchDB and demonstrates a simple single‑page app (node‑couch‑app, Samy JS, Futon Samy branch) loading ~3,000 run documents, using map/reduce views keyed by report type, system and numeric build ID to get sortable history. Builds a new‑failure detection view that queries two most recent builds per test to flag regressions; demo mostly works but has a filename‑emission bug that inflates “new” failures.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:15CouchDB for test results
- 3:23Test results database structure
- 7:10Single-page CouchDB app
- 10:00Viewing test results by type
- 14:17Finding new test failures
- 18:53Real-time failure queries
- 21:26Key takeaways and open discussion
related talks
40mTest trend analysis: towards robust, reliable and timely tests
45mAchievement Unlocked: Drive development, increase velocity, and write blissful...
08mFinding and Fixing Strange Builds
36mDeveloping Tools, While Testing
41mUsing ML to find value in your automated tests
01h 00mTesting Heresies
34mThe test report is dead, long live information!
42mPlanning To Fail