Testing in Legacy: From Rags to Riches
summary
An 11‑year-old Java codebase with no docs, ~100 database tables and 400 million records gets measurable fixes: heavy queries that took 3 minutes → 3 seconds after schema and query work, missing foreign keys are restored, and core domain classes are isolated for unit tests. The team adds PowerMock and Mockito for unit testing, builds Jenkins + JUnit integration suites and an automated deployment pipeline, which cuts manual testing and nightly deployment risk and makes incremental refactors safe and repeatable.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 3:54What is legacy
- 7:24Problems and first step
- 12:52Team knowledge
- 15:45Domain model breakdown
- 21:11Testing legacy systems
- 25:45Test automation
- 31:06Refactoring limits
- 37:11Deployment management
- 42:15Teamwork and management
- 45:10Summary and tips
related talks
49mMake your tests tell the story of your domain
45mTesting Legacy Code - What To Do When It Seems Untestable
48mLegacy, for lack of a better word, is good
55mModernizing Legacy Code - While Not Bringing Down Prod
28mAcceptable Upgrades
10mBeyond “try harder”: Effective strategies to tackle bugs
51mTest-Driven: from Zero to Hero
42mTest-Driven Repair