1000+ Compile-time Errors Later and Running Smoothly in Prod
summary
Claim is that pure FP pays: a critical HTTP service moves from upgraded http4s 0.16 → 0.20, survives CI and deploys to prod after 1000+ compile and test-compile errors. Migration replaces legacy effect with replaced scalaz Tasks with cats-effect IO, swaps HTTP/JSON stacks incrementally, and updates Dubey and fs2 compatibility. Changes land as small, reviewable PRs, unit tests pass, and the service runs in production with only one post-deploy bug from a streaming logic mistake.
AI-written summary · about
chapters (7)
- 0:00Start
- 0:10Agenda
- 0:58My Background
- 1:39Critical Web Service
- 2:20Problem (circa Feb-2018)
- 4:57Upgrades' Pull Requests (continued)
- 7:30Conclusion
related talks
31mBuilding a Sustainable CodeBase with FP
01h 02mFishbowl @ Functional Conf
27mTrivially bug free code for a greener planet
01h 15mFunctional Mechanical Sympathy
29mA Skeptic's Look at Scalaz Gateway Drugs
40mSmoothing the continuous delivery path- a tale of two architectures
33mChopping the monolith
52mBuilding Scalable Web Applications with Haskell