Episode goal: convert theory into muscle memory. Viewer watches a service go from nothing to a running endpoint, and sees the two-module structure appear.
Cold open
On camera: “Enough theory. Let’s generate a real Chenile service, build it, and call it — and watch the definition/implementation split appear on disk.”
Prereqs
On screen checklist: Node, GNU
make, Maven, a shell. “If you have these, you’re ten minutes from a running service.”Demo — setup
Terminal: clone
chenile-gen, cd app-gen, run make, add bin to PATH. Narrate: “make compiles the generator and grabs Mustache. One-time setup.”Demo — config
Run
gen.sh → “create a local config”. Open config/setenv.sh; set company and org. “These become your Java package names — set them once.”Demo — generate
Run
gen.sh → “create a normal service and monolith”. Service name stringdemo; monolith stringdemodeploy; accept defaults. Two folders appear.Payoff
Expand
stringdemo/: point at stringdemo-api and stringdemo-service. “There it is — the exact split from last episode, generated for you.”Demo — build
cd output/stringdemo && make build, then cd ../stringdemodeploy && make build. Narrate the api and service libraries building, then the deployable.Demo — run & call
Start the monolith; hit the generated endpoint with a curl script from
scripts/. Show the response, then the health endpoint. “Business logic, a health probe, and an HTTP binding — and we wrote none of the plumbing.”Point
Voiceover: “Look at the impl again — still just business logic. Everything else came from the framework. Next, we start adding the horizontal behavior as policies.”
Recap card
“Generated an
api + service pair and a deployable monolith, built and called it — no plumbing written by hand.”Companion reading: Get started · Definition vs. implementation