Lectures

We break down building in Go using go build: how Go builds not “files” but packages and dependencies, what the ./... target really means, when a binary appears (only for package main + func main ()), how to control the output name/folder with the -o flag, and why a build often “creates nothing” because of the build cache.

We walk through a practical way to “bake” build metadata directly into a Go binary at link time: pass version/commit/date via -ldflags and the -X option, store the values in the internal/buildinfo package (must be var, not const), and add a --version flag with an early exit and no side effects. At the end: common mistakes—import path from go.mod, quotes/spaces, and output to stdout/ stderr.

We break down the basic cross-compilation model in Go: how host differs from target, how to choose a build platform via the environment variables GOOS/ GOARCH, which combinations are most common, how to set them for go build, and how to “bake” the target into --version output via runtime. At the end — typical mistakes and ways to quickly diagnose build artifacts.

We break down how to package a CLI application into a “release bundle”: a predictable folder with a binary, README, an examples/ directory, and release notes ( RELEASE_NOTES.md/ CHANGELOG.md). We’ll discuss the distribution contract, artifact naming (the app_version_GOOS_GOARCH format), required diagnostic flags --help and --version, and common mistakes when preparing releases.

We break down practical build reproducibility in Go: which inputs affect the bytes of the resulting binary (sources, dependencies from go.mod/ go.sum, the toolchain version, GOOS/ GOARCH, -ldflags), how to embed a semantic fingerprint ( fingerprint) into the binary via --version, and how to verify file identity using sha256 (checksum).