Golang Pitfalls: Testing & Benchmarks

Source material: mistakes #82-90 (plus community fuzzing) from 100 Go Mistakes and How to Avoid Them (teivah/100-go-mistakes).

Apply these rules when writing Go tests and benchmarks.

82. Not categorizing tests (#82)

Categorize with build tags, environment variables, or short mode to separate unit vs. integration vs. long-running tests.
Example: gate slow tests behind -short (if testing.Short() { t.Skip(...) }) or separate files with //go:build integration.

Run tests with the race detector: go test -race ./....
It instruments memory accesses at runtime (not static analysis) and detects data races in concurrent tests, with added memory/time overhead — enable it locally and in CI, not production.
Reading a WARNING: DATA RACE report: it lists the concurrent goroutines, the offending read/write locations in code, and where the goroutine was created.
You can exclude a specific test file from race detection with //go:build !race (temporarily).