Thalarch Go

Thalarch Go

Use the Go version and module/tooling declared by the repository. Keep code simple and explicit; do not import patterns from Java/C++ when Go's standard library and conventions already solve the problem.

Preflight

Inspect go.mod, workspace files, generated-code markers, build tags, configured linters, Make/Task scripts, tests, and CI commands.

Idiomatic Go

• Keep interfaces small and consumer-owned when abstraction is genuinely needed.
• Prefer concrete types until multiple implementations or test seams justify an interface.
• Return useful errors rather than sentinel booleans that discard context.
• Wrap errors when the added context helps diagnosis while preserving errors.Is/As semantics.
• Use defer for clear ownership cleanup when its lifetime/cost is appropriate.
• Preserve zero-value usability when the existing API depends on it.

Avoid unnecessary getters, inheritance-shaped abstractions, and package-level mutable state.

Concurrency

Every goroutine needs an owner and termination path.

Review:

• context propagation/cancellation;
• goroutine leaks;
• channel close ownership;
• blocked sends/receives;
• data races;
• mutex scope and lock ordering;
• unbounded fan-out;
• timer/ticker cleanup;
• backpressure and worker limits.

Do not create a goroutine merely to make a synchronous call look concurrent.

HTTP/services

When building services, preserve:

• request context;
• timeout/cancellation behavior;
• body/resource cleanup;
• stable error/status semantics;
• bounded concurrency and connection reuse;
• structured logging/observability conventions already used by the repo.

Combine with thalarch-api, thalarch-security, and thalarch-data-sql when relevant.

Testing

Use table-driven tests when they make a behavior matrix easier to see, not as mandatory style. Use subtests, fuzz tests, race checks, integration tests, and benchmarks when they prove a real risk or regression.

Useful commands may include go test, go test -race, go vet, configured linters, and benchmarks — but run repository-native scripts/flags first.

Performance

Measure with available tools such as benchmarks, pprof, execution traces, allocation profiles, and application metrics. Fix algorithmic/I/O/concurrency bottlenecks before micro-tuning.

Verification

Run formatting plus the project's targeted tests/static checks/build. For concurrency-sensitive changes, include race detection when feasible and relevant.