[appleboy]

Summary

Apply idioms from the Modern Go Guidelines for this module's Go version (go 1.26.10). Detection: go fix -diff ./... on every module plus a manual scan for patterns go fix does not rewrite. No library/production code changes. Only tests, benchmarks, and _example programs are touched.

• testing_b_loop: replace b.N loops (and the now-redundant b.ResetTimer()) with for b.Loop() in job/benchmark_test.go, BenchmarkQueueTask, BenchmarkQueue, BenchmarkRingWithTask, the resize operations sub-benchmark, BenchmarkHandleRetry, and BenchmarkRingBoundary.
• atomic_types: thread_test.go uses atomic.Int32 instead of atomic.AddInt32 / LoadInt32. The failure message also reads the counter through Load() instead of reading the raw value.
• range_over_int (via go fix): _example/example01 and example02 use for i := range taskN.
• loopvar_capture: drop the redundant go func(i int) { ... }(i) parameter in the examples, because each iteration has its own i since Go 1.22.
• Intentionally not changed:
  • BenchmarkNewRing / testQueue and benchmarkQueueCompletion depend on b.N for queue sizing and batching, so they keep the b.N loop.
  • go fix also proposed adding //go:fix inline to job.Int64/Float64/Time/Bool. Those helpers are explicitly marked //nolint:modernize // pointer-returning helper kept intentionally, so that change was skipped.

Related issues

• Jira: N/A
• GitHub: N/A

AI authorship

• No AI was used
• AI was used
  • Tool / model: Claude Code (Claude Opus 5.5)
  • AI-authored files: _example/example01/main.go, _example/example02/main.go, benchmark_test.go, handle_performance_test.go, job/benchmark_test.go, ring_performance_test.go, ring_test.go, thread_test.go
  • Human line-by-line reviewed: None — not yet reviewed by a human.

Change classification

• Leaf change: tests, benchmarks, and examples only; no exported API or runtime behavior changes
• Core change

Plan reference

Goal: run the Modern Go Guidelines (list --go-version 1.26) across every Go module in the golang-queue org workspace and apply the guidelines that fit without changing behavior. Companion PRs were opened in the sibling repos (queue, benchmark, nats, nsq, rabbitmq, redisdb, redisdb-stream).

Verification

Setup

• Checkout target / working directory: branch refactor/modern-go-idioms @ fde8f0b, repository root
• Prerequisites: Go 1.26.10 (go version)
• Prepare and start: go mod download

Automated checks

Command (with working directory)	Behavior covered / expected success	Status	Observed result / reason not run
go vet ./...	Static analysis	Passed	No output
make lint	golangci-lint (repo config)	Passed	0 issues.
go test -race ./...	Unit tests incl. goleak	Passed	ok github.com/golang-queue/queue, ok github.com/golang-queue/queue/job
go test -run '^$' -bench . -benchtime 200x ./...	All benchmarks still execute	Passed	ok for both packages
cd _example/example01 && go vet ./... (and example02)	Examples compile	Passed	No output

Behavioral scenarios

Tests and benchmarks unchanged in behavior

• Acceptance condition: no library code changes; tests pass under -race; every converted benchmark still runs.
• Starting state: Go 1.26.10, repository root.

Step	Action	Expected observable result
1	go test -race -run TestRoutineGroupRun -v .	Both subtests PASS with no race report
2	go test -run '^$' -bench . -benchtime 200x ./...	All benchmarks print results; ok for queue and queue/job
3	cd _example/example01 && go run .	Prints 30 message: Hi Gopher, handle the job: NN lines with distinct NN

• Execution status: Steps 1–2 Passed; step 3 Not run (examples verified by go vet only)
• Observed result: see Automated checks.
• Cleanup: N/A — read-only.

Evidence

Web screenshots

N/A — no UI.

Performance report

N/A — benchmark harness modernization only; no performance claim is made. Absolute numbers from b.Loop() may differ slightly from the old b.N loop because setup is now excluded automatically.

Security check

• No secrets in the diff
• N/A - no external or security-sensitive interface changed

Risk and rollback

• Risk: Low. Only tests, benchmarks, and examples change.
• Rollback: revert this single commit.

Reviewer guide

• Read carefully: thread_test.go (atomic type change), benchmark_test.go (confirm b.N-sized benchmarks were left alone).
• Spot-check: mechanical for i := range n / go func() rewrites in _example/.
• Evidence not captured: N/A

🤖 Generated with Claude Code