tighten helm tests

[atkrad]

Summary

• Cover Helm list, get, delete, history, and rollback: invalid input, a missing release, an unreachable cluster, and whether delete waits for resources.
• Lock client timeout, debug chart cleanup, install dry-run, a canceled install, and chart-value edges (image digest, args, TLS off, invalid probe durations).
• Make chart-cache expiry deterministic by backdating the entry instead of waiting on the clock.

Testing principles addressed:

• Failure paths for invalid input, missing releases, an unreachable cluster, cancellation, and unsupported history.
• ErrorIs for sentinels, and ErrorContains only when the message is the contract.
• Determinism, per-test isolation, cleanup of chart directories, and t.Parallel except where a test calls t.Setenv.

Related

Test plan

• Unit tests added/updated
• Scenario under scenarios/ (if behavior changes)
• nix run .#lint / pre-commit clean
• Manual smoke (command + expected result), if user-facing

go test -race -shuffle=on -count=1 ./internal/helm/ passed.
golangci-lint run ./internal/helm/ reported 0 issues.
Pre-commit hooks passed on the commit.

Remaining

ListReleases, GetRelease, DeleteRelease, GetReleaseHistory, and RollbackRelease accept a context and ignore it. Cancellation of those calls stays untested until the methods honor ctx. That needs a production change, so it is left out of this PR.

Labels

• One of: kind/feature, kind/bug, kind/docs, kind/chore
• Add breaking-change if this breaks existing CLI or config behavior
• Add skip-changelog for internal-only PRs that should not appear in notes

Checklist

• Title is short and imperative (matches commit style)
• Docs / CLI help updated when user-facing (README.md, docs/cli/)
• No secrets or local-only paths in the diff

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.

📢 Thoughts on this report? Let us know!