Every KERN-01 through KERN-09 and CLI-01 contract has a behavioral test, including failure and non-TTY paths.
The final check runs `go vet`, `go test`, and `go test -race` in the framework, hello app and each hello plugin module.
A temporary app can run `summer build` and its generated binary; source edit under `summer dev` records measured rebuild time and restarts the child.
path
provides
scripts/check-phase1.sh
Repeatable phase-wide verification command
path
provides
compass/config_test.go
Config precedence regression tests
path
provides
festival/bus_test.go
Event dispatch regression tests
path
provides
examples/hello/hello_test.go
Built-app integration test
from
to
via
scripts/check-phase1.sh
examples/hello/go.mod
Checks nested app and plugin modules independently
from
to
via
examples/hello/hello_test.go
cmd/summer/main.go
Builds and executes a generated app binary
As a framework maintainer, I can run one repeatable check that proves the Phase 1 kernel and development loop before Phase 3 depends on them.
Purpose: Add the phase's dedicated unit and integration test plan, with broad behavior and failure coverage.
Output: Focused Go tests, a nested-module check script, a generated-binary smoke and completed Nyquist validation map.
@.planning/PROJECT.md
@.planning/ROADMAP.md
@.planning/REQUIREMENTS.md
@.planning/phases/01-framework-kernel-foundation/01-CONTEXT.md
@.planning/phases/01-framework-kernel-foundation/01-RESEARCH.md
@.planning/phases/01-framework-kernel-foundation/01-VALIDATION.md
@.planning/phases/01-framework-kernel-foundation/01-03-SUMMARY.md
Task 1: Cover config, lifecycle, services, events and context behavior
compass/config_test.go, party/registry_test.go, backpack/services_test.go, festival/bus_test.go, towel/context_test.go
compass/config.go, compass/env.go, compass/persist.go, party/registry.go, backpack/app.go, backpack/services.go, festival/bus.go, towel/context.go, compass/config_test.go, party/registry_test.go, backpack/services_test.go, festival/bus_test.go, towel/context_test.go, .planning/phases/01-framework-kernel-foundation/01-CONTEXT.md
Expand earlier smoke tests into table-driven behavioral tests for all config precedence levels, malformed YAML, snake_case env keys, `.env` non-override, dotted plugin namespace, typed section decode, Set/Persist/Reload and concurrent reads. Test plugin duplicate/missing/cycle errors, reordered manifest inputs, all-Register-before-any-Boot, HasConfig/HasCommands assertions, absent/present HasPlugin and typed service lookup with two separate App instances. Test three typed event dispatch modes, priority and stable ties, partial collect payload, joined errors, stop-on-error, owner-labelled panic recovery and concurrent independent app buses. Test actor/organization/collection/locale context accessors and nested context isolation. Fix any production defect exposed by these tests in the same task, keeping API behavior from D-01–D-13 intact. Report per-package coverage for these packages; cover contract branches rather than chasing an arbitrary percentage.
go test -race ./compass ./party ./backpack ./festival ./towel ./pact
- Each KERN-01/02/03/05/06/07/08 behavior above has at least one source-to-observable assertion.
- `go test -race ./compass ./party ./backpack ./festival ./towel ./pact` exits 0.
- No test asserts only implementation structure when a behavior can be observed.
The kernel's core behavioral and failure contracts are regression tested.
Task 2: Cover tool, output, watch loop and every workspace module
bonfire/output_test.go, bonfire/prompts_test.go, internal/build/build_test.go, internal/dev/watch_test.go, examples/hello/hello_test.go, scripts/check-phase1.sh, .planning/phases/01-framework-kernel-foundation/01-VALIDATION.md
bonfire/root.go, bonfire/output.go, bonfire/widgets.go, bonfire/prompts.go, internal/build/build.go, internal/build/manifest.go, internal/build/scaffold.go, internal/dev/watch.go, cmd/summer/main.go, examples/hello/summer.yaml, bonfire/output_test.go, bonfire/prompts_test.go, internal/build/build_test.go, internal/dev/watch_test.go, examples/hello/hello_test.go, scripts/check-phase1.sh, .planning/phases/01-framework-kernel-foundation/01-VALIDATION.md
Test namespaced command discovery and flag/argument parsing, injected output capture, all non-TTY widget/prompt fallbacks, `NO_COLOR`/`FORCE_COLOR`/`TERM=dumb` policy and closed stdin. In a temporary copied hello workspace, assert `make:plugin` minimal files, `plugin:add` idempotence, stable generated bytes, malicious ID/path rejection, `summer build` success, built hello command output and non-zero failure on invalid Requires. Use deterministic build/process hooks for debounce/cancellation tests, plus a real fsnotify temp-workspace edit for one successful rebuild/restart; capture and assert a `rebuild: ` line without assuming a threshold. Add `scripts/check-phase1.sh` that runs `go vet ./...`, `go test ./...` and `go test -race ./...` from root, `examples/hello`, and each nested plugin module; include built-binary integration. Update VALIDATION.md's per-task map/status and set `nyquist_compliant: true` only when every planned automated check exists and passes. Fix production defects the tests expose, while keeping this plan focused on verification.
bash scripts/check-phase1.sh
- `bash scripts/check-phase1.sh` exits 0 and names root, hello app, base, greeter and optional plugin modules.
- Test suite observes `summer build` and a separately executed hello binary, not only package compilation.
- Watch test observes a source-change rebuild, restart and measured latency line; generated-file edits do not loop.
- Non-TTY spinner/progress/table/prompt and color policy assertions run without a terminal.
- VALIDATION.md records passing automated checks and `nyquist_compliant: true` only after the full script passes.
The phase has a repeatable, race-enabled verification command and completed validation evidence.
<threat_model>
Trust Boundaries
Boundary
Description
Test fixture input → tool/process
Malformed manifests and source edits should fail safely during local tooling.
STRIDE Threat Register
Threat ID
Category
Component
Disposition
Mitigation Plan
T-01-09
Tampering
Manifest and scaffold inputs
mitigate
Add negative tests for invalid IDs, duplicate modules and path traversal.
T-01-10
Denial of service
Watch loop
mitigate
Test debounce, ignored outputs, failed build and child cleanup.
T-01-11
Information disclosure
CLI secret and config
mitigate
Assert secret prompt answer and config values do not leak into captured error/output streams.
</threat_model>
- Run `bash scripts/check-phase1.sh` and inspect its per-module output.
- Confirm `go test -race` passes in all five modules and `go vet` is green.
- Record measured single-plugin rebuild time in 01-04-SUMMARY.md.
<success_criteria>
All ten phase requirement IDs have automated behavioral evidence.
The final testing plan runs after implementation and passes across all workspace modules.
VALIDATION.md can be signed off without missing tests or broken Wave 0 references.
</success_criteria>
Create `.planning/phases/01-framework-kernel-foundation/01-04-SUMMARY.md` after completion.
