Charon/docs/plans/frontend_coverage_boost.md

Frontend Coverage Boost Plan (>=85%)

Current (QA): statements 84.54%, branches 75.85%, functions 78.97%. Goal: reach >=85% with the smallest number of high-yield tests.

Targeted Tests (minimal set with maximum lift)

• API units (fast, high gap)

  • src/api/notifications.ts: cover payload branches in previewProvider (with/without data) and previewExternalTemplate (id vs inline template vs both), plus happy-path CRUD wrappers to verify endpoint URLs.
  • src/api/logs.ts: assert getLogContent query param building (search/host/status/level/sort), downloadLog sets window.location.href, and connectLiveLogs callbacks for onOpen, onMessage (valid JSON), parse error branch, onError, and onClose (closing when readyState OPEN/CONNECTING).
  • src/api/users.ts: cover invite, permissions update, validate/accept invite paths; assert returned shapes and URL composition (e.g., /users/${id}/permissions).

• Component tests (few, branch-heavy)

  • src/pages/SMTPSettings.tsx: component test with React Testing Library (RTL).
    • Ensure initial render waits for query then hydrates host/port/encryption (flaky area); verify loading spinner disappears.
    • Save success vs error toast branches; Test Connection success/error; Send Test Email success clears input and error path shows toast.
    • Button disables: test connection disabled when host or fromAddress empty; send test disabled when testEmail empty.
  • src/components/LiveLogViewer.tsx: component test with mocked WebSocket and connectLiveLogs.
    • Verify pause/resume toggles, log trimming to maxLogs, filter by text/level, parse-error branch (bad JSON), and disconnect cleanup invokes returned close fn.
  • src/pages/UsersPage.tsx: component test.
    • Invite modal success when email_sent false shows manual link copy branch; toggle permission mode text for allow_all vs deny_all; checkbox host toggle logic.
    • Permissions modal seeds state from selected user and saves via updateUserPermissions mutation.
    • Delete confirm branch (stub confirm), enabled Switch disabled for admins, enabled toggles for non-admin users.

• Security & CrowdSec flows

  • src/pages/CrowdSecConfig.tsx: component test (can mock queries/mutations).
    • Cover hub unavailable (503) -> preset-hub-unavailable, cached preview fallback via getCrowdsecPresetCache, validation error (400) -> preset-validation-error, and apply fallback when backend returns 501 to hit local apply path and preset-apply-info rendering.
    • Import flow with file set + disabled state; mode toggle (crowdsec-mode-toggle) updates via updateSetting; ensure decisions table renders "No banned IPs" vs list.
  • src/pages/Security.tsx: component test.
    • Banner when cerberus.enabled is false; toggles toggle-crowdsec/toggle-acl/toggle-waf/toggle-rate-limit call mutations and optimistic cache rollback on error.
    • LiveLogViewer renders only when Cerberus enabled; whitelist input saves via useUpdateSecurityConfig and break-glass button triggers mutation.

• Shell/UI overview

  • src/pages/Dashboard.tsx: component test to cover health states (ok, error, undefined) and counts computed from hooks.
  • src/components/Layout.tsx: component test.
    • Feature-flag filtering (hide Uptime/Cerberus when flags false), sidebar collapse persistence (localStorage), mobile toggle (data-testid="mobile-menu-toggle"), nested menu expand/collapse, logout button click, and version/git commit rendering.

• Missing/low names from QA list

  • Summary.tsx, FeatureFlagProvider.tsx, useFeatureFlags.ts, LiveLogViewerRow.tsx: confirm current paths (may have been renamed). Add light RTL/unit tests mirroring above patterns if still present (e.g., summary widget rendering counts, provider supplying default flags).

SMTPSettings Deflake Strategy

• Wait for data: use await screen.findByText('Email (SMTP) Settings') and await waitFor(() => expect(hostInput).toHaveValue('...')) after mocking getSMTPConfig to resolve once.
• Avoid racing mutations: wrap vi.useFakeTimers() only if timers are used; otherwise keep real timers and await act(async () => ...) on mutations.
• Reset query cache per test (queryClient.clear() or QueryClientProvider fresh instance) and isolate toast spies.
• Prefer role/label queries (getByLabelText('SMTP Host')) over brittle text selectors; ensure toast mocks are flushed before assertions.

Ordered Phases (minimal steps to >=85%)

• Phase 1 (API unit bursts) — expected +0.30 to statements: notifications.ts, logs.ts, users.ts.
• Phase 2 (UI quick wins) — expected +0.50: SMTPSettings, LiveLogViewer, UsersPage.
• Phase 3 (Security shell) — expected +0.40: CrowdSecConfig, Security page.
• Phase 4 (Shell polish) — expected +0.20: Dashboard, Layout, any remaining Summary/feature-flag provider files if present.

Total projected lift: ~+1.4% (buffered) with 8–10 focused tests. Stop after Phase 3 if coverage already surpasses 85%; Phase 4 only if buffer needed.