test: add Send/Earn coverage with CI enforcement

[Ayash-Bera]

Summary

• Add Jest tests for Send helpers, SendForm component, Send index, Earn component, FeeConfigModal, utxo utils, and feature flags (165 tests across 22 suites)
• Add coverageThreshold to package.json (statements 38, branches 28, functions 30, lines 39)
• Update CI workflow to run npm test -- --coverage --watchAll=false and fail on threshold breach
• Add TESTING.md documenting provider wrapping, API mocking, and WebSocket test patterns

Coverage Baseline (pre-PR)

Before this PR, the only tests were App.test.tsx and Wallets.test.tsx (2 suites, 17 tests). The seven target files were absent from coverage output entirely — none are imported by those two test files.

Running CI=true npm test -- --coverage --watchAll=false on the upstream master produced:

Metric	Before	After
Statements	20%	38%
Branches	7%	28%
Functions	11%	30%
Lines	21%	39%

Files absent from coverage before this PR:

• src/components/Send/helpers.ts
• src/constants/features.ts
• src/components/utxo/utils.ts
• src/components/Send/SendForm.tsx
• src/components/Send/index.tsx
• src/components/Earn.tsx
• src/components/settings/FeeConfigModal.tsx

Coverage Threshold Justification

Thresholds are set at floor(measured_after) — the minimum that the new suite reliably meets. This keeps CI honest without inflating numbers: any regression that drops below the measured baseline fails CI, and the bar is not set higher than what the actual tests deliver.

Test Plan

• CI=true npm test -- --coverage --watchAll=false passes all 165 tests
• Coverage thresholds met (statements 38, branches 28, functions 30, lines 39)
• CI workflow enforces coverage on every push/PR

[Ayash-Bera]

Post-review update: reviewed PR #1222 (a parallel submission for the same project). A few things stood out that informed improvements here:

• Their per-folder coverage threshold proposal aligned with what I had in mind — I extended our thresholds to also cover Earn.tsx and settings/ paths (not just Send/), so regressions in those critical paths now fail CI independently of the global gate.
• Their test depth on validation logic was a useful reference point — pushed me to move our component tests from render/loading checks to actual behavioral assertions: postMakerStart call contract, postDirectSend API branching, FeeConfigModal save success/failure paths, and SendForm Formik validation blocking.
• Fixed a TypeScript build break in the process (SendForm.test.tsx wallet fixture missing displayName).

Coverage is now 46% statements / 39% branches (up from 38% / 28%), with 178 tests across 22 suites.

[Ayash-Bera]

Follow-up update after the latest test hardening commits:

I addressed the remaining core-flow gaps from my previous note and pushed additional behavioral coverage for both Send and Earn lifecycles.

What changed since last update

• Send flow

  • Added explicit coinjoin branch coverage in Send/index.test.tsx (postCoinjoin assertion path)
  • Added direct-send rejection coverage with UI error/state recovery assertions
  • Tightened selectors and removed brittle spots in send integration tests

• Earn flow

  • Added maker-start failure coverage (postMakerStart reject path)
  • Added stop-maker lifecycle coverage (getMakerStop assertion path)

Current status (latest local run)

• npm run lint ✅
• npm run build ✅
• CI=true npm test -- --coverage --watchAll=false ✅
• 22/22 suites, 182/182 tests
• Coverage now at 47.13% statements / 40.36% branches / 38.88% functions / 48.30% lines

These updates are intended to fully align with the project objective: deep coverage of Send/Earn/Fee-config critical paths plus enforced CI coverage gates to prevent silent regressions.