write-tests

Write Tests

Write tests that match project conventions and cover real edge cases.

Workflow

Discover → Understand → Design → Write → Validate

1. Discover

Find the testing landscape:

tl parallel \
  "testmap=tl test-map <file>" \
  "coverage=tl coverage <file>" \
  "example=tl example '*.test.*'" \
  "style=tl style"

2. Understand

Know the code before testing it:

tl parallel \
  "symbols=tl symbols <file>" \
  "exports=tl exports <file>" \
  "deps=tl deps <file>"
tl snippet <function> <file>   # Read each function to understand branches and edge cases

3. Design

For each exported function, identify:

• Happy path (normal input, expected output)
• Error cases (invalid input, missing data, exceptions)
• Edge cases (empty input, boundary values, null/undefined)

tl parallel "flow=tl flow <function> <file>" "types=tl types <file>"

4. Write

• Follow the pattern from existing test files found in step 1
• Match the project's test runner, assertion style, and file naming
• Colocate test files with source (or match existing convention)
• Mock external dependencies identified by tl deps, not internal functions

5. Validate

tl parallel \
  "new=tl run '<test command> <new-test-file>'" \
  "coverage=tl coverage <file>" \
  "full=tl run '<full test suite>'"

Decision tree

Write tests → Do tests already exist for this file?
  ├─ Yes (extend) → tl test-map to find the test file
  │   → tl symbols on the test file to see what's covered
  │   → tl coverage to find gaps
  │   → Add missing cases following existing patterns
  ├─ No (create) → tl example "*.test.*" to learn project patterns
  │   → tl exports on source to prioritize public API
  │   → tl deps to identify what to mock
  │   → Create test file matching project conventions
  └─ Coverage target → tl coverage for current numbers
      → tl symbols to list all functions
      → Prioritize: exported > complex > simple
      → Write tests until target met