Phase 06 Context

Goal

Not available

Codebase Map Available

Codebase mapping exists in .vbw-planning/codebase/. Key files:

• ARCHITECTURE.md
• CONCERNS.md
• PATTERNS.md
• DEPENDENCIES.md
• STRUCTURE.md
• CONVENTIONS.md
• TESTING.md
• STACK.md

Read CONVENTIONS.md, PATTERNS.md, STRUCTURE.md, and DEPENDENCIES.md first to bootstrap codebase understanding.

Changed Files (Delta)

• .vbw-planning/config.json
• .vbw-planning/milestones/03-coverage-analysis-quick-wins-critical-path-test-co/STATE.md
• .vbw-planning/phases/02-feed-reliability/01-SUMMARY.md
• .vbw-planning/phases/02-feed-reliability/02-SUMMARY.md
• .vbw-planning/phases/02-feed-reliability/03-SUMMARY.md
• .vbw-planning/phases/02-feed-reliability/04-SUMMARY.md
• .vbw-planning/phases/03-dashboard-pagination/03-04-SUMMARY.md
• .vbw-planning/phases/05-simplify-source-status/01-SUMMARY.md
• .vbw-planning/phases/05-simplify-source-status/03-SUMMARY.md
• .vbw-planning/phases/05-simplify-source-status/05-SUMMARY.md
• .vbw-planning/ROADMAP.md
• test/dummy/db/schema.rb
• test/dummy/Gemfile.lock

Code Slices

.vbw-planning/config.json (51 lines, first 30 shown)

{
  "effort": "thorough",
  "autonomy": "standard",
  "auto_commit": true,
  "planning_tracking": "manual",
  "auto_push": "never",
  "verification_tier": "standard",
  "skill_suggestions": true,
  "auto_install_skills": false,
  "discovery_questions": true,
  "context_compiler": true,
  "visual_format": "unicode",
  "max_tasks_per_plan": 5,
  "prefer_teams": "auto",
  "branch_per_milestone": false,
  "plain_summary": true,
  "active_profile": "default",
  "custom_profiles": {},
  "model_profile": "quality",
  "model_overrides": {},
  "agent_max_turns": {
    "scout": 15,
    "qa": 25,
    "architect": 30,
    "debugger": 80,
    "lead": 50,
    "dev": 75
  },
  "qa_skip_agents": [
    "docs"

.vbw-planning/milestones/03-coverage-analysis-quick-wins-critical-path-test-co/STATE.md (64 lines, first 30 shown)

<!-- VBW STATE TEMPLATE (ARTF-05) -- Session dashboard, auto-updated -->
<!-- Updated after each plan completion and at checkpoints -->

# Project State

## Project Reference

See: .vbw-planning/PROJECT.md (updated 2026-02-09)

**Core value:** Drop-in Rails engine for feed monitoring, content scraping, and operational dashboards.
**Current focus:** All phases complete

## Current Position

Phase: 4 of 4 (Code Quality & Conventions Cleanup)
Plan: 3 of 3 in current phase
Status: Built
Last activity: 2026-02-10 -- Phase 4 complete (3/3 plans done)

Progress: [##########] 100%

## Codebase Profile

- **Total source files:** 535
- **Primary language:** Ruby (330 files)
- **Templates:** ERB (48 files)
- **Tests:** 137 test files detected
- **Test suite:** 841 runs, 2776 assertions, 0 failures (up from 473 runs in Phase 1)
- **Coverage:** 86.97% line, 58.84% branch (510 uncovered lines, down from 2117)
- **CI/CD:** GitHub Actions (1 workflow)

.vbw-planning/ROADMAP.md (131 lines, first 30 shown)

# Roadmap

**Milestone:** rails-audit-and-modal

## Phases

- [x] Phase 01: Quick Wins & Security
- [x] Phase 02: Model Layer Hardening
- [x] Phase 03: Controller & Route Refactoring
- [x] Phase 04: Job & Pipeline Reliability
- [x] Phase 05: View Layer Extraction
- [ ] Phase 06: Test Infrastructure
- [ ] Phase 07: Ultimate Turbo Modal Integration

## Phase Details

### Phase 01: Quick Wins & Security

**Goal:** Fix high-priority security issues and low-effort quick wins from the Rails audit (findings M1, M3, M5, M12, C6, C9, V2, V8, V10, V11).

**Success Criteria:**
- ImportHistoryDismissalsController scoped to current user
- DashboardController uses explicit parameter allowlist (no .permit!)
- ScrapeLog has by_source, by_status, by_item scopes
- Scrape status badge logic extracted to helper
- MutationObserver disconnect leak fixed

**Plans:**
- [x] Plan 01: Security & Controller Fixes
- [x] Plan 02: Model & Job Fixes

test/dummy/db/schema.rb (449 lines, first 30 shown)

# This file is auto-generated from the current state of the database. Instead
# of editing this file, please use the migrations feature of Active Record to
# incrementally modify your database, and then regenerate this schema definition.
#
# This file is the source Rails uses to define your schema when running `bin/rails
# db:schema:load`. When creating a new database, `bin/rails db:schema:load` tends to
# be faster and is potentially less error prone than running all of your
# migrations from scratch. Old migrations may fail to apply correctly if those
# migrations use external dependencies or application code.
#
# It's strongly recommended that you check this file into your version control system.

ActiveRecord::Schema[8.1].define(version: 2026_03_13_120000) do
  # These are extensions that must be enabled in order to support this database
  enable_extension "pg_catalog.plpgsql"

  create_table "active_storage_attachments", force: :cascade do |t|
    t.bigint "blob_id", null: false
    t.datetime "created_at", null: false
    t.string "name", null: false
    t.bigint "record_id", null: false
    t.string "record_type", null: false
    t.index ["blob_id"], name: "index_active_storage_attachments_on_blob_id"
    t.index ["record_type", "record_id", "name", "blob_id"], name: "index_active_storage_attachments_uniqueness", unique: true
  end

  create_table "active_storage_blobs", force: :cascade do |t|
    t.bigint "byte_size", null: false
    t.string "checksum"
    t.string "content_type"

test/dummy/Gemfile.lock (409 lines, first 30 shown)

PATH
  remote: ../..
  specs:
    source_monitor (0.11.1)
      cssbundling-rails (~> 1.4)
      faraday (~> 2.9)
      faraday-follow_redirects (~> 0.4)
      faraday-gzip (~> 3.0)
      faraday-retry (~> 2.2)
      feedjira (>= 3.2, < 5.0)
      jsbundling-rails (~> 1.3)
      nokolexbor (~> 0.5)
      rails (>= 8.0.3, < 10.0)
      ransack (~> 4.2)
      ruby-readability (~> 0.7)
      solid_cable (>= 3.0, < 4.0)
      solid_queue (>= 0.3, < 3.0)
      turbo-rails (~> 2.0)

GEM
  remote: https://rubygems.org/
  specs:
    action_text-trix (2.1.16)
      railties
    actioncable (8.1.2)
      actionpack (= 8.1.2)
      activesupport (= 8.1.2)
      nio4r (~> 2.0)
      websocket-driver (>= 0.6.1)
      zeitwerk (~> 2.6)

Research Findings

phase: 06 title: Test Infrastructure Research - Phase 06 Audit Findings (T1-T17) type: research confidence: high date: 2026-03-14

Phase 06: Test Infrastructure Research

Current State

Test Infrastructure Overview

Framework & Setup:

• Minitest (no RSpec or FactoryBot)
• WebMock + VCR for HTTP stubbing
• SimpleCov for coverage tracking
• TestProf for optimized test runs (before_all, sampling)
• Thread-based parallelism (not process-based due to PG segfault)
• Rails 8 dummy app at test/dummy/ with Solid Queue and Hotwire

Key Files:

• test/test_helper.rb - Main test setup (65-128 lines)
• test/test_prof.rb - TestProf integration (38 lines)
• test/support/host_app_harness.rb - Host app test generation (395 lines)
• VCR cassettes: test/vcr_cassettes/source_monitor/ (4 cassettes found)

Test Count by Directory (estimated):

• Controllers: 6 test files
• Jobs: 3 test files
• Lib/: 50+ test files (architecture, configuration, fetching, items, jobs, logs, models, scrapers, scraping, setup, realtime, health, events, etc.)
• Models: 2 test files
• System: 6 test files
• Integration: 3 test files
• Tasks: 2 test files
• Mailers: 1 test file
• Helpers: 1 test file
• Examples: 4 test files (template tests)
• Total: ~80 test files

Test Count by Type:

• 1214 tests total (from project memory)
• Thread-parallel workers: N (number of processors) or PARALLEL_WORKERS env var
• Test order: random
• Per-test setup: SourceMonitor.reset_configuration! + SourceMonitor.config.http.retry_max = 0

Audit Findings Detail

T1: VCR Cassette Maintenance Not Documented

Finding: VCR is configured but usage patterns are not standardized across tests.

Current State:

• test/test_helper.rb (lines 59-63):
  • Cassette dir: test/vcr_cassettes/
  • Hooked into WebMock
  • ignore_localhost = true
• Only 4 cassettes found:
  • source_monitor/fetching/rss_success.yml
  • source_monitor/fetching/atom_success.yml
  • source_monitor/fetching/json_success.yml
  • source_monitor/fetching/netflix_medium_rss.yml
• Zero VCR.use_cassette calls found in test files (no grep matches)

Gap: Cassette strategy is not documented:

• When to record vs use cassettes
• How to regenerate stale cassettes
• Naming convention (only partial pattern exists: source_monitor/<module>/<descriptor>)
• Maintenance schedule for outdated recordings

Files Affected: test/test_helper.rb, test/vcr_cassettes/

T2: Fixture-Based Tests Tightly Coupled to User Model

Finding: Test helper fixtures :all loads all fixtures, creating implicit dependencies.

Current State:

• test/test_helper.rb (lines 51-57):
  rubyCopy

  if ActiveSupport::TestCase.respond_to?(:fixture_paths=)
    fixtures_root = File.expand_path("fixtures", __dir__)
    ActiveSupport::TestCase.fixture_paths = [ fixtures_root ]
    ActionDispatch::IntegrationTest.fixture_paths = ActiveSupport::TestCase.fixture_paths
    ActiveSupport::TestCase.file_fixture_path = fixtures_root
    ActiveSupport::TestCase.fixtures :all  # <-- loads ALL fixtures
  end
  

• fixtures :all loads every fixture, including users.yml if it exists
• No explicit fixtures :users found in grep (search returned no matches)
• But fixtures :all implicitly couples every test to user model if fixtures exist

Gap:

• Implicit coupling obscures test dependencies
• Tests that don't need fixtures still load them (performance hit)
• No isolated unit tests can avoid fixture loading without refactoring

Files Affected: test/test_helper.rb, test/fixtures/ (directory)

T3: No Isolated Unit Tests for Extracted Sub-Modules

Finding: Refactored sub-modules lack dedicated unit test files.

Current State:

• Phase 3/4 extracted FeedFetcher into sub-modules:
  • AdaptiveInterval - no dedicated test file found
  • SourceUpdater - no dedicated test file found
  • EntryProcessor - test file exists: test/lib/source_monitor/fetching/feed_fetcher/entry_processor_test.rb
• EntryProcessor has 1 test file
• Other sub-modules tested only as part of larger integration tests

Gap:

• Unit tests for AdaptiveInterval and SourceUpdater behavior don't exist
• Forces reliance on integration tests that exercise multiple layers
• Harder to debug failures (wider blast radius)
• Difficult to test edge cases in isolation

Files Affected:

• lib/source_monitor/fetching/feed_fetcher/adaptive_interval.rb (no test file)
• lib/source_monitor/fetching/feed_fetcher/source_updater.rb (no test file)
• lib/source_monitor/fetching/feed_fetcher/entry_processor.rb (test exists)

T4: Integration Test Coverage Gaps

Finding: Integration tests exist but coverage is incomplete for complex flows.

Current State:

• Integration tests found: 3 files
  • test/integration/engine_mounting_test.rb
  • test/integration/host_install_flow_test.rb
  • test/integration/release_packaging_test.rb
  • test/integration/navigation_test.rb (4 total)
• Cover: mounting, installation, release, navigation
• Missing: fetch pipeline end-to-end, scrape pipeline end-to-end, job error handling

Gap:

• No end-to-end fetch tests (HTTP → parsing → item creation → log entry → events)
• No end-to-end scrape tests (item selection → adapter → persistence → broadcast)
• Health check integration (source pause, recovery) not tested
• Retry logic and circuit breaker not tested

Estimated Test Count Needed: 8-12 additional integration tests

T5: System Tests Only Cover Happy Paths

Finding: System tests test success flows but avoid error scenarios.

Current State:

• System test files: 6
  • test/system/dashboard_test.rb
  • test/system/items_test.rb
  • test/system/logs_test.rb
  • test/system/mission_control_test.rb
  • test/system/sources_test.rb
  • test/system/dropdown_fallback_test.rb
• All tests use inline helpers (defined per-file)
• Tests create records, visit pages, assert UI renders
• No error state testing (404, 500, validation errors, timeout, network failure)

Gap:

• No error flow coverage
• No timeout/network failure UI handling
• No invalid state recovery
• No permission/auth error paths

Example Missing Test: "source deletion confirmation shows error if deletion fails"

T6: Duplicated Factory Helpers Across 20+ Test Files

Finding: Factory-like helpers are defined locally in each test file instead of centrally.

Current State:

• Centralized in test/test_helper.rb: only create_source! (lines 107-119)
• Local definitions found in:
  • test/lib/source_monitor/items/retention_pruner_test.rb:
    • build_source (lines 147-155)
    • create_item (lines 157-165)

Pattern: Each test file that needs item/source/log creation defines its own helper, e.g.:

# In retention_pruner_test.rb
def build_source(attributes = {})
  defaults = { name: "Source #{SecureRandom.hex(4)}", ... }
  create_source!(defaults.merge(attributes))
end

def create_item(source:, guid:, published_at:, title:)
  source.items.create!(guid:, url: "...", title:, ...)
end

Gap:

• Duplication across ~20+ test files (estimated, no full count)
• Inconsistent signatures
• Harder to maintain
• Scattered knowledge of factory patterns

Files Likely Duplicating:

• Any test file that uses Item.create!, Source.create!, ScrapeLog.create!, FetchLog.create!, etc.

T7: Inconsistent Mocking Approach

Finding: Tests use 3+ different mocking styles without standardization.

Current State:

• Style 1: Minitest::Mock (via .stub):
  rubyCopy

  # test/lib/source_monitor/fetching/advisory_lock_test.rb:40
  ActiveRecord::Base.connection_pool.stub :with_connection, ->(&block) { block.call(fake_connection) } do
    assert_raises(...)
  end
  

• Style 2: Class.new (anonymous class mocking):
  rubyCopy

  # test/lib/source_monitor/fetching/advisory_lock_test.rb:30-37
  fake_connection = Class.new do
    def exec_query(sql)
      # mock behavior
    end
  end.new
  

• Style 3: Inline stubs (direct method stubbing):
  rubyCopy

  # test/system/items_test.rb:99
  SourceMonitor::Scrapers::Readability.stub(:call, result) do
    # test code
  end
  

Gap:

• No consistency across test types
• Harder to teach/onboard (3 patterns to learn)
• Mixing implementation details with test code

T8: WebMock Stub Explosion in Some Tests

Finding: Some tests create large numbers of WebMock stubs inline.

Current State:

• test/test_helper.rb (lines 65): WebMock disables all external HTTP except localhost
• Stubs are created in test methods, not shared
• No centralized stub collection
• Stub naming not standardized

Gap:

• Tests with multiple HTTP calls create multiple stub_request calls
• Hard to understand which calls are being stubbed
• Difficult to debug stub mismatches
• No reusable stub library for common patterns

T9: System Test Setup/Teardown Not in Base Class

Finding: System tests define setup/teardown per-file instead of in a base class.

Current State:

• System tests require application_system_test_case.rb (no file found yet)
• Actual system tests define their own setup/teardown inline:
  rubyCopy

  # test/system/dashboard_test.rb:9-23
  def setup
    super
    SourceMonitor.reset_configuration!
    SourceMonitor::Jobs::Visibility.reset!
    SourceMonitor::Jobs::Visibility.setup!
    purge_solid_queue_tables
    SourceMonitor::Dashboard::TurboBroadcaster.setup!
  end
  
  def teardown
    SourceMonitor.reset_configuration!
    SourceMonitor::Jobs::Visibility.reset!
    purge_solid_queue_tables
    super
  end
  

• Helper methods defined in test files:
  • purge_solid_queue_tables (lines 167-184, dashboard_test.rb)
  • seed_queue_activity (lines 140-165, dashboard_test.rb)
  • apply_turbo_stream_messages (lines 186-226, dashboard_test.rb)
  • connect_turbo_cable_stream_sources (not shown but referenced)
  • capture_turbo_stream_broadcasts (not shown but referenced)
  • parse_turbo_streams (lines 229-245, dashboard_test.rb)
  • assert_item_order (test/system/items_test.rb:160-168)

Gap:

• Setup/teardown logic duplicated across files
• Shared helper methods not centralized
• No ApplicationSystemTestCase base class (missing file)
• Capybara wait configuration not centralized

Files Affected: All system test files (test/system/*.rb)

T10: Insufficient Controller Error Path Testing

Finding: Controller tests focus on happy paths; error scenarios are light.

Current State:

• Controller tests found: 6 files
  • test/controllers/source_monitor/application_controller_test.rb (55 lines, 4 tests)
  • test/controllers/source_monitor/source_fetches_controller_test.rb (44 lines, 2 tests)
  • test/controllers/source_monitor/health_controller_test.rb
  • test/controllers/source_monitor/logs_controller_test.rb
  • test/controllers/source_monitor/source_bulk_scrapes_controller_test.rb
  • test/controllers/source_monitor/source_health_checks_controller_test.rb
• Example test (source_fetches_controller_test.rb:14-27):
  rubyCopy

  test "queues a fetch and renders turbo stream" do
    source = create_source!(fetch_status: "idle")
    assert_enqueued_jobs 1 do
      post source_monitor.source_fetch_path(source), as: :turbo_stream
    end
    assert_response :success
  end
  

• Error test (lines 29-41): one error test per controller

Gap:

• One error test per controller is insufficient
• Missing: validation errors, permission denied, record not found, server errors
• No test for CSRF protection edge cases
• Toast/flash message error flow not tested

T11: Time-Dependent Tests Without Explicit Assertions

Finding: Tests use travel_to but assertions don't always validate the time-based behavior.

Current State:

• test/lib/source_monitor/items/retention_pruner_test.rb uses travel_to:
  rubyCopy

  test "removes items older than retention period" do
    travel_to Time.zone.local(2025, 10, 1, 12, 0, 0) do
      create_item(...)  # old item created
    end
    travel_to Time.zone.local(2025, 10, 10, 12, 0, 0) do
      create_item(...)  # recent item created
      result = RetentionPruner.call(source:)
      assert_equal 1, result.removed_by_age  # validates deletion
    end
  end
  

• Tests properly assert outcomes, but time travel setup is verbose
• No missing travel_back ensures (appear to be present)

Gap:

• Boilerplate time travel setup
• Harder to read test intent
• No test helper to simplify time-based test setup
• No ensure travel_back enforcer in base class

T12: Job Tests Mix Inline and Async Patterns

Finding: Job testing uses both :test and :inline adapters without clear convention.

Current State:

• Default adapter: :test (test_helper.rb:48)
  rubyCopy

  ActiveJob::Base.queue_adapter = :test
  

• Tests use with_inline_jobs when execution needed:
  rubyCopy

  # test/system/items_test.rb:101
  with_inline_jobs do
    click_button "Manual Scrape"
  end
  

• Some tests use assert_enqueued_with for :test adapter
• Other tests execute jobs inline directly

Gap:

• No clear convention for when to use each approach
• Naming with_inline_jobs suggests it should be in test_helper but it's in test_prof.rb
• No job test base class or pattern docs

T13: Counter Cache Tests Lack Atomicity Checks

Finding: Counter cache fields (items_count on Source) are not validated for atomicity.

Current State:

• Source model has counter_cache: true on has_many :items
• Tests create items and assert count:
  rubyCopy

  # test/system/items_test.rb:88
  initial_item_count = Item.count
  item = Item.create!(...)
  # Later assertion checks the count
  

• No tests for concurrency/race conditions
• No tests for counter cache recovery after corruption

Gap:

• No atomicity tests (concurrent item creation)
• No counter cache inconsistency recovery test
• Missing: reset_items_counter! testing

T14: Inconsistent Test Naming Style

Finding: Test names vary in style and specificity.

Current State:

• Verbose style: "removes items older than the configured retention period"
• Terse style: "filters items" (hypothetical)
• Mixed prefixes:
  • Some start with verbs: "respects configured scraper adapters"
  • Some start with nouns: "dashboard displays stats, job metrics, and quick actions"
• Test names in test_helper (lines 67-92) use comments for setup, not clear test intent

Gap:

• No naming convention documented
• Hard to scan test files for related tests
• Inconsistency makes test discovery harder

T15: No Shared Behavior Test Modules for Concerns

Finding: Concerns (Loggable, etc.) are tested inline; no shared test modules.

Current State:

• Loggable concern used by FetchLog, ScrapeLog, HealthCheckLog
• Each model tested separately (implied, no shared spec module)
• Example: test/lib/source_monitor/logs/entry_sync_test.rb tests log behavior directly

Gap:

• No shared test for Loggable behavior across 3 models
• Duplicated tests for concern behavior
• Harder to maintain concern contracts

T16: ApplicationSystemTestCase Missing Wait Configuration

Finding: ApplicationSystemTestCase doesn't exist or lacks Capybara wait config.

Current State:

• No test/dummy/app/test/application_system_test_case.rb file
• System tests hardcode waits: wait: 5, wait: 10, visible: :all, wait: 5
• Capybara default wait: 2 seconds (Rails default)
• Tests use explicit waits for async behavior:
  rubyCopy

  # test/system/dashboard_test.rb:86
  assert_selector "turbo-cable-stream-source", visible: :all, wait: 5
  

Gap:

• No centralized wait time configuration
• Arbitrary wait values (5s, 10s) scattered across tests
• No Capybara screenshot/save_page on failure setup
• No JavaScript driver logging
• No Capybara::Minitest integration configuration

T17: No Temp File Cleanup in Test Teardown

Finding: Tests that create temp files don't have centralized cleanup.

Current State:

• test/test_helper.rb includes cleanup for engine tables but not temp files
• test/support/host_app_harness.rb manages temp directories for integration tests
• No explicit FileUtils.rm_rf in test teardown blocks
• System tests may leave temp screenshots/artifacts

Gap:

• Temp files accumulate in test/tmp/
• test/dummy/tmp/ grows unbounded
• No automatic cleanup of failed test artifacts
• Disk space issues in long test runs

Existing Patterns

Factories (Centralized & Local)

Centralized in test_helper.rb:

def create_source!(attributes = {})
  defaults = {
    name: "Test Source",
    feed_url: "https://example.com/feed-#{SecureRandom.hex(4)}.xml",
    website_url: "https://example.com",
    fetch_interval_minutes: 60,
    scraper_adapter: "readability"
  }
  source = SourceMonitor::Source.new(defaults.merge(attributes))
  source.save!(validate: false)
  source
end

Local Patterns (per-file):

• build_source(attributes) - wraps create_source! with defaults
• create_item(source:, guid:, title:, published_at:) - creates Item records
• Direct Model.create! calls for other models (FetchLog, ScrapeLog, etc.)

Mocking Patterns

1. Minitest Stub (.stub):

   rubyCopy

   Object.stub(:method, return_value) { code }
   

2. Anonymous Class Mocking:

   rubyCopy

   fake = Class.new { def method; ...; end }.new
   

3. Direct Method Stubbing:

   rubyCopy

   Class.stub(:method, value) { code }
   

WebMock & VCR

• WebMock blocks all external requests except localhost
• VCR configured but underutilized (4 cassettes total)
• Tests typically stub individual requests inline:
  rubyCopy

  stub_request(:get, "https://example.com/feed.xml")
    .to_return(status: 200, body: File.read(file_fixture("feeds/rss_sample.xml")))
  

Test Setup/Teardown

Standard Setup (test_helper.rb:82-91):

setup do
  SourceMonitor.reset_configuration!
  SourceMonitor.config.http.retry_max = 0
end

System Test Setup (per-file):

def setup
  super
  SourceMonitor.reset_configuration!
  SourceMonitor::Jobs::Visibility.reset!
  SourceMonitor::Jobs::Visibility.setup!
  purge_solid_queue_tables
  # dashboard-specific setup
end

Test Isolation

Thread-Based Parallelism:

parallelize(workers: worker_count, with: :threads)

• Avoids PG fork segfault
• Each test gets fresh SourceMonitor config atomically

Scoped Assertions:

assert_equal 1, SourceMonitor::Item.where(source: source).count  # GOOD
assert_equal 1, SourceMonitor::Item.count  # BAD - counts across parallel tests

Recommendations

Priority 1: Foundational Infrastructure (Start Here)

1. Create ApplicationSystemTestCase base class (T16)

   • File: test/dummy/app/test/application_system_test_case.rb
   • Configure Capybara waits: default 5-10 seconds for async
   • Add screenshot/save_page on failure
   • Centralize helper methods: purge_solid_queue_tables, seed_queue_activity, apply_turbo_stream_messages, assert_item_order
   • Include cleanup hooks for temp files (T17)

2. Centralize Factory Helpers (T6)

   • Create test/test_helper_factories.rb or module
   • Consolidate: create_source!, build_source, create_item, create_fetch_log, create_scrape_log
   • Include in ActiveSupport::TestCase
   • Document signatures and defaults
   • Audit 20+ test files for duplicates and consolidate

3. Document VCR Strategy (T1)

   • Create test/VCR_README.md
   • Define naming convention: source_monitor/<domain>/<scenario> (existing partial convention)
   • Document when to record vs use cassettes
   • Add regeneration guide for stale cassettes
   • Implement cassette recording check in CI

Priority 2: Unit Test Gaps (Execute Phase 06)

4. Add Unit Tests for Sub-Modules (T3)

   • Create test/lib/source_monitor/fetching/feed_fetcher/adaptive_interval_test.rb
     • Test interval calculation logic, edge cases
   • Create test/lib/source_monitor/fetching/feed_fetcher/source_updater_test.rb
     • Test source state mutations, log creation, event firing

5. Add Integration Tests for Pipelines (T4)

   • Create test/integration/fetch_pipeline_integration_test.rb
     • HTTP fetch → parsing → item creation → log entry → events
   • Create test/integration/scrape_pipeline_integration_test.rb
     • Item selection → adapter → persistence → broadcast

Priority 3: Test Coverage Completeness (Execute Phase 06)

6. Add Error Path Tests for Controllers (T10)

   • Add 3-5 error tests per controller
   • Cover: 404, 422, 500, permission denied, timeout
   • Use stubs to simulate failures

7. Add Error Scenario System Tests (T5)

   • "Fetch fails with HTTP timeout" → user sees error toast
   • "Source deletion fails" → shows error, reverts UI
   • "Item scrape times out" → shows pending state, recovers

8. Standardize Mocking Approach (T7)

   • Adopt single pattern: Minitest stub (most concise)
   • Create mocking guide in CLAUDE.md
   • Refactor Class.new mocks to use stub

Priority 4: Consistency & Maintainability (Execute Phase 06 or later)

9. Remove Fixture Loading (T2)

   • Replace fixtures :all with explicit per-test setup_fixtures for tests that need them
   • OR: Move to factory-only approach (no fixtures)
   • Benchmark performance impact

10. Add Concern Test Modules (T15)

    • Create test/support/shared_loggable_tests.rb
    • Include in FetchLog, ScrapeLog, HealthCheckLog tests
    • Ensure consistent Loggable behavior contract

11. Standardize Test Naming (T14)

    • Convention: Start with action verb (imperative mood)
    • Format: "action [condition] [expected outcome]"
    • Example: "removes items older than retention period"

12. Enforce Time Travel Cleanup (T11)

    • Create travel_to_at(time) { code } helper in test_helper
    • Automatically calls travel_back in ensure
    • Reduces boilerplate, prevents travel leaks

13. Add Counter Cache Atomicity Tests (T13)

    • Test reset_items_counter! correctness
    • Add concurrency test (2 threads creating items)
    • Validate counter cache matches actual count

14. Centralize Job Testing Convention (T12)

    • Document when to use :test vs :inline adapter
    • Move with_inline_jobs to test_helper (out of test_prof.rb)
    • Add with_test_adapter helper for explicit test-mode tests

15. Add Temp File Cleanup Hook (T17)

    • Extend ApplicationSystemTestCase to clean test/tmp/* in teardown
    • Add Rails test log cleanup for failed tests

Effort Estimate

Recommended Phase 06 Scope:

• Implement ApplicationSystemTestCase + factories (Priority 1) = 4-6 hours
• Add sub-module unit tests (Priority 2) = 3-4 hours
• Add core integration tests (Priority 2) = 4-5 hours
• Add error path tests (Priority 3) = 3-4 hours
• Subtotal: 14-19 hours (1-2 weeks at 10 hours/week)

Future Phases:

• Priority 4 cleanup work (9-13 hours) as separate phase if time permits

Key Metrics & Baselines