Skip to content

ARCHITECTURE.md

Latest commit

 

History

History
74 lines (48 loc) · 3.74 KB

ARCHITECTURE.md

File metadata and controls

74 lines (48 loc) · 3.74 KB

Architecture

Directory overview

tests/automation/
├── setup/        # App launch and typed test fixtures
├── utilities/    # Reusable action helpers (send message, create contact, etc.)
├── locators/     # Page Object Model — all UI selectors
├── constants/    # Shared test data and configuration values
└── types/        # TypeScript type definitions

Typed test builders

Tests are defined using typed builder functions that handle window launch, user creation, and teardown. The function name encodes the scenario:

test_Alice_1W_Bob_1W(
  'some test name',
  async ({ alice, aliceWindow1, bob, bobWindow1 }) => {
    // windows are open, users are created
  },
);

Naming convention:

  • User names: Alice, Bob, Charlie, Dracula
  • 1W / 2W = number of windows for that user; 2W is a linked device (second window using the same recovery phrase)

Teardown (force-close all windows) runs in a finally block. test_Alice_1W_no_network skips waiting for the network to be ready.

All builders are in tests/automation/setup/sessionTest.ts.

Test context

Builders accept an optional TestContext object that injects conditions into the Electron process at launch:

Field Effect
dbCreationTimestampMs Sets DB_CREATION_TIMESTAMP_MS — mocks the account's database creation time (used to simulate account age, e.g. for donation CTA tests)
networkPageNodeCount Sets SESSION_MOCK_NETWORK_PAGE_NODE_COUNT — mocks the node count shown on the network page (1–10)

Locators

Selectors are centralised in tests/automation/locators/index.ts, organised into classes by page or feature area. testId() targets data-testid attributes; className() and hasText() are also available.

Utilities

Multi-step actions (send message, create contact, set disappearing messages, etc.) are in tests/automation/utilities/.

Playwright project split

playwright.config.ts defines two projects:

Project Match pattern Parallelism
Community tests **/*community*tests.spec.ts Sequential (workers: 1)
All other tests everything else Parallel (workersCount() workers)

Community tests run sequentially because they share the same community server. Parallel execution causes incoming messages to trigger UI jumps that close open context menus, making tests unreliable.

Localization

tests/localization/ is a git submodule pointing to Session Desktop's localization strings. pnpm sync runs sync-localization.sh, which anchors the submodule pointer to the strings version matching the current Session Desktop build.

Maintenance notes

Locators break when the app UI changes. Update the relevant class in tests/automation/locators/index.ts.

Baseline screenshots need updating when intentional UI changes ship. Delete the affected files in screenshots/ and rerun the affected tests — Playwright auto-saves a new baseline when none exists. Baselines are platform-specific (-darwin / -linux); update on both platforms before committing.

Dependabot is configured in .github/dependabot.yml