Commit Graph

4 Commits

Author SHA1 Message Date
Alexander Brown
f2d5bfab73 test(browser): refactor browser tests for reliability and maintainability (#8510)
## Summary

Major refactoring of browser tests to improve reliability,
maintainability, and type safety.

## Changes

### Test Infrastructure Decomposition
- Decomposed `ComfyPage.ts` (~1000 lines) into focused helpers:
- `CanvasHelper`, `DebugHelper`, `SubgraphHelper`,
`NodeOperationsHelper`
- `SettingsHelper`, `WorkflowHelper`, `ClipboardHelper`,
`KeyboardHelper`
- Created `ContextMenu` page object, `BaseDialog` base class, and
`BottomPanel` page object
- Extracted `DefaultGraphPositions` constants

### Locator Stability
- Added `data-testid` attributes to Vue components (sidebar, dialogs,
node library)
- Created centralized `selectors.ts` with test ID constants
- Replaced fragile CSS selectors (`.nth()`, `:nth-child()`) with
`getByTestId`/`getByRole`

### Performance & Reliability
- Removed `setTimeout` anti-patterns (replaced with `waitForFunction`)
- Replaced `waitForTimeout` with retrying assertions
- Replaced hardcoded coordinates with computed `NodeReference` positions
- Enforced LF line endings for all text files

### Type Safety
- Enabled `no-explicit-any` lint rule for browser_tests via oxlint
- Purged `as any` casts from browser_tests
- Added Window type augmentation for standardized window access
- Added proper type annotations throughout

### Bug Fixes
- Restored `ExtensionManager` API contract
- Removed test-only settings from production schema
- Fixed flaky selectors and missing test setup

## Testing
- All browser tests pass
- Typecheck passes


<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **Tests**
* Overhauled browser E2E test infrastructure with many new
helpers/fixtures, updated test APIs, and CI test container image bumped
for consistency.

* **Chores**
* Standardized line endings and applied stricter lint rules for browser
tests; workspace dependency version updated.

* **Documentation**
* Updated Playwright and TypeScript testing guidance and test-run
commands.

* **UI**
* Added stable data-testids to multiple components to improve
testability.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

---------

Co-authored-by: Amp <amp@ampcode.com>
Co-authored-by: GitHub Action <action@github.com>
Co-authored-by: github-actions <github-actions@github.com>
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
2026-02-03 12:29:40 -08:00
Alexander Brown
82bacb82a7 test: add Playwright test tags for filtering (@smoke, @slow, @screenshot, domains) (#8441)
## Summary

Adds structured test tags to all 54 Playwright test files to enable
flexible test filtering during development and CI.

## Tags Added

| Tag | Count | Purpose |
|-----|-------|---------|
| `@screenshot` | 32 files | Tests with visual assertions
(`toHaveScreenshot`) |
| `@smoke` | 5 files | Quick essential tests for fast validation |
| `@slow` | 5 files | Long-running tests (templates, subgraph,
featureFlags) |
| `@canvas` | 15 files | Canvas/graph rendering tests |
| `@node` | 10 files | Node behavior tests |
| `@ui` | 8 files | UI component tests |
| `@widget` | 5 files | Widget-specific tests |
| `@workflow` | 3 files | Workflow operations |
| `@subgraph` | 1 file | Subgraph functionality |
| `@keyboard` | 2 files | Keyboard shortcuts |
| `@settings` | 2 files | Settings/preferences |

## Usage Examples

```bash
# Quick validation (~16 tests, ~30s)
pnpm test:browser -- --grep @smoke

# Skip slow tests for faster CI feedback
pnpm test:browser -- --grep-invert @slow

# Skip visual tests (useful for local development without snapshots)
pnpm test:browser -- --grep-invert @screenshot

# Run only canvas-related tests
pnpm test:browser -- --grep @canvas

# Combine filters
pnpm test:browser -- --grep @smoke --grep-invert @screenshot
```

## Implementation Details

- Uses Playwright's native tag syntax: `test.describe('Name', { tag:
'@tag' }, ...)`
- Tags inherit from describe blocks to child tests
- Preserves existing project-level tags: `@mobile`, `@2x`, `@0.5x`
- Multiple tags supported: `{ tag: ['@screenshot', '@smoke'] }`

## Test Plan

- [x] All existing tests pass unchanged
- [x] Tag filtering works with `--grep` and `--grep-invert`

┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-8441-test-add-Playwright-test-tags-for-filtering-smoke-slow-screenshot-domains-2f76d73d36508184990ec859c8fd7629)
by [Unito](https://www.unito.io)

---------

Co-authored-by: Amp <amp@ampcode.com>
Co-authored-by: GitHub Action <action@github.com>
2026-01-29 16:34:56 -08:00
Benjamin Lu
bc4549244e test(e2e): align test default menu to Top; make legacy specs explicit (#5746)
## Summary
UseNewMenu has been defaulted to Top in the app for over a year;
Playwright’s test default lagged behind. This PR aligns the test default
with reality and keeps legacy specs stable.

## Changes
- tests(e2e): default to 'Top' via fixture; specs that previously relied
on the old implicit default now explicitly set 'Comfy.UseNewMenu' to
'Disabled'.
- docs(browser-tests): remove outdated README note suggesting tests set
'Top' manually.

## Review Focus
- Intentional uses of 'Top' and 'Bottom' remain unchanged.
- Confirm ComfyPage default remains 'Top' (see
browser_tests/fixtures/ComfyPage.ts).

## Screenshots (if applicable)
N/A

┆Issue is synchronized with this [Notion
page](https://www.notion.so/PR-5746-test-e2e-align-test-default-menu-to-Top-make-legacy-specs-explicit-2786d73d365081218d06c1346f3ae18e)
by [Unito](https://www.unito.io)

---------

Co-authored-by: github-actions <github-actions@github.com>
2025-09-24 18:43:22 -07:00
Simula_r
c74c1c014f Feat/adaptive lod threshold (#5249)
* feat: use a physical min font size as the LOD threshold instead of an abritrary zoom level that is different on different screens

* feat: min 1px font size, max 24ox font size

* refactor: settings text

* refactor: settings text

* refactor: version

* fix: default font size from 10 to 8

* feat: cache the threshold and move it's calculation out of the render loop

* refactor: comments

* refactor: removed package-lock

* refactor: improve how we manage deprecated settings, and removed any types

* refactor: how the migration settings formula works so we get prev settings closer to the new font size setting

* test: add in zoom and settings test for LOD

* refactor: tests to use best practices

* Update test expectations [skip ci]

---------

Co-authored-by: github-actions <github-actions@github.com>
2025-09-03 12:38:44 -07:00