Golden Toolkit
This project contains APIs and utilities that build upon Flutter's Golden test functionality to provide powerful UI regression tests.
It is highly recommended to look at sample tests here: golden_builder_test.dart
Table of Contents
Key Features
GoldenBuilder
The GoldenBuilder class lets you quickly test various states of your widgets given different sizes, input values or accessibility options. A single test allows for all variations of your widget to be captured in a single Golden image that easily documents the behavior and prevents regression.
Consider the following WeatherCard widget:
You might want to validate that the widget looks correct for different weather types:
testGoldens('Weather types should look correct', (tester) async {
final builder = GoldenBuilder.grid(columns:2, widthToHeightRatio: 1)
..addScenario('Sunny', WeatherCard(Weather.sunny))
..addScenario('Cloudy', WeatherCard(Weather.cloudy))
..addScenario('Raining', WeatherCard(Weather.rain))
..addScenario('Cold', WeatherCard(Weather.cold));
await tester.pumpWidgetBuilder(builder.build());
await screenMatchesGolden(tester, 'weather_types_grid');
});
The output of this test will generate a golden: weather_types_grid.png that represents all four states in a single test asset.
A different use case may be validating how the widget looks with a variety of text sizes based on the user's device settings.
testGoldens('Weather Card - Accessibility', (tester) async {
final widget = WeatherCard(Weather.cloudy);
final builder = GoldenBuilder.column()
..addScenario('Default font size', widget)
..addTextScaleScenario('Large font size', widget, textScaleFactor: 2.0)
..addTextScaleScenario('Largest font', widget, textScaleFactor: 3.2);
await tester.pumpWidgetBuilder(builder.build());
await screenMatchesGolden(tester, 'weather_accessibility');
});
The output of this test will be this golden file: weather_accessibility.png:
See tests for usage examples: golden_builder_test.dart
DeviceBuilder
DeviceBuilder class is like the GoldenBuilder except that it constrains scenario widget sizes to Device configurations. This removes the need to specify a column or grid based layout.
It will generate a widget that lays out its scenarios vertically and the Device configurations of those scenarios horizontally. All in one single golden png file.
testGoldens('DeviceBuilder - multiple scenarios - with onCreate',
(tester) async {
final builder = DeviceBuilder()
..overrideDevicesForAllScenarios(devices: [
Device.phone,
Device.iphone11,
Device.tabletPortrait,
Device.tabletLandscape,
])
..addScenario(
widget: FlutterDemoPage(),
name: 'default page',
)
..addScenario(
widget: FlutterDemoPage(),
name: 'tap once',
onCreate: (scenarioWidgetKey) async {
final finder = find.descendant(
of: find.byKey(scenarioWidgetKey),
matching: find.byIcon(Icons.add),
);
expect(finder, findsOneWidget);
await tester.tap(finder);
},
)
..addScenario(
widget: FlutterDemoPage(),
name: 'tap five times',
onCreate: (scenarioWidgetKey) async {
final finder = find.descendant(
of: find.byKey(scenarioWidgetKey),
matching: find.byIcon(Icons.add),
);
expect(finder, findsOneWidget);
await tester.tap(finder);
await tester.tap(finder);
await tester.tap(finder);
await tester.tap(finder);
await tester.tap(finder);
},
);
await tester.pumpDeviceBuilder(builder);
await screenMatchesGolden(tester, 'flutter_demo_page_multiple_scenarios');
});
This will generate the following golden:
flutter_demo_page_multiple_scenarios.png
multiScreenGolden
The multiScreenGolden assertion is used to capture multiple goldens of a single widget using different simulated device sizes & characteristics.
The typical use case is to validate that a UI is responsive for both phone & tablet.
testGoldens('Example of testing a responsive layout', (tester) async {
await tester.pumpWidgetBuilder(WeatherForecast());
await multiScreenGolden(tester, 'weather_forecast');
});
This will generate the following two goldens:
weather_forecast.phone.png
weather_forecast.tablet_landscape.png
You can also specify the exact device configurations you are interested in:
await multiScreenGolden(
tester,
'weather_forecast',
devices: [
const Device(
name: 'strange_device',
size: Size(100, 600),
),
const Device(
name: 'accessibility',
size: Size(375, 667),
textScale: 2.5,
)
],
);
Getting Started
A Note on Golden Testing:
Goldens aren't intended to be a replacement of typical behavioral widget testing that you should perform. What they provide is an automated way to provide regression testing for all of the visual details that can't be validated without manual verification.
The Golden assertions take longer to execute than traditional widget tests, so it is recommended to be intentional about when they are used. Additionally, they can have many reasons to change. Often, the primary reason a golden test will fail is becaue of an intentional change. Thankfully, Flutter makes it easy to regenerate new reference images.
The rule of thumb that the eBay Motors team has used is to minimize our "full-screen" goldens and reserve them for high-level visual integration tests. Often, we only have one multiScreenGolden() test per screen.
For testing variations, edge cases, permutations, etc, we tend to use the GoldenBuilder focused on smaller widgets that have fewer reasons to change.
Setup
If you are new to Flutter's Golden testing, there are a few things you might want to do
Add the failures folder to .gitignore
When golden tests fail, artifacts are generated in a failures folder adjacent to your test. These are not intended to be tracked in source control.
# don't check in golden failure output
**/failures/*.png
Add a "golden" tag to your project
Add a dart_test.yaml file to the root of your project with the following content:
tags:
golden:
This will indicate that goldens are an expected test tag. All tests that use testGoldens() will automatically be given this tag.
This allows you to easily target golden tests from the command-line.
Configure VS Code
If you use VSCode, we highly recommend adding this configuration to your .vscode/launch.json file in the root of your workspace.
{
// Use IntelliSense to learn about possible attributes.
// Hover to view descriptions of existing attributes.
// For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
"version": "0.2.0",
"configurations": [
{
"name": "Golden",
"request": "launch",
"type": "dart",
"codeLens": {
"for": ["run-test", "run-test-file"]
},
"args": ["--update-goldens"]
}
]
}
This give you a context menu where you can easily regenerate a golden for a particular test directly from the IDE:
Loading Fonts
By default, flutter test only uses a single "test" font called Ahem. This font is designed to show black spaces for every character and icon. This obviously makes goldens much less valuable if you are trying to verify that your app looks correct.
To make the goldens more useful, we have a utility to dynamically inject additional fonts into the flutter test engine so that we can get more human viewable output.
In order to inject your fonts, we have a helper method:
await loadAppFonts();
This function will automatically load the Roboto font, and any fonts included from packages you depend on so that they are properly rendered during the test.
Material icons like Icons.battery will be rendered in goldens ONLY if your pubspec.yaml includes:
flutter:
uses-material-design: true
Note, if you need Cupertino fonts, you will need to find a copy of .SF UI Display Text.ttf, and .SF UI Text.ttf to include in your package's yaml. These are not included in this package by default for licensing reasons.
The easiest and recommended way to invoke this, is to create a flutter_test_config.dart file in the root of your package's test directory with the following content:
import 'dart:async';
import 'package:golden_toolkit/golden_toolkit.dart';
Future<void> testExecutable(FutureOr<void> Function() testMain) async {
await loadAppFonts();
return testMain();
}
For more information on flutter_test_config.dart, see the Flutter Docs
Caveats
Unfortunately, Flutter's font loading support for testing is fairly limited.
At the moment, it is only possible to load a single .ttf file for a font family. This means it may not be possible to get the proper visuals to appear. For example, different font weights may not work:
Additionally, in some instances, it is not possible to replace the "Ahem" font. There are specific places in the Flutter codebase, such as rendering the "debug banner" where no explicit font family is specified. In these instances, the engine will use Ahem in a test context, with no way to override the behavior.
testGoldens()
It is possible to use golden assertions in any testWidgets() test. As the UI for a widget evolves, it is common to need to regenerate goldens to capture your new reference images. The easiest way to do this is via the command-line:
flutter test --update-goldens
By default, this will execute all tests in the package. In a package with a large number of non-golden widget tests, we found this to be sub-optimal. We would much rather run ONLY the golden tests when regenerating. Initially, we arrived at a convention of ensuring that the test descriptions included the word 'Golden'
flutter test --update-goldens --tags=golden
However, there wasn't a way to enforce that developers named their tests appropriately, and this was error-prone.
Ultimately, we ended up making this testGoldens() function to enforce the convention. It has the same signature as testWidgets but it will automatically structure the tests so that the above flutter test command can work.
Additionally, the following test assertions will fail if not executed within testGoldens:
multiScreenGolden()
screenMatchesGolden()
Pumping Widgets
Flutter Test's WidgetTester already provides the ability to pump a widget for testing purposes.
However, in many cases it is common for the Widget under test to have a number of assumptions & dependencies about the widget tree it is included in. For example, it might require Material theming, or a particular Inherited Widget. Often this setup is common and shared across multiple widget tests.
For convenience, we've created an extension for WidgetTester with a function pumpWidgetBuilder to allow for easy configuration of the parent widget tree & device configuration to emulate.
pumpWidgetBuilder has optional parameters wrapper, surfaceSize, textScaleSize
This is entirely optional, but can help reduce boilerplate code duplication.
Example:
await tester.pumpWidgetBuilder(yourWidget, surfaceSize: const Size(200, 200));
wrapper parameter is defaulted to materialAppWrapper, but you can use your own custom wrappers.
Important: materialAppWrapper allows your to inject specific platform, localizations, locales, theme and etc.
Example of injecting light Theme:
await tester.pumpWidgetBuilder(
yourWidget,
wrapper: materialAppWrapper(
theme: ThemeData.light(),
platform: TargetPlatform.android,
),
);
Note: you can create your own wrappers similar to materialAppWrapper
See more usage examples here: golden_builder_test.dart
Configuration
There are global settings that can be configured by calling the following API:
GoldenToolkit.runWithConfiguration()
Currently, the primary option is to allow consumers to holistically skip golden assertions. For example, perhaps you only want to perform golden assertions on certain platforms.
See here as an example: flutter_test_config.dart
License Information
Copyright 2019-2020 eBay Inc.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
- Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
- Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
- Neither the name of eBay Inc. nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
3rd Party Software Included or Modified in Project
This software contains some 3rd party software licensed under open source license terms:
-
Roboto Font File: Available at URL: https://github.com/google/fonts/tree/master/apache/roboto License: Available under Apache license at https://github.com/google/fonts/blob/master/apache/roboto/LICENSE.txt
-
Icons at: Author: Adnen Kadri URL: https://www.iconfinder.com/iconsets/weather-281 License: Free for commercial use
-
OpenSans Font File: Available at URL: https://github.com/googlefonts/opensans License: Available under Apache license at https://github.com/googlefonts/opensans/blob/master/LICENSE.txt