Skip to content

0.5.0
Compare
Choose a tag to compare
@coreysprague coreysprague released this 25 Jun 13:29
· 163 commits to master since this release
0.5.0
386d611
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

More intelligent behavior for loading assets

A new mechanism has been added for ensuring that images have been decoded before capturing goldens. The old implementation worked most of the time, but was non-deterministic and hacky. The new implementation inspects the widget tree to identify images that need to be loaded. It should display images more consistently in goldens.

This may be a breaking change for some consumers. If you run into issues, you can revert to the old behavior, by applying the following configuration:

GoldenToolkitConfiguration(primeAssets: legacyPrimeAssets);

Additionally, you can provide your own implementation that extends the new default behavior:

GoldenToolkitConfiguration(primeAssets: (tester) async {
 await defaultPrimeAssets(tester);
 /* do anything custom */
});

If you run into issues, please submit issues so we can expand on the default behavior. We expect that it is likely missing some cases.

New API for configuring the toolkit

Reworked the configuration API introduced in 0.4.0. The prior method relied on global state and could be error prone. The old implementation still functions, but has been marked as deprecated and will be removed in a future release.

The new API can be invoked in flutter_test_config.dart to be applied for all tests within a given folder (or the entire package). Additionally, if there is a need to override configuration at a narrower scope, this API can be invoked in-line as well.

GoldenToolkit.runWithConfiguration((){/* callback() */}, config: GoldenToolkitConfiguration(/* custom config here */));

Added the ability to customize the generated filenames

When using screenMatchesGolden or multiGoldenFile, you can now supply your own functions for controlling the naming of the files. This can be done using the configuration API mentioned above.

GoldenToolkit.runWithConfiguration((){ /* callback() */}, config: GoldenToolkitConfiguration(fileNameFactory: (filename) => '' /*output filename*/));

There are two methods that can be overridden:

  • fileNameFactory is used for screenMatchesGolden
  • deviceFileNameFactory is used for multiScreenGolden

Future releases will likely consolidate these APIs.

Thanks to @christian-muertz for this enhancement.

Added additional utility functions for preparing for goldens

Extracted out some public extension methods that were previously private implementation details of multiScreenGolden & screenMatchesGolden

Added the following extensions. These can be used with any vanilla golden assertions and do not require multiScreenGolden, screenMatchesGolden, or GoldenBuilder.

// configures the simulated device to mirror the supplied device configuration (dimensions, pixel density, safe area, etc)
await tester.binding.applyDeviceOverrides(device);

// resets any configuration applied by applyDeviceOverrides
await tester.binding.resetDeviceOverrides();

// runs a block of code with the simulated device settings and automatically clears upon completion
await tester.binding.runWithDeviceOverrides(device, body: (){});

// convenience helper for configurating the safe area... the built-in paddingTestValue is difficult to work with
tester.binding.window.safeAreaTestValue = EdgeInsets.all(8);

// a stand-alone version of the image loading mechanism described at the top of these release notes. Will wait for all images to be decoded
// so that they will for sure appear in the golden.
await tester.waitForAssets();

Misc Changes

A few API / parameters were marked as deprecated and will be removed in future releases.

Assets 2
Loading