Because we’re making UI text changes, I’ve forked our views into “long” and “short” versions that preserve the previous experience while allowing the shorter experience to be different.
One limitation of the current approach is that the screenshot tests are limited to the current configuration, so there is some risk of ‘bit rot’ with the screenshot tests for the longer onboarding. For this PR, I manually switched the feature flags and re-ran the screenshot tests to make sure they still worked.
---------
Co-authored-by: Honza <rychnovsky.honza@gmail.com>
Twitter had a set of Compose rules, which have been forked into Detekt, Ktlint, and Android Lint implementations.
I see ktlint as mostly being formatting, Detekt being more for multiplatform static analysis, and Android Lint for Android-specific issues. Android lint is slow, so I didn’t want to use it as the first choice.
I went with the Detekt implementation which can be useful if we leverage multiplatform for our Composables in the future.
Carter Jernigan