Right now we split the data set according to the requested number of columns at composables and use remember to cache that.
Obviously we use the UI thread for doing that, which can be avoided.
Moving the calculation back to ViewModel allows us to pick whichever dispatcher to do that, and we can include the results as a part of unit test (which in case we finally cover view model tests).
This is a low-priority work. Generally, it won't be done before we conclude this project in July 2024.
Right now we split the data set according to the requested number of columns at composables and use
remember
to cache that. Obviously we use the UI thread for doing that, which can be avoided.Moving the calculation back to ViewModel allows us to pick whichever dispatcher to do that, and we can include the results as a part of unit test (which in case we finally cover view model tests).
This is a low-priority work. Generally, it won't be done before we conclude this project in July 2024.