I'm looking at the comment docs
/// An optional [buildWhen] can be implemented for more granular control over
/// how often [BlocBuilder] rebuilds.
/// [buildWhen] should only be used for performance optimizations as it
/// provides no security about the state passed to the [builder] function.
/// [buildWhen] will be invoked on each [bloc] `state` change.
/// [buildWhen] takes the previous `state` and current `state` and must
/// return a [bool] which determines whether or not the [builder] function will
/// be invoked.
and
/// The [selector] function which will be invoked on each widget build
/// and is responsible for returning a selected value of type [T] based on
/// the current state.
final BlocWidgetSelector<S, T> selector;
And it reads to me like in case of buildWhen the check will be performed all the time, while the rebuilt won't happen if it returns false.
In the Selector case it reads like the build method will be called every time, but the state will be the same if the selector returns same result.
This does not really add up to what I'm seeing when debugging the code, when the builder is only called when the Selector state changes, same as when buildWhen returns true.
So what are the actual differences in this 2 approaches? Are there any at all except for the syntax?