Improvements to Optional<T> analysis

[findbugsproject]

Improvements to Optional analysis

from findbugs.

[KengoTODA]

Maybe these links are related with this issue:

• http://sourceforge.net/p/findbugs/feature-requests/297/
• http://sourceforge.net/p/findbugs/feature-requests/302/

[bgourlie]

Ahh, which bug tracker is the official one?

[amaembo]

@bgourlie this one is preferred now. I totally disagree as there are would be too much false positives. Simple case:

//input: List<int[]> data;
data.stream().map(arr -> IntStream.of(arr).max()).filter(OptionalInt::isPresent)
                     .map(OptionalInt::getAsInt).toArray();

Both isPresent and getAsInt are called, but at different stages of pipeline. How can this be tracked?

Another case:

public static int min(int[] data) {
    if(data == null || data.length == 0) return Integer.MAX_VALUE;
    return Arrays.stream(data).min().getAsInt();
}

Perfectly safe, but isPresent was not called. This is similar to complain about unchecked dereference of Map.get(): while you may catch some bugs, tons of false-positives would arise.