Migrating methods in Collections

Doug Lea dl at cs.oswego.edu
Thu Dec 24 18:36:15 UTC 2015


On 12/24/2015 11:01 AM, Brian Goetz wrote:

> I still think the logical signatures here would be:
>      E[] toArray()
>      <U super E> toArray(IntFunction<U[]> generator)

My thought was: Let them eat streams!
Which already supports:

     Object[] toArray();
      <A> A[] toArray(IntFunction<A[]> generator);

On the other hand, IntStream adds:  int[] toArray();

And similarly for LongStream, DoubleStream, but nothing that
accommodates arbitrary Value-Stream, so the issue might need
to be addressed somewhere somehow.

>
>>      boolean removeAll(AnyCollection<? extends E> c);   // not
>> Collection<?>
>>      boolean retainAll(AnyCollection<? extends E> c);   // not
>> Collection<?>
>
> Arguably better handled by remove(Predicate)

Sure. I agree there's no good rationale for these in
Collection super-interface. (On the off chance that we
actually split them...)

>
>>      long findFirst(E e);
>>      long findLast(E e);
>
> Arguably better as findXxx(Predicate)

I agree about adding predicate versions.
But people will complain about requiring a lambda in plain version
that otherwise ought to be simple enough to compile into
optimal machine instruction loop for ArrayLists



More information about the valhalla-spec-experts mailing list