Possible bug in KTypeHashSet.add(KType key)

[Olivier G.]

Hello,

First thanks for this great library.

I think I found a bug in code or in JavaDoc: JavaDoc for ObjectHashSet.add(KType key) says when there is an already existing element with same key, it is replaced with the one provided in argument. But it is not the case.

I made a test to reproduce the problem (using JUnit and HPPC 0.7.2).

@Test

public void add() {

ObjectHashSet<Integer> set = new ObjectHashSet<Integer>();

set.add(new Integer(0));

Integer newKey = new Integer(0);

set.add(newKey);

Integer keyInSet = set.iterator().next().value;

assertSame(newKey, keyInSet);  // <-- Fails.

}

The key is not replaced, and KTypeHashSet source code seems to confirm that.

Regards,

Olivier

[Dawid Weiss]

Thanks Olivier, this looks like a bug. I'll take a look.

Dawid

> --
> You received this message because you are subscribed to the Google Groups
> "High Performance Primitive Collections for Java" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to
> java-high-performance-primi...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[Dawid Weiss]

I fixed this in the documentation, Olivier. The behavior here should
be I think consistent with HashMap (leaving existing entry in the
set).

http://issues.carrot2.org/browse/HPPC-164

[Olivier G.]

I think you meant: the behavior of HPPC KTypeSet.add(E) should be consistent with JCF Set.add(E) (that is leave the set unchanged).
Because for hash maps, JCF and HPPC are already consistent: put(K,V) replaces the existing value and returns the old one.

Olivier

>> java-high-performance-primitive-collections+unsub...@googlegroups.com.

>> For more options, visit https://groups.google.com/d/optout.

>> java-high-performance-primitive-collections+unsub...@googlegroups.com.

>> For more options, visit https://groups.google.com/d/optout.

>> java-high-performance-primitive-collections+unsub...@googlegroups.com.

[Dawid Weiss]

Yup, exactly.

D.

>> >> java-high-performance-primi...@googlegroups.com.

>> >> java-high-performance-primi...@googlegroups.com.

>> >> java-high-performance-primi...@googlegroups.com.

>> >> For more options, visit https://groups.google.com/d/optout.
>

> --
> You received this message because you are subscribed to the Google Groups
> "High Performance Primitive Collections for Java" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to

> java-high-performance-primi...@googlegroups.com.