Fix HashMap when keys are objects that may override toHash and/or opEquals

[n8sh]

Includes a unittest demonstrating code that fails without this patch.

[n8sh]

The current test failure appears to be spurious so I'll just force-push to make the tests repeat until all pass.

[wilzbach]

Just ignore the spurious test failures. No point in wasting time with fighting against them.

[wilzbach]

Why not?

Suggested change

	return () @trusted { return a is b ? true : (a !is null && (cast(Object) a).opEquals(cast(Object) b)); }();
	return () @trusted { return a is b ? true : a == b; }();

[n8sh]

I wanted to avoid it because it does a bunch of work solely to support improperly designed class hierarchies that can have asymmetric opEquals:

https://github.com/dlang/druntime/blob/60d992da13027948b590e7804238e60bcefafde2/src/object.d#L1095-L1115

bool opEquals(Object lhs, Object rhs)
{
    // If aliased to the same object or both null => equal
    if (lhs is rhs) return true;


    // If either is null => non-equal
    if (lhs is null || rhs is null) return false;


    // If same exact type => one call to method opEquals
    if (typeid(lhs) is typeid(rhs) ||
        !__ctfe && typeid(lhs).opEquals(typeid(rhs)))
            /* CTFE doesn't like typeid much. 'is' works, but opEquals doesn't
            (issue 7147). But CTFE also guarantees that equal TypeInfos are
            always identical. So, no opEquals needed during CTFE. */
    {
        return lhs.opEquals(rhs);
    }


    // General case => symmetric calls to method opEquals
    return lhs.opEquals(rhs) && rhs.opEquals(lhs);
}