Parrot_PMC_get_pmc : Does it work correctly?

[dukeleto]

I added the following test here

 https://github.com/parrot/parrot/commit/ac2ad789fc094d33a9db1f984e85ed100047be49

# TODO: Is this test correct?
extend_vtable_output_is(<<'CODE', <<'OUTPUT', "Parrot_PMC_get_pmc");
    Parrot_PMC_set_integer_native(interp, pmc, 42);
    Parrot_PMC_push_pmc(interp, rpa, pmc);
    Parrot_PMC_push_pmc(interp, rpa, pmc);
    Parrot_PMC_push_pmc(interp, rpa, pmc);

    pmc3 = Parrot_PMC_get_iter(interp, rpa);
    pmc2 = Parrot_PMC_get_pmc(interp, pmc3);
    Parrot_printf(interp,"%P\n", pmc2);
CODE
3
Done!
OUTPUT

It currently passes, but is it correct? i.e. If get_pmc is called on a RPA iterator, should it return the number of elements in the RPA?

[cotto]

After a bit of digging, here's what's happening. VTABLE_get_pmc on ArrayIterator returns the array-like PMC being iterated, which is sensible. The confusion comes from the fact that Parrot_printf (after many layers of magic and gotos) calls VTABLE_get_string in the array, which returns a stringification of the number of elements in the array. In the example above, looking at the type of pmc2 will make it clearer what's happening.

[dukeleto]

This ended up being me not understanding stuff. No bugs here, move along, nothing to see here.