On 04/04/14 18:19, Jason Merrill wrote:
I also notice that the opt_storage destructor is implicitly declared and
trivial; it seems excessive to warn in that case.
I disagree, I think that's exactly a case that should be warned about -- because
it's easy to overlook.
I'm also somewhat skeptical
about extending the warning to non-polymorphic bases at all, at least in the
absence of -Weffc++.
That's certainly plausible.
Since this change is causing problems, I'm inclined to back it out until after
4.9 branches.
That too.
I'm testing a patch that makes the test in the loop:
if (TREE_PUBLIC (base_binfo)
&& (TYPE_POLYMORPHIC_P (basetype) || warn_ecpp)
&& accessible_nvdtor_p (basetype))
but I agree about reverting the patch if there's still further discussion
following that suggested change.
nathan
