From: "NARUSE, Yui" <naruse@...>
Date: 2010-07-15T15:45:33+09:00
Subject: [ruby-core:31295] Re: [Bug #3562] regression in respond_to?

Hi,

2010/7/15 Marc-Andre Lafortune <ruby-core-mailing-list@marc-andre.ca>:
> Hi,
>
> On Tue, Jul 13, 2010 at 11:35 PM, NARUSE, Yui <naruse@airemix.jp> wrote:
>> So I want revert this change to ease maintaining RubySpec.
>
> RubySpec is already fixed, with the exception of lib/delegate; I'm
> waiting for decisions to be made to ask wether delegates should
> forward protected methods or not (and thus change the specs & lib
> accordingly).

I'm talking about that two failures.

> I am not saying to revert it or not, just that maintaining RubySpec is
> not a problem.
>
> It's quite unfortunate that no decision was made in time for 1.9.2.

It is needless to say that it won't merged in 1.9.2 because of feature freeze.

> Did we consider adding a warning in 1.9.2 if respond_to(:foo, false)
> is called for a protected method :foo?
>
> Another solution would be to revert the change and introduce
> `respond_to_public?`. If the flag of respond_to? defaulted back to
> true, then neither `send(:foo) if respond_to?(:foo)` nor
> `send_public(:bar) if respond_to_public?(:bar)` would raise a
> NoMethodError.
>
> I'll adapt RubySpec when a decision is made; I also understood that
> the change was for 1.9.3 and coded accordingly.

Matz says the change won't be in 1.9.3 but 2.0 in [ruby-core:31244].

-- 
NARUSE, Yui
naruse@airemix.jp