Hi Vladimir, hi Urabe,

Most of these 57 rubyspec failures are caused by the convert_type
change, that made convert_type check for private methods, too. This
change has already been reverted in 1.8.6 and 1.8 HEAD (even in
1.8.5), but not yet in 1.8.7.

Simply backporting http://svn.ruby-lang.org/cgi-bin/viewvc.cgi?view=3Drev&r=
evision=3D17396
into the 1.8.7 branch should fix these and many, many more failures
that we are currently seeing when running the latest 1.8.7 versions
against the rubyspecs.

Thanks, Arthur

On 2 Jul., 16:51, "Vladimir Sizikov" <vsizi... / gmail.com> wrote:
> Hi Urabe,
>
> * 1.8.6 HEAD has 7 rubyspec falures for StringIO
> * 1.8.7 HEAD has *57* rubyspec failures for StringIO
>
> Take a look here (for 1.8.7 faliures)http://pastie.org/226299
>
> Would it be possible to fix at least 1.8.7 branch?
>
> Please note that 1.8 HEAD passes all 100% StringIO rubyspecs.
>
> Thanks,
> =A0 --Vladimir
>
>
>
> On Wed, Jul 2, 2008 at 12:41 PM, Urabe Shyouhei <shyou... / ruby-lang.org> =
wrote:
> > Hello, I think current 1.8.6/1.8.7 is stable than p230/p22, so I decide=
d
> > to release them again. Expected date for that is Jul. 4 (in Japanese
> > Timezone).
>
> > Can you check to see if they works on your machine? =A0I don't want to
> > release again and again (that make users confuse.) =A0I want to make ne=
xt
> > release acceptable for you.