Hi,

On Fri, Jul 25, 2008 at 11:50 AM, Jeremy Henty <onepoint / starurchin.org> wrote:
> On Fri, Jul 25, 2008 at 05:08:30PM +0900, Shugo Maeda wrote:
>> Hi,
>>
>> 2008/7/25 Federico Builes <federico.builes / gmail.com>:
>> > If you see any other problem please let us know.
>>
>> Currently ruby_1_8_6 has 2 rubyspec failures.
>
> ... and ruby_1_8_7 has exactly the same failures.

Yes, there was a change in RubySpecs yesterday, that triggers these
two failures, we'll deal with them in RubySpecs. The problem is that
different implementations have different policy/behavior w.r.t. to
(insignificant) spaces in yaml. Some produce them, and some others
don't. Nothing really serious/critical.

Also, since RubySpec is always evolving, with the commits daily, there
is always a chance that all tests passed 5 minutes ago and now there
are some failures due to new commits.

In JRuby (and Rubinius) we deal with this fluid situation by
introducing the notion of "stable" or "frozen" specs, fixing particual
revision to avoid newly-introduced changes and to reduce instability.

From time to time, the "stable/frozen" pointer gets updated to newer
revision (and the excludes are updated, and bugs are filed if needed)
so that there are no failures.

Thanks,
  --Vladimir