On 03/07/2008, Urabe Shyouhei <shyouhei / ruby-lang.org> wrote:
> Charles Oliver Nutter wrote:
>  > There should not be a release until the failures in rubyspec are
>  > addressed.
>
>
> OK, so all of you can wait until rubyspec passes?  I think I cannot
>  review every failures in a 24 hours so releases should be delayed. Is it
>  really OK?
>
>  I'm neutral whether we should make an immediate release, or we need more
>  brush-up to release.

While it is nice to have a release on the planned date I find the
quality of the release much more important.

The security release of 1.8.6p230 which contained a lot of untested
changes caused a lot of headache.

Thus delaying the release when problems are found is much better then
releasing known broken code. People who want to start testing at the
scheduled date can always check out from svn.

It may or may not be feasible to set up automated testing of the code
as each commit is done but testing at least the release is important
in my view.

Thanks

Michal