From: Charles Oliver Nutter <headius@...>
Date: 2013-04-12T16:40:30+09:00
Subject: [ruby-core:54214] Re: Encouraging use of CommonRuby

On Thu, Apr 11, 2013 at 3:50 PM, Marc-Andre Lafortune
<ruby-core-mailing-list@marc-andre.ca> wrote:
> My thought is that all feature requests for trunk should be considered
> Common Ruby.
>
> I'm not sure I see the point of having two separate "projects". What would a
> "bug" in Common Ruby mean?

Do you agree that Ruby and MRI are separate things? If so, we need to
avoid filing bugs or features against MRI that would lead to changes
in other Rubies.

I gave examples of what would *not* be ComnonRuby in my repliy to
naruse. If any implementation would need to implement said feature to
be "compatible", it should be in ComnonRuby. If it can be reasonably
defined as an MRI-specific change (env vars, RubyVM methods, build
changes, cross-platform behavioral differences, ...) it would be fine
to call a ruby-trunk issue.

I recognize there's a grey area here. I also recognize that feature
requests filed against ruby-trunk do not encapsulate similar visible
feature changes that happen in other Redmine projects. We need one
place where we can look up implenentation-impacting changes.

> Can't you filter on the subject of the message for "Feature"?
>
> If so, I would simply merge Common Ruby into trunk

I think there's a very clear line between MRI features and "Ruby"
features. I need to know the latter in order to make JRuby as
compatible as possible. I don't really care about the former.

- Charlie