Ruby-Core Speaks on Ruby 1.8.8, 1.9.3 and 2.0
You may imagine that the ruby-core mailing list is a 24/7 programming disco dealing with core Ruby implementation topics.. but no, it's usually a low-traffic list with calm discussion of bugs and patches. This week, however, some Interesting Stuff™ has happened and the kimono has been lifted on a few issues including, notably, a potential Ruby 1.8.8.
Yui Naruse, one of the core Ruby committers, was trying to dispel myths that the Japanese-language ruby-dev mailing list is a hotbed of illicit Ruby implementation activity by inviting people to ask any implementation-related questions they liked in
ruby-core. From the results of this, we can pick up a few tidbits:
What's going on with Ruby 1.9.1?
Still alive but is less active because of yugui's resource. I think, it will die soon.
When will Ruby 1.9.3 be released?
Not decided but we want to release 1.9.x every 1.5 years (to prevent 1.9.2 become too stable one like 1.8.6). So it may be released in 2011-12-25 or 2012.
What about Ruby 2.0?
See Matz's presentation at RubyConf 2010. [..] But when we relase Ruby 2.0 is not decided. I think it takes several years [PC: emphasis mine].
Yes, several years. Ruby 2.0 is our Perl 6.
Ruby 1.8.8 - Another Update to the 1.8 Branch?
Yui also let slip that that there could be a Ruby 1.8.8 waiting in the wings and the next day Shota Fukumori posted Let's begin a talk for "1.8.8", an English translation of a Japanese ruby-dev post. It's a little hard to read but the general points are:
- Ruby 1.9.0 came out 3 years ago and Ruby 1.8.7 will be 3 years old in June 2011. This is the longest it has been between significant releases.
- Ruby 1.9.2 is recommended for new development but, it is claimed, many "can't use 1.9.x in actuality" because "extension libraries" do not work [I'm not au fait with libraries used in Japan but most of the mainstream ones now are fine IME].
- Ruby 1.8 must not become "uncontrolled" [Unmaintained, I'm assuming..]
From these points, Shota proposes three ways forward:
- 1.8.8 could be created as a "better 1.8.7" with a release date of Summer 2011.
- 1.8.8 can be worked on until it is "ideal" and users are not affected. A release date of 2020 is proposed. [This is the joke entry, I take it.]
- No 1.8.8 ever.
The ultimate question, though, is "What do we do?" Ultimately it's for the community to decide. Matz himself has said:
We haven't have any consensus yet. Maybe it's good time to start discussion. I myself don't think 1.8.8 is needed. But there might be demand for 1.8.8 or later.
So if this is something you care strongly for, get discussing. IMHO, I want to encourage people onto 1.9.2 as much as possible. The water's warm, the libraries are (finally) friendly, and Ruby 1.9 is better in many ways.
And if you're not yawning yet..
ruby-core news, Lucas Nussbaum kicked off a heated debate about the virtues of moving MRI development to Git (it ain't gunna happen) and Ryan "zenspider" Davis proposes that gem_prelude should die (based on the RubyGems team's experiences with Ruby 1.9.2 and the new RubyGems 1.4).