It’s just data

Ruby 2.0 -= Continuations?

I see that the rumor that Ruby 2.0 is dropping continuations is spreading.  Can anybody provide a definitive reference?  The best I can find is this, and the decision doesn’t seem quite as final to me.

I was joking with Tim Bray yesterday over IM that I should take on a binding of REXML to libxml2 as a Christmas project, make sure that the cornerstone of the implementation depends on continuations, and deliver it a full year before Ruby 2.0 comes out.  I would of course make sure that it has better spec compliance than the current REXML regex based implementation (my personal hot button), but I suspect that the draw would be the performance benefits.


Just before RubyConf Matz made a stop here at BYU. During a question and answer session with the BYU RUG he was asked about continuations.

He said that they were difficult to include into YARV and so he decided that the 1.9.1 release would not include continuations even though he was very sad to see them go. He expressed a desire to include continuations in Ruby at some later date when there was enough time to get them integrated into YARV. He had no idea of when that could be.

Unfortunately I don’t have a link for you.  We did record the meeting but it was very hard to hear so I don’t have high hopes of having the video turn out.

Posted by Devlin at

At RubyConf this question was also asked, and it was very clear that it’s a matter of priority and difficulty.  Essentially Matz said that if we cared about continuations, we should talk to ko1.

I think your plan for REXML has great merit.  :-)

Posted by Glenn Vanderburg at

Anxiety In Ruby Community

The many implementations of Ruby have created elation for some and anxiety for others. While it does indicate that the language is popular it is easier for someone to create a fork creating multiple versions. So far, no one has shown any intention,...

Excerpt from iface thoughts at

As mentioned above, it’s been a question of priority.  Given a strong (non continuations web framework) reason for continuations, it could happen sooner.  Given some solid coding by someone to implement efficient continuations for one or more of the implmentations (YARV or JRuby especially), it would happen even sooner.

Evan Phoenix has worked on continuations based on the smalltalk bluebook model for his rubinius implementation, and that’s looking promising.

Posted by Pat Eyler at

In Smalltalks like VisualWorks and Squeak continuations require no VM support whatsoever. And yes, these VMs  use byte code (no byte codes for continuations) and continuations are fast enough to run continuation based web servers.

I don’t know what’s so hard about this. That stuff is more than 20 years old.

Posted by anonymous at

Daily Spit

Ruby might be losing continuation in 2.0. I predict Ruby will lose much of its forward momentum if this is true because, IMHO, continuation was the source of much of Ruby’s fun and brevity. Too bad. Yahoo continues its healthy ascent although I am...

Excerpt from Don Park's Daily Habit at

Libxml-ruby ([link]) is coming along nicely now - you could implement a REXML api clone on top of that, and maybe even squeeze continuations into a SAX strategy or similar :)

Posted by Ross Bamford at

Sun Frees Its JDK

I’m travelling today. I’m in Seattle, Washington for a few days of whole day meetings. I didn’t get a chance to surf the internet until 8:30pm local time. The first thing I noticed is that I got a lot of hits on some of my old “Free Java” blog...

Excerpt from Weiqi Gao's Observations at

Related discussion

Posted by Sam Ruby at

[link]...

Excerpt from del.icio.us/taq/continuations at

Sam Ruby: Ruby 2.0 -= Continuations?

[link]...

Excerpt from del.icio.us/subscriptions/whipsnap at

From Lang.NET 2008

Monday 0815 So here I am. It snowed last night, so the drive down was a little crazy. I’ve arrived. They say I represent “Collier Technologies” instead of “Amazon.com.” I hope nobody at work minds. The network is up,...

Excerpt from The Pædantic Programmer at

Add your comment