Hi,

So, once again, the question of which should be the minimum Java version to run JRuby. We are currently at 1.4.2, but the proposal is to move to 5 for trunk and all major releases from now on. This does not concern the 1.0.x-series.

I wrote a blog post about it, and got quite many replies. Almost all of them very positive about going to 5, but there were a few that represents my gut feel about it. Take this for example, from Jesper: I think for a lot of users one of the most compelling aspects of JRuby is it's ability to sneak Ruby into the enterprise. If you decide to ditch 1.4, many enterprises will not be able to run JRuby (and Rails) on their existing infrastructure - and these guys *don't* just upgrade. Besides from Websphere 6 and Weblogic 9 also Oracle Application Server 10g is using 1.4.2. For now, the cutting edge shops are running CRuby/Mongrel and in the enterprise shops we can sneak in a little magic with JRuby. So for now, please stay on 1.4.


I also asked on one of the internal Java discussion lists for TW, and the results were mixed to. What was obvious was that we needed some way to handle backwards compatibility, and retroweaver wouldn't be enough. Further, staying on 1.0.x may not be an option either, since we're doing substantial changes in trunk now.

I'm feeling deeply torn, especially since I know many of the features of Java 5 will make our lives easier. Can we go the middleground? Development and building has to be done with Java 5 or later, meaning that the source code can contain annotations, which we can use to do code generation with apt. We can force the resulting jars to be 1.4-compliant anyway. We could also make dual distributions, where the 5+ distribution contains real concurrency instead of the backport, and so forth.

I don't really know how to solve this problem in a good way.

Cheers

--
Ola Bini (http://ola-bini.blogspot.com) JRuby Core Developer
Developer, ThoughtWorks Studios (http://studios.thoughtworks.com)

"Yields falsehood when quined" yields falsehood when quined.



---------------------------------------------------------------------
To unsubscribe from this list please visit:

   http://xircles.codehaus.org/manage_email

Reply via email to