I have finally been able to actually get some work done on JvYAML again. The plan is to have the 0.2 release out in two weeks time. There are three components left to finish. I have already completed the Emitter and the Serializer. What's missing is the Representer, JavaBean support and JRuby integration. But when this is finished, we will have a complete pure Java YAML library for JRuby
I really want Mongrel to work in JRuby. We have been spending some time getting the HTTP-parser ported to Java, and this is more or less finished now. Earlier today I managed to run all the Mongrel test cases for the parser successfully in JRuby, so we're clearly on the way. What we need right now is to get the first parts of the Extension API into JRuby proper. When that's done we can start testing Mongrel for real.
I have a goal to rewrite parts of this to abstract away differences between different drivers (we have some specialized code for Oracle, some for MySQL, and that's no good in the long run.) I'm also planning on writing some test cases using Migrations, so we can test the driver against many different databases. What I really want to do is to test MySQL, Oracle, MS SQL Server and PostgreSQL enough to say that JRuby can use them. Especially MS SQL would be good, since MRI only can use ODBC if not running production on a Win32 machine.
- JRuby performance and bug correction:
Right now these are the big points. We need to correct as many bugs as possible, nibble away at performance, and also improve documentation. I will try to take a look at these issues if time allows.
This is a new project that I have been thinking about and planning for the last few months. I won't say much more about it right now, but that it is something I'm really passionate about (even more than JRuby). Expect more information in a blog post soon.