Hi JRubyists,
I voiced some concerns about the state of some JRuby libraries on
Twitter a few days ago[1]. Being someone that uses many nuances of Ruby,
I still run into JRuby<>MRI problems on a regular basis, mostly due to
library issues in JRuby-specific libraries like
activerecord-jdbc-adapter, not due to the base system itself. Often,
those problems are already known or reported, but unfixed - and
sometimes need some effort. Also, some of those libraries require
special knowledge e.g., database specifics or details of JRuby.
I would like to start a series of Ruby Hackdays to improve the Ruby
ecosystem under the umbrella of Ruby Berlin e.V… For the reasons stated
above, I would like to pick the JRuby ecosystem as our first project to
hack on. The goal of the first day is to get new hackers educated about
JRuby, provide them with worthwhile projects to work on, get them up to
speed and also to motivate them to work on the “boring” parts[2] from
time to time. If the day is a success, I’d like to do rerun it
regularly.
This is going to happen soon: We do have a room and will have a date set
by the next Ruby U.group Berlin meeting (10th of January) - expect it
to be at the end of January, beginning of February.
What we do need is support by project maintainers providing tasks and
maybe documenting their pain points with the project[3]. If you read
this mail and lead a JRuby project or a project with JRuby issues,
please get in contact with me - I’d like to compile a list of things for
attendees to get started with quickly. Also, depending on the date, it
would be nice if some support on IRC or github issues could be given to
those getting started with new things.
My current list is:
- activerecord-jdbc-adapter, especially on Oracle (in that case, along
with oracle-enhanced) is somewhat behind the current Rails
state-of-the-art and sometimes buggy because of that - warbler has a long issue list
- Generally, I find the documentation about Java-to-JRuby-integration
lacking. While there are big projects using it around (like Nokogiri),
it is hard to find guidelines. - On a similar note, I find documentation of actual differences between
MRI and JRuby lacking. - Finally, Kernel.spawn is a fake in JRuby ;).
- At least in my case, Bundler and the asset pipeline seem slower than
on MRI, I’d like to investigate that
Regards,
Florian G. [email protected]
[1] https://twitter.com/Argorak/status/279650537652248576
https://twitter.com/Argorak/status/279676648700399616
[2] Yes, Oracle is not fun :).
[3] Just a statement like: “Does not work well in 1.9-mode” or
something, nothing big.
