JRuby & Rails 2.2

We just started converting things over to Rails 2.2.2 and I thought I should say that the one feature that really kicks some ass is its thread safety. JRuby 1.1.5 (very soon to be 1.1.6), which takes advantage of Java's native threads, combined with this latest version of Rails finally makes for a very good deployment story.

Rails' traditional single threaded model, while conceptually nice for development, never really made things easy for deployment. In MRI-land, you always needed multiple processes, translating into multiple instances of Mongrel. In JRuby we needed multiple instances of the JRuby runtime. While it worked, it was definitely a bit of a memory pig. With Rails 2.2 we can now run with only one JRuby runtime in a JEE web container that was already multi-threaded. Now memory consumption is several fold smaller than what it was before and we now have a much better story to tell everyone. Thanks to everyone on the JRuby and Rails teams!

Now, I shouldn't just end with a big hurrah to everyone without mentioning that we did run into a problem with some thread unsafe code in our own app. It took awhile to find it (a week) so the one word of caution is be careful about any shared state in your application (in our case it was a rather obvious global variable that we seemed to keep overlooking).

In any case, Rails and JRuby are a great match for enterprise web app development. I've been hesitant to say that in the past because of a few rough edges like the one I mentioned before but now I think it's about time everyone should look at this fantastic combination.

UPDATE:

Thought I'd add a link to a recent Ruby shootout. It's reporting on the performance of JRuby vs other Ruby implementations and the importance of threading.

Heroku, RubyGems and Git

My daughter and I started a little Ruby on Rails project and I thought it'd be nice to host it somewhere for her to play with. I have an account on Heroku that I haven't used in months and thought it'd be nice to try it out again. I signed back in (after resetting my forgotten password) and was impressed to discover that they've integrated Git. After creating the project and navigating to the Rails start page I found some instructions:

gem install heroku
heroku clone stuffies
cd stuffies
[..local edits..]
git add .
git commit -m "local changes"
git push

So I ran off and installed the heroku gem. But had trouble with step 2. Fortunately a couple of google searches had me back up and running. One to deal with "Permission denied (publickey)" upon heroku clone. And another to figure out how to create the RSA key. Now I have a local Git clone of my project that I can work on in NetBeans and then I can quickly publish the changes to Heroku with a commit and a push. Sweeeet!!!

Calling Java from JRuby to do the dirty work

I recently had to try and read a UTF-16LE-BOM encoded text file from Ruby. I couldn't figure out how to get Ruby to deal with the double byte characters in the file. But I did know that Java has very good UTF support so I decided to let Java do the heavy lifting for me. I ended up with something like this:

require 'java'

import java.io.InputStreamReader
import java.io.ByteArrayInputStream
import java.io.BufferedReader

data = ''
File.open("text.txt", 'rb'){|f| data = f.read}

#strip the BOM (Byte Order Marker)
data.slice!(0..1)
#let Java deal with the UTF-16 encoding
reader = BufferedReader.new(
 InputStreamReader.new(
   ByteArrayInputStream.new(data.to_java_bytes), 'UTF-16LE'))

while ((s = reader.read_line) != nil)
 puts s
end

It may not be the perfect solution (I'd like to see a regular Ruby solution) but I suppose leveraging the Java libraries was the whole point of JRuby in the first place wasn't it!

(Executing File.open("text.txt", 'rb'){|f| puts f.read} looked fine on my Mac, except for the BOM, but looked terrible in the console on Windows. The solution above actually converts the text from double byte to single byte characters)

Really Achieving Your Childhood Dreams

By Randy Pausch. Great lecture from Carnegie Mellon's School of Computer Science. Here's the iTunes link and YouTube as well. Watch out for the head fake. Great stuff.

Lessons from my first Rails project

I’m finally being paid for working on an “Enterprise” Ruby on Rails pilot project. I think I helped persuade our IT group to give Rails a try. But unfortunately I wasn’t available to help the project get started. Instead I joined the project several months in. So while it’s cool to finally be doing Ruby and Rails full-time it’s also pretty painful to be living with some of the poor decisions made before I got here. I can’t say that if I’d been involved from the get-go that we would have avoided some of these problems (sometimes it runs deeper than the technology) but what I can offer is a list of things I’d suggest you watch out for (in no particular order):

Be realistic
In the early days of Rails I remember people saying things that made it sound that Rails was ten times more productive than Java. That’s a wild exaggeration and everybody should know it. There’s no way you’re going to be 10X more productive, end of story. Hope for “10% more productive” after you’ve got one or two projects under your belt. At least that’s realistic and achievable and if you get more than that then great! But for the love of God don’t start off with that idea. A big part of the software game is managing expectations. Don’t dig yourself a hole you have to try and climb out of. Despite what you read on the web, your first Rails project will probably take you more time, not less. New language, new tools, new framework, new infrastructure, new bugs, and new deployment problems. My project started with unrealistic expectations and now we’re over budget and over schedule. That’s not a great way to get management on your side.

Deploy early and often
That’s an agile thing and not unique to Rails, but getting over the hurdles of deployment shouldn’t be taken for granted. We’re primarily a Java shop so when people saw JRuby and WAR deployment, they figured everything would be easy-peasy. Not so. Part of being “Enterprise” is that quite often you’re several months, if not years, behind the technical curve. The first thing we needed was to get past Java 1.4 and an old version of Weblogic. We needed at least Java 6 and Tomcat 6. Trust me, getting “comfortable” IT infrastructure people to upgrade their technology can be very painful and slow.

Now assuming you got the basic app server running on a reasonably recent version of Java, go and figure out how to build the WAR (warbler gem), get to Oracle (activerecord-jdbc gem) through JNDI, get your Rails log files out of the exploded war directory (logging gem), integrate with ActiveDirectory for Authentication and you’ve got your hands full of a bunch of other problems that will eat away at your time. Budget for it.

Now once you got all that, deploy often! Our project has periodically deployed something outside of our dev environment but not frequently enough to gather any kind of momentum to satisfy our users. Without gaining the users’ trust you get into a whole other set of problems.

Unit test (RSpec seems pretty awesome)
Sounds like a reasonable thing to do... but we didn’t. I won’t go into all the reasons why testing is good, but to fail to take advantage of the testing infrastructure that Rails gives you out of the box is pretty ridiculous. Even if all you’re doing is executing the code and never asserting the results are correct at least you’re executing the code and will find runtime problems that a language like Java would have given you at compile time. I started using RSpec for my stuff and I really like it, but it sucks when the other developers won’t add to the tests or even run yours. But my strategy is to lead by example. Maybe sooner or later the other will pick it up too. If you can get buy-in early and be ruthless about keeping your testing effort on track.

Figure out the non-standard Rails stuff
Beyond the deployment stuff I talked about earlier, figure out the other places you must configure rather than adopting the Rails conventions. For example, we have a rule that says you must access the database from an account other than the schema owner. That means we needed to call set_table_name(‘schema.table’) in all of our models. We also needed to create the permissions for the non-schema owner. And don’t forget proper referential integrity constraints in your migrations. These are all things you can ignore for a while but if you have any hope of putting you application in your users’ hands you’re going to have to figure it out. Personally, I hate surprises like this late in the project. My recommendation is to sort them out early.

Don’t underestimate security (a.k.a., read the requirements doc)
We have users who don’t trust anyone and have dreamed up one of the most elaborate security mechanisms you can think of. Unfortunately we went through five months of development and not one of the developers before me even seriously looked at it. Sadly, I can understand why. Every Java developer (including me) thinks they got that one figured out and just skips that part of the requirements doc. Except that someone should have actually read ours and said Woah! It’s not always easy to retrofit something like this if you just assumed it was an orthogonal concern. If someone wrote a requirements document for you (no matter how badly written) be sure to read through the whole thing. You never know what problems may be lurking.

Read ‘The Rails Way’
In general, get some training. Or in lieu of that, get some good books. I bought “The Rails Way” by Obie Fernandez (great book btw) after I got assigned to the project. To my way of thinking the project should have already bought a whole bunch of Rails books and made them available to the developers. But it’s up to you to be prepared. If there are parts of the framework that are still a bit fuzzy start reading some books (go to the source if you need to).

Evaluate your editor
So like I said before, our shop is primarily a Java shop, so when I got on the project people were using IntelliJ to edit their code. Now IntelliJ may be a kick-ass Java IDE but from what I’ve seen it’s not great for doing Rails. So when I started, I used NetBeans. Now NetBeans has never in my mind been a great Java IDE, and quite frankly it’s not my dream Ruby IDE either but Sun hired the JRuby dudes and Tor Norbye, the principal NetBeans guy at Sun seems to understand Ruby and the needs of Rails developers pretty well. As near as I can tell, unless your a Mac-toting Textmate user, NetBeans is the best game in town for Rails development (I even use it on my Mac). So my recommendations is to get off your butt and seriously evaluate the tools you’re using. Maybe IntelliJ is the right thing for you, but you don’t really know unless you look around.

Use an issue tracking system
We didn’t have an issue tracking system when I started on this project. So not only could we not tell how far we were by test/spec status, we weren’t even tracking the change requests and bugs properly. Wow. In my mind this is part of the scope problem we’re having. The users just ask for something and without visibility to the whole team and management, it just gets magically added and the schedule slips.

Web 1.0
AJAX is pretty attractive to a Web 1.0 developer. But food is awfully attractive to a starving man too, but that doesn’t mean you should let him eat himself to death. Take it slowly. IMHO, do it old school on your first (and maybe your second or third) Rails project. Throw in the odd autocompleting field if you want but this AJAX stuff and all the pretty visual effects can get out of hand pretty quickly.

Use people who want it to succeed
Java and .NET are pretty typical technologies in IT and unfortunately a lot of the developers who’ve invested a lot of time learning this stuff seem to forget that the technology is only a means to an end and not the end itself. They’ll be threatened by new technology. They’ll be threatened by Ruby and Rails. So if you can, staff your project with the developers who are open minded and wouldn’t mind if a Rails application succeeded. If you push these hostile developers onto a Rails project their negativity will bring everyone else down.

Understand RESTful Rails
I don’t remember when exactly David Heinemeier Hansson really started pushing the idea of a resourceful/RESTful way to develop web applications with Rails, but it’s damn good stuff and you should really figure out what it all means before you start building a whole bunch of stuff the old way. Figure out Rails’ routing system and what map.resources does and then you’ll be able to go with the flow and let Rails do a lot more of the work for you. I’ve seen more than one developer write a lot more code than they need to.

Separate HTML, CSS, & Javascript
I can’t believe these so called IT web developers who still use all the old FONT, BORDER, WIDTH, ONCLICK etc. HTML elements in their templates. C’mon! HTML is mark-up, CSS is styling and JS is behaviour. Separate these aspects of your view layer the same way you separate models, views, and controllers. I’ll make some allowances for surgical uses of RJS but if you want things to be maintainable sort this out already. (I know this isn’t just about Rails but it bugs me so much I figured I’d rant about it anyway ;-))

If you’re about to embark on a Rails project, get ready to enjoy yourself, just don’t lose your head in all the hype and follow some good common sense.

My first JRuby contribution

I've reported lots of bugs before but this is the first time I submitted a fix back to an open source project.  The fix was to the JRuby-Extras ActiveRecord-JDBC project to allow you to do something like this for a Rails model class against Oracle:  

class Person

  set_table_name "department.people"

end

Accessing the database from someone other than the schema owner and therefore fully qualifying the table with schema is a rule I needed follow in order to deploy a webapp in my current IT department.  

I posted the original fix quite a while ago but nobody really took notice. Then Jesse Hu came along and made it better (and probably independently I might add) and also added a couple other fixes for things I hadn't noticed.  

But unfortunately neither one of us seemed to really know how to submit the fix so it got ignored for awhile.  Originally I created a patch and just pasted the text into the body of my comment.  Jesse attached the entire Java file.  A few days ago when Thomas Enebo asked for things we'd like to see fixed in the next release of JRuby I jumped at the chance.  After I got a hint about what to do from Charles Nutter, I got the latest version of the project from SVN, created a .patch file from Jesse's source, and then submitted it back to the JRuby-Extras Tracker and then finally back to JRuby's JIRA issue.

Today I got notice from Nick Sieger that it'll be in the next version.  Very cool.

Unfortunately I have to admit that I still a need a bit of work on my testing skills because I wasn't able to figure out where or how to insert some unit tests to prove that my fix worked.  Something to save for the next time. I'm just happy I can finally get rid of my custom version of this library.

Rails on Windows

I've been working on a Rails project at work for the last few weeks.  We've been targeting JRuby for deployment and therefore I've been working primarily in JRuby for development.  This week I thought it'd be nice to get running with the regular MRI with MySQL instead of Derby.  I installed cygwin and tried to get back to doing things at the command line (I've been using NetBeans)...  I was getting tired of the performance hit I was taking every time I ran a rake task or a generator or started up Mongrel.  I was under the impression that the MRI ran as well on Windows as it does on my Mac.  Not even close.  Performance still sucks and managing the differences between my JRuby installation and my cygwin hosted Ruby isn't much fun.  I'm about ready to bail and just go back to JRuby.  Now I know why people buy Macs to do Rails development.  It rocks.  Rails development on Windows on the other hand truly sucks.  Thank goodness JRuby and NetBeans are there to hand me a few niceties.  Otherwise I'm not sure I'd have the patience for it.

The iPhone SDK

I don't own an iPhone and since I'm in Canada I can't even buy an iPhone from Apple even if I wanted to, but that hasn't tempered my enthusiasm for the new iPhone SDK.  Like so many thousands of others I downloaded the SDK as soon it was available, installed it, and then started coding away.  I'm sure more experienced Cocoa developers were able to quickly bang out something cool, but not me.  It took me a bit longer to get the mandatory "Hello World" running, but nonetheless, my previous Objective-C/Cocoa investigations definitely came in handy.  I think one of my mistakes was thinking that the videos Apple put up would help me.  IMHO, don't bother.  Go right for the developer documentation in XCode instead.

Alas, I can't really say that much about the SDK.  I clicked through one of those license agreements and I'm sure that there's something in there that forbids me from revealing anything too juicy.  But I think it suffices to say that I'm still enthusiastic about where this is going.  I think the iPhone (even though I don't have one yet) is a great bit of technology and once in the hands of people in the enterprise who suffer at the other craptaculous mobile devices out there, they'll definitely start demanding stuff.  And that's where I fit in.  I'm excited to see how this all evolves and hope I'm ahead of the curve at least momentarily. ;-)

Javascripting with ExtJS

Back in the early days of my work in Java I adopted the typical Java developer's bias against Javascript. Whenever someone would mention Javascript I would cringe. Afterall, why would I choose Javascript when I had Java? But I always felt I was missing something.

So last year I decided to do something about it and integrated the Prototype Javascript library into my Struts application.  I quickly learned that I didn't hate Javascript but hated the inconsistencies in the browser DOMs and learned to properly direct my hate  (Internet Explorer, I'm looking at you!).  

Despite a couple hiccups, that project was really quite enlightening and I went on to expand my knowledge with a great series of videos at the YUI Theatre. Douglas Crockford has three series of videos that I highly recommend to anyone wanting to learn JS. He goes over the history of Javascript and plainly spells out Javascript's problems and strengths (e.g., it supports functions as first class objects with support for closure, hashes are all powerful, prototypical inheritance is different than classical inheritance but is quite workable, and be careful of the default global namespace).

So now armed with a proper understanding of the language, I started a building new web application in January and realized that there were requirement that would be best served by AJAX and some nice browser widgets.  In enters ExtJS. After a couple weeks writing a spike through Struts and ExtJS (still using Prototype underneath), I made the choice to adopt ExtJS as a the library for my application's user interface.  Two months later, I'm more than satisfied with the choice.  So are my users.  

In the end, I've shed a bias I really shouldn't have been harbouring, I embraced a "new" language and added the ability to better server my users.  I still won't choose Javascript for a lot of things but I'm much happier knowing that I can properly apply it as part of an overall solution.