Scott Patten's Blog

A Truly Minimum Viable Kitchen

2014-02-24T00:00:00-08:00

I read Matt Maroon’s Minimum Viable Kitchen post (and the ensuing Hacker News discussion) a while ago, and I’ve been thinking about it off and on since then.

Not about the post per-se, but about what a minimum viable kitchen was, and whether there was anything interesting in the concept.

In the original post, Matt goes into great detail on a huge list of things that are needed for a Minimum Viable Kitchen. It quickly becomes obvious that Matt is not a minimalist. The list includes things like a stand mixer, a scale, and a 10-piece pot and pan set. Nothing wrong with that. I own a stand mixer and a scale, and I use them all the time, but I certainly wouldn’t put them on my “desert-island kitchen list”. I don’t own 10 pots and pans. I do agree that you need good quality ones, but 10 seems like a bit much, even for a non-minimal kitchen. I’d rather have the extra cupboard space.

Truly Minimal

So, let’s scale it back a bit. What is a truly minimalist kitchen, and what can you make with it?

I’m going to assume that you are living in a standard house that comes with a fridge, stove and oven. Let’s pretend you are just moving into your first place and have nothing else. What do you need to cook with? I’m going to ignore anything you use to eat, we’re just talking about getting food to the table here.

Okay how’s this for a first shot.

A cutting board
A chef’s knife
Stuff to sharpen the knife (honing steel and sharpener)
A 12” cast-iron frying pan with a lid
A medium-sized pot
A metal spatula and a metal spoon

With this and a few ingredients, you can make a lot of great meals.

A simple omelette with cheese, mushrooms and bacon
Broccoli and Beef Stir-Fry.
Roast Chicken surrounded by Onions and Garlic (Use the roasted garlic and drippings to make a gravy in the pan when the chicken is done.)
Chickpea curry with coconut sauce.
Rice pilaf
Fajitas
A frittata
This amazing Puerto Rican style Pork Shoulder
Pancakes
Home-fries
Corn Bread
Roasted beets on a bed of greens with fresh goat cheese, dressed with olive oil and balsamic vinegar

This is about as minimal as I can imagine. It’s almost viable, and probably would be fine for a while.

Why bother with a minimal kitchen anyways?

The whole point of this Minimum Viable Product (oops, sorry, kitchen) thing is to learn and learn quickly. So try this for a while and find out what you’re missing. Use all of your tools and listen carefully to yourself, and you’ll know what you’re missing. Think about what you’re making, what you can’t make that you’re craving, and what about your setup drives you crazy.

Do that for a while, and (carefully! cautiously!) buy a few things, and you’ll have a truly viable Kitchen.

No Excuses

2012-04-23T00:00:00-07:00

I wrote this in response to a post on HN. I could link to the post, but it doesn’t matter. There are posts like this all the time. Stop making excuses. Go build something.

This is going to sound more dramatic than it was, but here’s one way of summarizing the last five years of my life:

I was 35, working at a great job as a physicist. Amazing team (I was usually the dumbest guy in the room), amazing project, but I was bored. I knew that if this didn’t make me happy, I needed to get out. I read a few PG essays, caught the startup bug, and started coding in the evenings.

Within 8 months, with a brand-new daughter and another that was two, I was gone from that job, working on my own. In that 8 months I had taught myself web-dev from scratch.

I knew nothing of Ruby, HTML, CSS or JavaScript. I had a smattering of Unix knowledge and had been coding data-analysis in Python and C for a while, but the amount I had to learn was staggering. I bought tons of books, learned tons, and (luckily) had no idea how much I still had to learn.

The first startup idea (online graphing software for physicists – I didn’t exactly reach far for the idea) was a complete flop. That’s okay, though. I turned to consulting for a while. Gave some talks. Wrote part of a book. Met my co-founder by consulting for him for a while.

Now, things are good. Leanpub is growing steadily, and I couldn’t be happier.

The point is, that you can and should teach yourself some new skills. You’re way closer to being employable than I was, that’s for sure. It may take some evenings and weekends, but they’ll be some of the most fun evening and weekends you ever have. Find something that you want to do and build it. Fall in love with the idea, but be ready to dump it for something better.

The market for people with even a smattering of Rails knowledge is huge. iPhone dev is bigger. or Erlang. or fancy JavaScript using Socket.IO.

Location doesn’t matter. Age doesn’t matter (I’m 40 now, in Vancouver, not in SV, and I bet I could find 10 jobs at the drop of a hat). My blog isn’t huge. My startup hasn’t had “an exit”.

There are excuses, and there is getting out there and having a blast learning and making your life better.

Seems like an easy choice. Go kick some ass.

Getting up and running with Amber

2011-12-08T00:00:00-08:00

I was quite excited by Yehuda Katz’ announcement of Amber.js this morning. I had been intrigued by SproutCore, but hadn’t really liked how big it felt. So, when Yehuda said “If you played with SproutCore and liked the concepts but felt like it was too heavy, give Amber a try”, I listened.

The following is mostly just notes of what I learned. It’s not at all complete, but it should get you up and running and get you something to play with quickly.

Getting it on your machine

Assuming that you have git, Ruby 1.9.2 and a recent version of bundler installed, this is pretty straightforward. So, use rvm use 1.9.2 or the rbenv equivalent if you need to. These commands should do it:

    git clone https://github.com/amberjs/amber.js.git
    cd amber.js
    bundle install
    rake dist

This will create sproutcore.js and sproutcore.min.js in the dist directory.

Creating Hello World

Now that we’ve got the Javascript files, we can make a new project skeleton for a hello world app. We want to end up with an helloworld.html file where we put our HTML, and a javascripts directory where we put some JS files to include.

mkdir hello_world
cd hello_world
mkdir javascripts
touch helloworld.html
cd javascripts
wget http://code.jquery.com/jquery-1.7.1.js
ln -s jquery-1.7.1.js jquery.js
ln -s <path-to-sproutcore.js>
cd ..

Okay, we’re all set now.

The first thing we’re going to do is create a SC application (it’s still called SC right now, presumably this will be renamed to Amber sometime soon) and a Planet model. We’ll then instantiate an instance of this model and show it off on the page.

Edit helloworld.html to look like this:

<html>
<head>
<title>Hello World</title>

<!-- JavaScript includes -->
<script type="text/javascript" src="javascripts/jquery.js"></script>
<script type="text/javascript" src="javascripts/sproutcore.js"></script>

<script type="text/javascript">
HelloWorld = SC.Application.create();

HelloWorld.Planet = SC.Object.extend({
  name: null,
  star: null,
  position: 0
});

HelloWorld.currentPlanet = HelloWorld.Planet.create({name: 'Earth', star: 'Sol', position: 3})
</script>

</head>
<body>

<script type="text/x-handlebars">
  Hello !
</script>

</body>

</html>

Let’s go through that step-by-step.

First, we load up jquery.js and sproutcore.js. Make sure to load jquery.js first or things will blow up.

Next, we set up our SproutCore (again, presumably soon to be Amber) application with

HelloWorld = SC.Application.create();

For now, I’m thinking of this as a name-space for the app.

Next, we make our Planet class.

HelloWorld.Planet = SC.Object.extend({
  name: null,
  star: null,
  position: 0
});

This sets up a planet with name, star and position attributes. The values of the hash are the default values.

Finally, we create an instance of a planet with HelloWorld.Planet.create, passing in a hash of attributes and assigning it to HelloWorld.currentPlanet.

That’s the end of the script.

Now comes the crazy part. We put a Handlebars template in the body tag, and it gets processed by Amber and stuck in to the DOM. The part inside the double-curly-braces gets replaced with what you would expect: the name of the planet we created. Also, it is bound to HelloWorld.currentPlanet, so any time the name of the planet changes, our view will change.

Play around with the demo.

Try opening up your JavaScript console and changing the name of the planet (hint: HelloWorld.currentPlanet.set('name', 'Venus')).

Hooking up a form

Okay, that’s useful, but we can’t expect our users to open up JavaScript consoles. Let’s hook our model up to a form.

To do this, we’ll create a view and a template. Notice that we’re taking a short-cut here and skipping the ‘C’ part of the MVC triumvarate. This is not a good idea, of course, but I’m getting tired and I don’t want to make this post too long.

The view is going to be a simple text field, so we’re going to extend SC.TextField. Put this at the bottom of your script block in the head of the HTML:

HelloWorld.PlanetNameView = SC.TextField.extend({
  insertNewline: function() {
    var value = this.get('value');

    if (value) {
      HelloWorld.currentPlanet.set('name', value);
      this.set('value', '');
    }
  }
});

And put this below the handlebars template in the body of the HTML:

<br />
<script type="text/x-handlebars">
Rename the planet: 
</script>

Here’s a demo to play around with.

I think that’s it for now. So far, this is looking pretty fun and magical.

Speeding up Jekyll Generation

2011-11-18T00:00:00-08:00

One of the only complaints you’ll see out there about Jekyll is that when sites get bigger, it starts to slow down.

Luckily, this is easy to fix. Unfortunately, it’s apparently not obvious as you still see people complaining about it. Here’s my attempt at proselytizing for Jekyll :).

If you run jekyll --help, you’ll see that there’s a --limit_posts option. If you set this to 1, then you’ll only re-generate your most recent post when you save. This is usually exactly what you want. If you’re working on a slightly older post, then bump it up to 3 or 4.

Like this:

 jekyll --limit_posts 1

I’ve put this in a rake task

    namespace :jekyll do

      desc "start the jekyll server in auto mode"
      task :server, :num_posts do |t, args|
        num_posts = args[:num_posts]
        cmd = "jekyll --auto --server --pygments"
        cmd += " --limit_posts #{num_posts}" if num_posts
        puts "running #{cmd}"
        exec(cmd)
      end

    end

So I can start Jekyll like this:

rake jekyll:server[1]

and only see the current post, and it’s blazing fast.

Opening up ports to your security group on EC2

2011-11-18T00:00:00-08:00

Say you have a cluster of EC2 instances that you want to be able to talk to each other, but you don’t want everyone in the world to be able to join in on the conversation. For example, I was just setting up a typical cluster of servers:

A rails app server
A DB server
A daemon server
A DB slave

I want all of these servers to be able to talk to each other over port 3306 (the MySQL port), but I don’t want the whole world to be able to connect over port 3306.

You need two things:

A security group
Your EC2 user id.

Assuming you have your ec2 command line tools set up already, here’s how you would do it. This will create a group called yoursecuritygroup with ports 22 (ssh), 80 (http) and 443 (https) open to the world, but with all other ports only open to other computers in the same security group.

$> ec2-create-group --description "yoursecuritygroup production" yoursecuritygroup
$> ec2-authorize yoursecuritygroup -p 22
$> ec2-authorize yoursecuritygroup -p 80
$> ec2-authorize yoursecuritygroup -p 443
$> ec2-authorize yoursecuritygroup -o yoursecuritygroup -u 1234-1234-1234

You need to add your user id here in place of 1234-1234-1234. You can find this by going to https://aws-portal.amazon.com/gp/aws/developer/account?ie=UTF8&action=access-key and scrolling to the bottom. You want your AWS Account ID.

Now when you spin up your instances, make sure to start them in the yoursecuritygroup group using the --group argument:

ec2-run-instances --key your-key --group yoursecuritygroup --block-device-mapping /dev/sda1=:100:false --instance-initiated-shutdown-behavior stop --disable-api-termination --instance-type m1.small  ami-a7f539ce

And you should be all set.

Multiple Google Analytics Tracking Codes on one Page

2011-11-11T00:00:00-08:00

Say I want to make it so that Leanpub authors can get Google Analytics information about visits to their page while still sending the info to our Leanpub analytics account. This is pretty easy to do, but finding out how took me a bit of googling.

Your google analytics script for a normal, single tracking code looks something like this:

      var _gaq = _gaq || [];
      _gaq.push(['_setAccount', 'UA-1234567-1']);
      _gaq.push(['_trackPageview']);

      (function() {
        var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
        ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
        var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
      })();

That’s all fine and dandy, but what if you want to have two codes in there?

Turns out it’s pretty simple. You add another set of _gaq.push lines, right after the current ones and before (function() {. Something like this:

      _gaq.push(['account2._setAccount', 'UA-1234567-1']);
      _gaq.push(['account2._trackPageview']);

The account2 part can be anything you want. In fact, you can stick in a bunch of them if you want, resulting in something like this:

      var _gaq = _gaq || [];
      _gaq.push(['_setAccount', 'UA-1234567-1']);
      _gaq.push(['_trackPageview']);
      _gaq.push(['account2._setAccount', 'UA-1234568-1']);
      _gaq.push(['account2._trackPageview']);
      _gaq.push(['account3._setAccount', 'UA-1234569-1']);
      _gaq.push(['account3._trackPageview']);
      _gaq.push(['account4._setAccount', 'UA-1234570-1']);
      _gaq.push(['account4._trackPageview']);

      (function() {
        var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true;
        ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
        var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s);
      })();

Nice!

More info can be found here: http://code.google.com/apis/analytics/docs/tracking/asyncUsageGuide.html#MultipleCommands.

How big is a Googol, anyway?

2011-05-14T00:00:00-07:00

A friend of mine told his son that a Googol is a huge number. My kids caught on, so now I hear googol used in conversation all the time.

You know, like this:

I could eat 100 chocolate Easter eggs!

I could eat 1000!

Oh Yeah? Well, I could eat a googol!

A googol is 10¹⁰⁰. That’s 1, followed by 100 zeroes. It’s a ridiculously big number. It’s a number purely designed to awe the imagination. It’s not really useful in real life.

After hearing googol-bombs dropped all the time like this, you sort of get inured to it. A couple of days ago, though, I started thinking about how big a googol really was. My friend (the same one who brought googol in to my life) and I tried to think of something where googol was a reasonable scale, and came up blank.

Huh.

So, the question I want to ask is: is there anything that has a googol of it in the universe?

Let’s start with some standard “big quantity” things.

The number of grains of sand on the earth.

I don’t really care about accuracy here. So, I’ll just google it. The first answer says

There are seven quintillion five quadrillion grains of sand on all the beaches of the world. That’s a 75 with 17 zeros following!

So, in scientific notation that’s 7.5 x 10¹⁸. It’s a big number, but we’re nowhere near close.

How many stars are there?

Okay, let’s up the ante a bit. Another quick google search yields this answer:

With this simple calculation you get something like 10²² to 10²⁴ stars in the Universe. This is only a rough number, as obviously not all galaxies are the same, just like on a beach the depth of sand will not be the same in different places.

Obviously, these numbers are nowhere near big enough. Surely the number of atoms is going to be bigger.

How many atoms are in the universe?

From the Wikipedia article on the Observable Universe (which is a wealth of numbers for making these kinds of calculations):

two approximate calculations give the number of atoms in the observable universe to be close to 10⁸⁰.

We’re getting closer!

Well, it may feel like it, but we’re still ridiculously far off. We’re off by a factor of 10²⁰. That’s part of the mind-bend of playing around with googol. The biggest number we’ve come up so far with is 10⁸⁰. That means we’d need 10²⁰ universes just like ours to have a googol of atoms. Yikes!

How many particles are in the universe?

So, if we ask how many particles are in the universe, the number of fermions (things that make up matter, like protons and neutrons and electrons) isn’t going to get much larger than the number of atoms. Most of the universe is made of hydrogen, which has one proton and one electon. Even if we pretended that the atoms in the universe were made of some heavy element (say, Uranium), and then break the protons and neutrons up into their component quarks, we’d only add two or three orders of magniture to the number.

Okay, so what about photons and neutrinos and stuff like that? Well, I found something here:

Of course, besides material particles there are also lots of photons and neutrinos flying around the Universe. It is estimated that there are about 1e9 times as many photons and neutrinos as atoms in the Universe.

So, that gives us ~ 10⁸⁹. Closer, but still off by a factor of 10¹¹.

It looks to me like our initial goal of finding something tangible that had more than a googol of it has failed.

Just for fun, let’s make one more calculation.

Packing the universe with sand

Okay, let’s go crazy here. Remember the grains of sand in the earth question we started with? Let’s pretend we pack the whole universe with grains of sand. Yeah, the whole thing. Ignore questions of gravitational collapse and turning the whole universe into a giant black hole of what used to be sand and all that.

So, from that same Wikipedia page on the observable universe:

The observable universe is thus a sphere with a diameter of about 28 billion parsecs (93 billion, or 9.3 × 10¹⁰, light years). Assuming that space is roughly flat, this size corresponds to a comoving volume of about 3 × 10⁸⁰ cubic meters.

Cool! So that means, based on our number of atoms in the universe number above, there’s 1 atom for every three cubic meters in the observable universe. That’s not enough! Let’s see what happens if we pack the whole thing with sand.

This page here gives a reasonable estimate for the size of a grain of sand. The number they come up with is 8 per cubic millimeter. Let’s make life easier and round that up to 10. That means we’ll get 10 grains/mm³ X 10⁹ mm³/m³ = 10¹⁰ grains/m³. Multiplying that by the volume of the universe (and if you’re feeling bored, think about how crazy that phrase is…) gives us 10⁹⁰ grains of sand.

That’s right, if we fill the entire universe with grains of sand we’d still need 10¹⁰ universes to get a googol grains of sand.

Hmph. Okay, but how many atoms would that be?

To make things easier, let’s just make the universe one giant silicon crystal. Silicon has a density of 2.3 g/cm³ (again, Wikipedia is our friend), or 2.3 x 10⁶ g/m³. Silicon has an atomic mass of 28 g/mol (the stable isotope of Silicon has 14 protons and 14 neutrons).

So, a cubic meter of silicon has

 2.3 x 10⁶ g/m³ / 28 g/mol X 6.022 10²³ atoms/mol

 =

 4.95 X 10²⁸ atoms/m³

Ooo, we may be on to something here.

Multiply that by the volume of the universe:

 4.95 X 10²⁸ atoms/m³ X 3 × 10⁸⁰ m³ = 1.48 X 10¹⁰⁹ atoms

= 1.48 X 10⁹ googol atoms!

We did it!

We had to make a ridiculous assumption that could never happen, but in the end managed to get a number that was bigger than a googol.

Phew.

What did I miss?

Thanks for reading this far.

If you’re interested, upvote it on hacker news.

What did you think? Is there any reasonable number of real-live countable things in our current universe that has more than a googol? Perhaps if we include gravitons, or dark matter, or something like that?

Also, this post was written in about an hour in a coffee shop, so I’m sure I’ve made errors. Let me know if you find any!

Booting an Alestic EC2 Image with a Larger Root Disk

2011-05-06T00:00:00-07:00

I always start up my new EC2 images using Alestic’s awesome Ubuntu images.

I typically pick the EBS boot images. My only complaint has been that the EBS size is too small. They’re only 8 GB by default.

In the past, the fix has been to follow these directions: start up a server, terminate it, detach the volume, make a snapshot of it, create a new AMI from the snapshot and then start up a new server with that AMI.

Too much work for me.

Today I was starting up another server, and I did a bit of googling and found out that with the latest images, you can just pass in a parameter to your ec2-run-instances command and set the boot disk size to anything you want. Woo hoo!

Here’s the google groups thread. Look for the message from Scott Moser.

Here’s the command I used to create a new instance with a 100GB boot disk:

ec2-run-instances --key your-ec2-key \
 --block-device-mapping /dev/sda1=:100:false \
 --instance-initiated-shutdown-behavior stop \
 --disable-api-termination --instance-type m1.small  ami-06ad526f

The magic part is the --block-device-mapping /dev/sda1=:100:false. This takes three params. The first (which is blank here) allows you to pick a volume to mount as the root. If it’s blank, then a new EBS volume will be created.

The second (100) is the size of the boot disk. Change the 100 there to be whatever you want the size of the boot disk to be, in GB.

The third is whether or not to destroy the volume when you kill the instance, and can be either true or false. Unless this is a throwaway instance, you’re going to want this to be false.

SLL Subscribers Vs Time

2010-09-10T00:00:00-07:00

First, if you’re not a fan of Eric Ries’ Startup Lessons Learned, this post won’t be very fun for you. Luckily, there’s a very simple remedy for that: go there right now and read a few articles! There’s some amazing stuff, and I guarantee that you’ll learn something new and thought provoking.

If you’re intrigued and want to read all of the posts, then I also humbly recommend you go and buy his blog converted to a PDF on Leanpub: http://leanpub.com/startuplessonslearned. It’s much easier to read it as a PDF than clicking through all of the blog posts, and you get updates every month or so with all of the new posts.

Okay, now that that’s out of the way, let’s get down to the data!

Every once in a while, Eric asks his subscribers to do a survey for him so he can find out about them. The title of the survey is always something like “Lo, my 50 subscribers, who are you”. The latest one is Lo, my 57692 subscribers, who are you?.

Just for fun, I decided to plot the number of subscribers Vs time. Here are the results:

Hockey stick graph? I think so!

That naturally leads to the question: can Eric legitimately say that he has exponential growth? Let’s take a look…

Sorry, Eric, but the data does not support that conclusion. You can say that you have second order polynomial growth.

I know, not quite as sexy, but I guess it’ll just have to do.

The Myth of the Zone

2010-08-28T00:00:00-07:00

I’ve been thinking about how I work best for a while, and I was inspired by Peter’s post on his new workflow/schedule to write something up about it.

First, the provocative title is only partly link-bait. For me, the zone really is a myth. I keep hearing and reading about coders getting into zones and coming out of the mist 10 hours later with reams of awesome code (and, presumably, a really full bladder).

I’ve never had this happen.

Maybe it’s just me, and everyone else does get into a zone regularly. Maybe the zone isn’t as universal as hacker literature makes it out to be. Maybe the zone is a young hacker thing, and I’m getting towards the old side of the scale. Or, maybe it really is a myth.

Whatever the answer is, my version of “The Zone” looks like this:

Start working on something.
Hit a wall where I can’t find a solution or the solution I’ve come up with is too complicated.
Realize “Hey, I’ve hit a wall. I should change the scenery” (The realization is the hard part).
Go for a walk / go to a coffee shop / leave my home office and head to the “real” office / etc.
On the way there, mull over the problem. Or don’t bother, and let my subconscious do that for me.
Arrive at my destination, unpack my laptop, and find out that I’ve solved the problem or, at the very least, have come up with a new way to look at solving the problem.
Happily code away until I get stuck again, whereupon I go for another walk.

Luckily, I live close to a lot of coffee shops. Also, our new office is about a 25 minute walk from home, and a lot of those coffee shops are en route.

This means that I can start working at home (or, preferably, at a coffee shop. That way I don’t spend half an hour surfing the web in the morning). Then, when I hit that inevitable wall, I continue my commute and arrive at the office with a fresh solution to whatever problem I’ve been working on.

Here are a few more thought on this in no particular order.

The hard part really is coming to the realization that I’m banging my head against the wall. I’ve started listening more closely to those little voices in the back of my head saying “There must be an easier way to do this” or “Do you really need this feature?” or “Sheesh, I need a break.” Once I started doing this, I spent a lot less time going down rat holes.
Part of the problem is that I tend to dive into the details really quickly and don’t spend long enough looking at the big picture. I’m working on this. I’d suspect that most (not all, but most) programmers do this too.
It’s a heckuva lot more productive to go for a 10 minute walk than to bang your head against a wall for an hour.
This would be tough to do in a typical office environment – I don’t think I would’ve been able to go out for a walk and a coffee a few times a day while I was working at Creo (and Creo was far from being a bureaucratic, top-down organization).
Pair programming or describing the problem to someone or Rubber Ducking also get you out of those ruts. You don’t get those nice walks, though. Also, you may break someone out of their zone if you interrupt them to describe the problem to them…
Surfing the web does not give you the same results as going for a walk. You absolutely have to do something active.
Jamis Buck recommends getting a standing desk to help you walk away from the desk more easily. I may have to try this.

I’d love to hear any thoughts you have on the zone, whether or not it’s a myth, and how you stop yourself from banging your head against the wall.

Custom Validators with Validatious

2010-08-10T00:00:00-07:00

When I talked about Validatious previously, I had written almost no Javascript and had no need for any validations besides making sure that fields were required and of a certain minimum length.

On another project, I needed to validate some North American phone numbers and Canadian postal codes. It turned out to be really simple to add custom validations. Here’s how it’s done.

First, here’s a form:

<form id="test_form" class="validate">
  <div class="formset">
  <label for="address">Address</label><br />
  <input type="text" id="address" name="address" class="required" />
  </div>
  <div class="formset">
    <label for="postal_code">Postal Code</label><br />
    <input type="text" id="postal_code" name="postal_code" class="required cndpostalcode" />
  </div>
  <div class="formset">
    <label for="phone_number">Phone Number</label><br />
    <input type="text" id="phone_number" name="phone_number" class="required phone" />
  </div>
</form>

Note that I’ve added a class of cdnpostalcode to the postal code input and a class of phone to the phone number input. You still need the required class to turn on the validation.

This won’t do anything other than require that the inputs be filled out yet, as we haven’t added our custom validators in. To register a validator, you do something like this:

v2.Validator.reg('validatorClassName', function(field, value, params) { 
   return true /* some call back that returns true or false */ 
})

Most of the time, the callback is a regular expression. Let’s tackle the Canadian Postal Code one first, as they’re a bit simpler.

A Canadian Postal Code looks like A1A 1A1 or S7J 3A2 or H0H 0H0 ( bet you didn’t know that Santa lived in Canada! ). It’s letter-number-letter space number-letter-number. The space is optional, too. So, a regular expression to match a Canadian Postal Code might look like this:

/^\w\d\w\s*\d\w\s$/

We’ll register the class cdnpostalcode to trigger this validation like this:

v2.Validator.reg('cdnpostalcode', function(field, value, params) {
  return /^\w\d\w\s*\d\w\d$/.test(value)
})

The phone number validator is exactly the same, except the regular expression is a bit more hairy. This regex is from regexlib.com (I added in the part that accepts an extension).

v2.Validator.reg('phone', function(field, value, params) {
  return /^([0-9]( |-)?)?(\(?[0-9]{3}\)?|[0-9]{3})( |-)?([0-9]{3}( |-)?[0-9]{4}|[a-zA-Z0-9]{7})[ x0-9]*$/.test(value)
})

If we add those two chunks of Javascript to the HTML that calls our form, then they will validate the postal code and phone number correctly. The only problem is that the validation error message will be something like “Postal Code does not pass cdnpostalcode validator”. Not very pretty. There are a few more options to Validator.reg that will allow us to fix that.

The signature for Validator.reg looks like this (it’s in src/core/validator.js):

reg: function(name, fn, params, message, aliases, acceptEmpty) {
}

So far, we’ve only used the first two parameters (name and fn). The params argument would allow us to specify the parameters used in the validator. We don’t need that, so we’ll pass in null. Then, we can set the default message. Like this:

var postal_code_message = "You must provide a valid Canadian Postal Code"
v2.Validator.reg('cdnpostalcode', function(field, value, params) {
  return /^\w\d\w\s*\d\w\d$/.test(value)
}, null, postal_code_message)

similarly for the phone validator:

var phone_message = "You must provide a valid phone number"
v2.Validator.reg('phone', function(field, value, params) {
  return /^([0-9]( |-)?)?(\(?[0-9]{3}\)?|[0-9]{3})( |-)?([0-9]{3}( |-)?[0-9]{4}|[a-zA-Z0-9]{7})[ x0-9]*$/.test(value)
}, null, phone_message)

Here’s what the full HTML looks like. You can see this form in action here. If you’re intrigued, make sure to go check out Validatious.

<html>
  <head>
    <title>Demo: Custom Validations with Validatious</title>
    <style type="text/css">
      ul.messages {
        list-style-type: none;
        padding: 0;
        margin: 15px 0 0 0;
      }

      .error ul.messages li {
        color: red;
        padding: 0;
        margin: 0;
      }
    </style>

  </head>
  <body>
    <h1>Demo: Custom Validations with Validatious</h1>
    <form id="test_form" class="validate">
      <div class="formset">
      <label for="address">Address</label><br />
      <input type="text" id="address" name="address" class="required" />
      </div>
      <div class="formset">
        <label for="postal_code">Postal Code</label><br />
        <input type="text" id="postal_code" name="postal_code" class="required cdnpostalcode" />
      </div>
      <div class="formset">
        <label for="phone_number">Phone Number</label><br />
        <input type="text" id="phone_number" name="phone_number" class="required phone" />
      </div>
    </form>
    <a href="/2010/08/custom-validators-with-validatious.html">Back to the blog post</a>

    <script src="/demos/javascripts/validatious.min.js" type="text/javascript"></script>
    <script type="text/javascript">
      v2.Field.prototype.instant = true; // make the validations happen instantly

      var postal_code_message = "You must provide a valid Canadian Postal Code"
      v2.Validator.reg('cdnpostalcode', function(field, value, params) {
        return /^\w\d\w\s*\d\w\d$/.test(value)
      }, null, postal_code_message)

      var phone_message = "You must provide a valid phone number"
      v2.Validator.reg('phone', function(field, value, params) {
        return /^([0-9]( |-)?)?(\(?[0-9]{3}\)?|[0-9]{3})( |-)?([0-9]{3}( |-)?[0-9]{4}|[a-zA-Z0-9]{7})[ x0-9]*$/.test(value)
      }, null, phone_message)
    </script>

  </body>
</html>

Validating with Validatious: A Quick Tutorial

2010-08-03T00:00:00-07:00

I had a need for some client side validation for a project, so I looked into a few validation plugins. We’re using JQuery on the project (of course!), so I looked into the standard jquery validation plugins (jquery.validate looked nice, and so did jQuery Validation).

In the end, though, I ended up going with the totally unobtrusive Validatious.

I’m totally happy with my choice, but I ran into a gotchas and figured out a few things that I thought I should mention.

First, though, a quick overview of Validatious.

A quick intro to Validatious

Validatious has two different APIs: a Javascript DSL and an HTML extension. I used the HTML extension.

To use it, you load the script with a standard script tag

<script src="/javascripts/validatious.min.js" type="text/javascript"></script>

Then, you add some class names to elements. For example, to validate a form that looks like this:

<form id='donation_form' action='/donate'>
  <label for='first_name'>First Name</label>
  <input type='text' name='first_name' id='first_name' />
  <label for='last_name'>Last Name</label>
  <input type='text' name='last_name' id='last_name' />
  <label for='amount'>Amount to donate</label>
  <input type='text' name='amount' id='amount' />
</form>

You need to add a class of validate to the form element, and a validation class to each of the form elements that you want to validate. Validation classes are things like required (this element is required) or min-val_3 (this element has a minimum value of 3). Your final file might look something like this:

<html>
  <head>
    <title>Validatious Demo</title>
  </head>
  <body>
    <h1>Validatious Demo</h1>

    <form id='donation_form' action='#' class='validate'>
      <label for='first_name'>First Name</label>
      <input type='text' name='first_name' id='first_name' class='required' /><br />
      <label for='last_name'>Last Name</label>
      <input type='text' name='last_name' id='last_name' class='required'/><br />
      <label for='amount'>Amount to donate</label>
      <input type='text' name='amount' id='amount' class='required min-val_3' title='The donation amount must be at least $3'/><br />
      <input type='submit' value='submit!' />
    </form>
    <script src="/demos/javascripts/validatious.min.js" type="text/javascript"></script>
  </body>
</html>

All I’ve done is:

include the javascript.
add a class of validate to the form element.
add a class of required to the inputs.
add a class of min-val_3 to the final input (I want a donation of at least $3 or it’s not worth my time!).
add a title to the donation input to set a custom validation error.

Notice what I didn’t do: write any Javascript. I like that :).

You can see the form in action. Note that you have to click on submit to see the validation. Also note that the form and the validation errors are really ugly . We’ll fix all of that before the end of this.

There’s lots more you can do here. A reasonable place to start learning is on the Validatious features page, which contains a bunch of mini tutorials to get you started.

The Gotcha

I mentioned a gotcha. The problem, which almost made me give up on this plugin right away, is that

if you try to validate a form element with no name or id, it blows up with a cryptic error message like this:

     Uncaught Error: SYNTAX_ERR: DOM Exception 12

If you see this, go through and make sure that all of your form elements have ids and names.

Go see it in in-action.

Validating on blur

In the original demo, you had to click the submit button to see the validations. There’s been some research done that shows that validating on blur (just after the user has entered the value into the input box) is more effective. Luckily, that’s easy enough to do. Just add a script like this to your form:

v2.Field.prototype.instant = true; // make the validations happen instantly

Putting validations closer to the input field

One thing I didn’t like was that the validation errors were all getting piled up on the top of the form. It turns out that Validatious inserts the errors at the beginning of the parent of the input element that has an error. So, the fix is to wrap all of your input elements in a div tag. Not perfect, I know, but it’s reasonable if you give it a class name like formset and use it to hook some CSS in as well.

That’s semantic, right? It works, and that’s semantic enough for me. Don’t tell the standardistas.

Here’s what that looks like:

<form id='donation_form' action='#' class='validate'>
  <div class='formset'>
     <label for='first_name'>First Name</label>
     <input type='text' name='first_name' id='first_name' class='required' /><br />
  </div>
  <div class='formset'>
    <label for='last_name'>Last Name</label>
    <input type='text' name='last_name' id='last_name' class='required'/><br />
  </div>
  <div class='formset'>
    <label for='amount'>Amount to donate</label>
    <input type='text' name='amount' id='amount' class='required min-val_3' /><br />
  </div>
  <input type='submit' value='submit!' />
</form>

Validating via Javascript

As per usual, I wanted to do something a little different with my form. It was a very complex form, and I needed to do a few things before submitting. So, I needed to intercept the click on the submit button. Also, I wanted to style the button as an image with hover and active states, which is easier to do with an anchor tag than an submit input.

This all meant that I needed to trigger the validations from Javascript. You do this with a Javascript line like this:

var validate = v2.Form.forms['form_id'].validate()

replacing form_id with the id of the form you are validating. The validate variable will be true if the form is valid, false otherwise.

In the end, that’s all of the JavaScript I had to use. Everything else was as simple as adding a class to the form and then to all of the form elements I wanted to validate. I’m definitely using Validatious next time I need some client side validation.

Bringing it home

Here’s a final version of the form, with the formset divs, validation on blur, validating via Javascript and even a bit of styling:

<html>
  <head>
    <title>Validatious Demo</title>
    <style type="text/css">
      ul.messages {
        list-style-type: none;
        padding: 0;
        margin: 15px 0 0 0;
      }

      .error ul.messages li {
        color: red;
        padding: 0;
        margin: 0;
      }
    </style>
  </head>
  <body>
    <h1>Validatious Demo</h1>

    <form id='donation_form' action='#' class='validate'>
      <div class='formset'>
        <label for='first_name'>First Name</label>
        <input type='text' name='first_name' id='first_name' class='required' /><br />
      </div>
      <div class='formset'>
        <label for='last_name'>Last Name</label>
        <input type='text' name='last_name' id='last_name' class='required'/><br />
      </div>
      <div class='formset'>
        <label for='amount'>Amount to donate</label>
        <input type='text' name='amount' id='amount' class='min-val_3' /><br />
      </div>
      <a href='#' id='submitter'>submit!</a>
    </form>
    <script src="/demos/javascripts/validatious.min.js" type="text/javascript"></script>
    <script src="/demos/javascripts/jquery.js" type="text/javascript"></script>
    <script type="text/javascript">
      v2.Field.prototype.instant = true; // make the validations happen instantly
      $('#submitter').click(function() {
        var validate = v2.Form.forms['donation_form'].validate()
        // if validate is true then submit via AJAX or something ...
      })
    </script>
  </body>
</html>

You can see it in action here.

If you’re interested, go check Validatious out.

Update: I’ve also written a post on writing custom validators with Validatious.

My Git Aliases

2010-07-27T00:00:00-07:00

After you’ve been using Git for a while, you start to get kind of bored with typing git commit -a -m "some message" all the time, and you never want to see this message ever again:

  $> git stastus
  git: 'stastus' is not a git-command. See 'git --help'.
  
  Did you mean this?
        status

(why can’t you just do it for me!)

Ken had set up some nice aliases for git on his machine, so I totally stole them. My only regret is that it took me so long to do it. (I am glad that I didn’t do this right when I started using Git, though – it was worth while getting the long commands deeply ingrained into my memory before I started taking short cuts).

Here’s what I came up with:

   alias gpom='git push origin master'
   alias st='git status'
   alias gppp='git pull peter master'
   alias gcam='git commit -a -m '
   alias gd='git diff '

Obviously you don’t want the git pull peter master one, but the rest are probably close to something you use all the time.

I also wanted one that replaces

  git commit some_file some_other_file -m "commit message"

with something like

 gcom some_file some_other_file "commit message"

but my bash-fu wasn’t strong enough, so I resorted to a quick Ruby script instead:

#!/usr/bin/env ruby
USAGE = 'gcom <file 1> <file 2> "commit message" ==> git commit <file 1> <file 2> -m "commit message"'
(puts USAGE; exit) if(ARGV.length == 0)
message = ARGV.pop
exec "git commit #{ARGV.join(' ')} -m \"#{message}\""

These are still a work in progress, and there’s nothing earth shattering here, but I thought I’d throw up a post just for the heck of it.

How scottpatten.ca came to be

2010-07-27T00:00:00-07:00

I was happily working away on Leanpub one day when I got a phone call. It was some guy from the city of Vancouver asking why I hadn’t paid for my business license for spatten design.

I explained that I was no longer running spatten design as a business, so there was obviously no need to pay the $200 per year fee. Seemed pretty straight forward to me.

Unfortunately, he wasn’t buying any of that. He had gone all CSI on me and done a Google search for spatten design and found a (gasp!) running web site at http://spattendesign.com. There was a website running called spattendesign.com, therefore I was still in business.

Q E fucking D.

He was quite proud of his cyber sleuthing skills, and didn’t want to hear that the fact that the website was running had no bearing on whether or not I was running it as a business.

The front page of the site looked like this at the time:

Note a few things: 1. It says I work full time at Ruboss 2. The latest blog post was on April 20th. That’s April 20th, 2009. 3. There’s no mention that I do any consulting work as spatten design (and I don’t! Trust me, running a startup and a consultancy is enough work for me.)

I tried to explain that there were useful things on that site that people still used (mostly the S3 synch plugin), but he didn’t really care. The only thing I could do to avoid paying $200/year for a business license for a business that does not exist was take down the web site.

After a long and rather unfruitful discussion, I basically hung up on him.

So, here we are. I’ve made spattendesign.com redirect to scottpatten.ca. Hopefully that’s close enough to taking down the site that I don’t get another call from the Googler extraordinaire from the City of Vancouver.

On the plus side, I finally registered scottpatten.ca, and I had a great time getting this site up and running with Jekyll, and got rid of the old blog, which had been running on an old and creaky install of Mephisto. Also, Jekyll makes me actually want to blog (like a hacker).

Don’t worry, this is as close as I get to ranting. Your very irregularly scheduled technical posts will be coming back soon.

The S3 Cookbook

2009-04-20T00:00:00-07:00

One of the many things I’ve been doing while ignoring this blog has been writing a book.

It’s called The S3 Cookbook, and I’m more than happy to announce that it’s available for sale on Sopobo at http://leanpub.com/thes3cookbook.

Here’s a bit about the book:

Amazon’s Simple Storage Service (S3) has been described as “Storage in the cloud”. I like to think of it as Amazon lending you some hard-drive space on their servers, and giving you lots of interesting ways to put data there and share it with others.

Getting started with S3 is pretty straightforward, but even with its simple interface (really, because of its simplicity), there are a lot of things you can do with it. The S3 Cookbook will give you a quick introduction to S3’s architecture, get you set up with some tools, and then let you explore.

The recipes provide you with the code to do all of the everyday things that people are doing with S3, as well as the more esoteric ones. They will help you get the simple things done quickly, and inspire you with ideas of your own.

You will learn to

get started with S3: set up your account, get some tools and get coding
upload and download data from S3
list all of the files you have in your S3 Buckets
add metadata to your objects
serve your objects as bit torrents
share your data with others (everyone, just a chosen few or nobody at all)
backup and restore your pictures, databases or any directory on any computer to S3
log who accesses your objects and how often
serve assets such as images, MP3s or the static files on your website

S3 is built for people like you and me: programmers. If you feel the itch to create your own library to talk to S3, then chapters 4 and 5 are for you. Chapter 4 walks you through the S3 authentication process, and builds an S3 authentication library using Test Driven Development. Chapter 5 builds the beginnings of an S3 library, giving you everything you need to build one of your own, or take the code that’s in the book and go from there.

For a complete list of the recipes, see the table of contents

Interested? Download a sample of the book.

All of the code in the book is MIT licensed and available for download.

NOTE:

The S3 Cookbook is still in beta. This means that:

the book is not done
Things are still a little rough, but they will improve
There will be more recipes coming
I need your feedback!

When you buy the book, you get updates for free for the lifetime of the book.

I'll be speaking at RailsConf

2009-04-20T00:00:00-07:00

I’m doing a “Birds of a Feather” session at RailsConf called Amazon S3 and Rails – Take the First Step into the Cloud.

The talk will be 8:00pm on Wednesday, May 6th in Conference Room 10

Here’s the blurb about the talk:

Learn how to make your apps faster, more scalable, bullet-proof and just plain cooler with Scott Patten, author of The S3 Cookbook.

Scott Patten will get you set up to use S3 and get you creating and downloading objects using s3sh, and give a whirlwind tour of its architecture and features.

We’ll then talk about S3 and its uses for web applications: serving user generated data and assets, backing up your app, authenticated downloads and whatever other crazy things you are using it for. Then we’ll open the floor to a discussion of S3 tips, tricks and pitfalls.

Hope to see you there!

Searching Beast and WordPress from a Rails app

2008-05-17T00:00:00-07:00

I did some work for ecolect recently, integrating search across their main site, a Beast forum and a WordPress blog. It was pretty straightforward once I had it figured out, but I couldn’t find a walkthrough on the net.

So, I decided to write one.

The Search Engine

Ecolect was already using acts_as_ferret for their main site search, so it was a no-brainer to keep using it. For a fun and thorough introduction to acts_as_ferret, see the Rails Envy Tutorial

Searching the Beast Forum

(Note: the beast Forum search isn’t live at the moment as the Forums are not fully functioning yet.)

Beast setup

In Beast, each Forum has many Topics (saved in the topics table), and each Topic has many Posts (saved in the posts table).

The Beast forum was set up by Shanti Braford as described in option #3 of this great article on integrating a beast forum into a Rails app.

In this setup, the Beast tables are added to the main site’s database. So, the posts and topics tables that we want to search are already in the main site’s database. This made things pretty easy: you can search the Beast forum just like you would any database table.

Searching a non-integrated Beast forum

If you don’t have the Beast tables integrated in to your main site’s database, you can still search them. You just need to point the Topic and Post models to the correct database. This is a two step process.

First, set up a database entry for your Beast forum in config/database.yml. Something like this:

beast:
  adapter: mysql
  database: beast_forum
  username: app
  password: your_password
  host: localhost

Then, in the bottom of config/environment.rb, add the following lines:

Post.establish_connection "beast_forum"
Topic.establish_connection "beast_forum"

I haven’t actually tried this, so let me know if you get it working or if you needed to make any changes to what I’ve written here.

The Topic Model

Even with the integrated setup there were a few wrinkles. First, although the Beast tables are in the database, there are no models associated with them. I wanted to search post body and topic titles, so I created Topic and Post models.

Here’s the Topic model. It’s only here so that the Post model can search a posts’s titles, so there’s not much to it.

class Topic < ActiveRecord::Base
  
  has_many :posts
    
end

The Post Model

The Post model is a bit more complicated.

class Post < ActiveRecord::Base
  
  belongs_to :topic
  
  SEARCH_FIELDS = [:scrubbed_body, :topic_title]
  
  acts_as_ferret :fields => { :scrubbed_body => {:boost => 0, :store => :yes}, 
                              :topic_title => {:boost => 3, :store => :yes}}  
  
  extend FullTextSearch
  
  # remove the html tags from body
  def scrubbed_body
    body_html.gsub(/<\/?[^>]*>/, "")
  end
  
  def topic_title
    topic.title
  end
      
  # Construct the url to the post    
  def url
    File.join("http://forums.ecolect.net", "forums", topic.forum_id.to_s, 
              "topics", topic.id.to_s + "#post_#{id}")
  end
  
end

acts_as_ferret

The acts_as_ferret declaration makes the Post model searchable. Notice that the actual fields being searched are not taken directly from the database; they are both manipulated in some way. acts_as_ferret doesn’t really care if the stuff it is indexing is coming directly from the database or from methods you have added to your model.

scrubbing the html tags

The post body is stored with HTML tags in them, so I wanted to search and show the posts with tags scrubbed out of them. This is done using the Post#scrubbed_body method, which is just an ugly regexp that takes out anything between < and > signs.

the url method

I also wanted to link to the posts, so I created a Post#url method which is used in the view.

Finally, the actual search is done using the FullTextSearch mixin, which adds a class method Post::full_text_search to Post. The FullTextSearch mixin is described in more detail below.

Searching the WordPress Blog

The only table from the WordPress Blog that you really care about is the wp_post table. To get access to it in your Rails app, make a WpPost model and point it at your WordPress db.

First, create a database entry in config/database.yml that looks like this:

wordpress:
  adapter: mysql
  database: wordpress
  username: app
  password: your_password
  host: localhost

Then, add the following line to the bottom of config/environment.rb

WpPost.establish_connection "wordpress"

Here’s the model:

class WpPost < ActiveRecord::Base
  
  primary_key = "ID"
  
  SEARCH_FIELDS = [:scrubbed_title, :scrubbed_content]
                   
  acts_as_ferret :fields => { :scrubbed_title => {:boost => 3, :store => :yes},
                              :scrubbed_content => {:boost => 0, :store => :yes}}
  
  extend FullTextSearch
  
  def id
    read_attribute(:ID)
  end
  
  # title and content need to have the html tags removed from them, and should
  # only be searchable if the post has been assigned a url (the guid) and the post 
  # is actually a post and not an asset.
  def scrubbed_title
   post_title.gsub(/<\/?[^>]*>/, "") if post_type == "post" and !guid.empty?
  end
  
  def scrubbed_content
    post_content.gsub(/<\/?[^>]*>/, "") if post_type == "post" and !guid.empty?
  end
end

There are a few things to note here:

WordPress uses ID, rather than id, as its primary key. The line primary_key = "ID" lets Rails know about that. You also need to add an id method that returns ID to get ferret indexing things properly.

You will need to scrub the html tags from the content and title; that’s what the scrubbed_title and scrubbed_content methods do.

Finally, you don’t want search results to index assets (which are stored in the wp_post model as well) or any un-published posts.

Real posts will have a post_type of 'post'.
An unpublished post won’t have its guid set.

This is taken care of by only returning titles or content if post_type == "post" and !guid.empty?.

The FullTextSearch mixin

This is based on code by Roman Mackovcak’s article on full text search in Rails. All I did was extract the method he provides out in to a mixin so I could use it in multiple models.

To use the mixin in a model, the model needs to define SEARCH_FIELDS and have an acts_as_ferret declaration. SEARCH_FIELDS is an array of symbols giving the model fields to be searched.

module FullTextSearch
  
  ##
  # FERRET SEARCH METHOD
  # This method requires that you set the following in the model:
  # SEARCH_FIELDS: 
  # a list of symbols giving the fields to be searched.
  # E.g., SEARCH_FIELDS = [:post_title, :post_content]
  #
  # The acts_as_ferret declaration.  
  # Use :store => :yes for each field if you want to use highlighting for that field.
  # E.g., 
  # acts_as_ferret :fields => { :post_title => {:boost => 3, :store => :yes},
  #                             :post_content => {:boost => 0, :store => :yes}}
  ##
  DEFAULT_PER_PAGE = 10
  
  def full_text_search(q, options = {})
     return nil if q.nil? || q == ""
     
     default_options = {:limit => FullTextSearch::DEFAULT_PER_PAGE, 
                        :page => 1, 
                        :lazy => self.const_get(:SEARCH_FIELDS)
                       }
     options = default_options.merge options

     # Get the offset based on what page we're on
     options[:offset] = options[:limit] * (options.delete(:page).to_i-1)  
     # Now do the query with our options
     results = self.find_by_contents(q, options)

     return [results.total_hits, results]
  end  
  
  
end

You use it like this:

WpPost.full_text_search('test')
WpPost.full_text_search('test', :limit => 30, :page => 2)

The full_text_search method returns an array of length two. The first value in the array is the number of search results, and the second value the actual search results.

Re-indexing the non-local models

As Ruben pointed out in the comments, I forgot to mention how I deal with re-indexing the WordPress and Beast database tables. This is necessary as these tables have data that is modified by another application, so your Rails app doesn’t know that changes have been made to them.

To deal with this, I wrote a simple Rake task that reindexes the WpPost, WpArticle and Post models, and then added a line to the crontab to run it hourly. Here’s the rake file, which I put in RAILS_ROOT/lib/tasks/ferret.rake:

namespace :ferret do
  
  desc "rebuild the non-local Ferret indices, as these need to be done manually."
  task :rebuild_nonlocal_indices => :environment do
    [WpPost, WpArticle, Post].each do |model|
      puts "rebuilding #{model} index"
      model.rebuild_index
    end    
  end
  
end

and, for posterity, here’s the crontab line:

20 * * * *  cd <rails_root> && /usr/local/bin/rake ferret:rebuild_nonlocal_indices >> <log_directory>/ferret_reindex.log  2>&1

Rubyize this: Live in Vancouver. Refactoring #1.

2008-01-27T00:00:00-08:00

At rubycamp today, I ran a little Rubyize This workshop. Rubyize This is a game invented by François Lamontagne. The idea is that someone puts up a chunk of code that is written in Ruby, but in a not very Rubyish way. Then, everyone in the audience gets to Rubyize it!

We tried to use the most excellent RefactorMyCode.com, but the WiFi was too saturated so we had to resort to posting the code on my blog and people plugging their laptops in to the projector to present their results.

So, this post and the following two will show the original code. Keep in mind that this is intentionally ugly!

We had some great refactorings. If you were at the workshop, please post your refactorings in the comments. I’ll make the pretty and put them in the posting.

The other two refactorings are at Rubyize this #2 and Rubyize this #3.

Without further ado, here’s the first refactoring. The code is supposed to create a file filled with random numbers, one number per line.

#!/usr/bin/env ruby

# create num random numbers between min and max 
# and print them to a file called file_name, one per line
def create_random_numbers(file_name, min, max, num)
  file = File.open(file_name, 'w')
  n = 0
  while n < num
    r = rand(max - min) + min
    file.puts(r)
    n += 1
  end
  file.close   
end

create_random_numbers('random.txt', 0, 10, 1000)

Here’s the first refactoring:

def create_random_numbers(file_name, min, max, num) 
  File.open(file_name, 'w') do |f| 
    1.upto(num) {|i| f.puts rand(max-min)+min }
  end 
end

Here’s one from Sam Livingston-Gray

#!/usr/bin/env ruby

# create num random numbers between min and max 
# and print them to a file called file_name, one per line

class Range
  def random_member
    offset = rand(max - min)
    min + offset
  end
end

def create_random_numbers(file_name, range, num)
  File.open(file_name, 'w') do |file|
    num.times { file.puts(range.random_member) }
  end
end

create_random_numbers('random.txt', (0..10), 1000)

Here’s what I came up with:

def create_random_numbers(file_name, min, max, num)
  File.open(file_name, 'w') do |file|
    num.times {file.puts(rand(max - min) + min)}
  end
end

create_random_numbers('random.txt', 0, 10, 1000)

Rubyize This: Live in Vancouver. Refactoring #3

2008-01-27T00:00:00-08:00

Here’s the final refactoring from the Rubyize This workshop. See the first refactoring for an explanation of what’s going on and why this code is so darn ugly! The second refactoring is worth checking out as well.

#!/usr/bin/env ruby

def delete_large_files(directory, max_size)
  
  # Make sure directory ends in a slash
  if directory !~ /\/$/
    directory += '/'
  end
  
  # Find all of the files in the directory 
  files = Dir.glob(directory + '*')
  
  # Delete all files with size larger than max_size
  for file in files
    size = File.size(file)
    if size >= max_size
      File.delete(file)
    end
  end
  
end

max_size = 1024 * 50 # 50 kb
directory = './files_to_delete'
delete_large_files(directory, max_size)

This was the final refactoring, and the end of the conference, so we sort of ran out of time.

I didn’t get a change to show off my solution. Here it is.

def delete_large_files(dir, max_size)

  files_to_delete = Dir.glob(File.join(dir, '*')).select do |file|
    File.size(file) > max_size
  end

  File.delete(*files_to_delete)

end

What’s going on with that File.delete call?

First, File.delete takes multiple arguments and deletes all of them.

Second, I used the asterisk operator. From here:

“The asterisk operator may also precede an Array argument in a method call. In this case the Array will be expanded and the values passed in as if they were separated by commas.”

I don’t think I’ve every actually used the asterisk operator in production code, but it sure came in handy here.

Rubyize This: Live in Vancouver. Refactoring #2

2008-01-27T00:00:00-08:00

Here’s the second refactoring from the Rubyize This workshop. See the first refactoring for an explanation of what’s going on and why this code is so darn ugly! Don’t forget to check out the third and final refactoring

This script loads the file full of random numbers from the first refactoring and makes a beautiful ascii-art histogram from it.

#!/usr/bin/env ruby

require 'pp'

BIG_NUMBER = 32768

# Get data from a file, turn it in to a float, and find the max
data = File.readlines('random.txt')
max = -BIG_NUMBER
for n in (0 ... data.length)
  data[n] = data[n].chomp.to_f
  if data[n] > max
    max = data[n].to_i + 1 # max is ceil(max(data[n]))
  end
end

# Create the empty histogram
histogram = []
for n in (0 .. max)
  histogram.push(0)
end

# Fill the histogram
for n in (0 ... data.length)
  histogram[data[n].to_i] += 1
end

# Print the histogram
pp histogram
puts
for n in (0 .. max)
  puts "*" * histogram[n]
end

Here’s a very concise one-liner from the crowd:

0.upto((data=File.readlines('random.txt').collect {|e| e.chomp!.to_f}).max.to_i) {|i| puts i.to_s + " " + data.select{|a| a==i}.size.to_s}

Here’s Owen’s refactoring:

require 'pp' 
histogram = [] 
File.readlines('random.txt').each do |value| 
  i = value.chomp.to_i 
  histogram[i] ||= 0 
  histogram[i] += 1 
end 

# Print the histogram 
pp histogram 
puts 
histogram.each { | v | puts "*" * v }

Here’s Sam Livingstone Gray’s refactoring

#!/usr/bin/env ruby

require 'pp'

# Get data from a file, turn it in to a float, and find the max

module Enumerable
  def value_counts
    h = Hash.new(0)
    each { |e| h[e] += 1 } 
    h
  end
end

lines = File.readlines('random.txt')
histogram = lines.map { |line| line.chomp.to_i }.value_counts

# Print the histogram
pp histogram
puts
histogram.keys.sort.each { |n| puts '*' * histogram[n] }

I really like this one. Sam created a simple extension to Enumerable that I can see using over and over again.

Here’s what I came up with, with some debugging help from the group

#!/usr/bin/env ruby

require 'pp'

data = File.readlines('random.txt').collect {|datum| datum.chomp.to_f}
max = data.max

histogram = data.inject([0] * (max + 1)) do |histogram, datum|
  histogram[datum.to_i] += 1
  histogram
end

pp histogram
histogram.each do |bucket|
  puts "*" * bucket
end

Adding json responses to all of your controller actions

2008-01-14T00:00:00-08:00

Gerald has decided to use JSON responses for the mobile phone app
we’ll be building at the intermediate Rails
workshop.
I’m a lazy guy by nature, so there was no way I was going to paste in

format.json  { render :json => @bookmark }

in to very controller method.

Now, the easiest way to do this would be to add JSON response to the scaffolding. The templates are in RAILS_ROOT/vendor/rails/lib/rails_generator/generators/components/scaffold/. Open up the controller.rb template and add in the JSON response shown above to the index, show, update and destroy methods and you’re done.

Unfortunately, I had already mostly developed the API when I realized that I had to add in the JSON responses (It’s not Gerald’s fault, he told me long ago). So, I decided to write a quick rake task to do the job for me. To use it, create a file called add_json.rake in RAILS_ROOT/lib/tasks and copy the following code in to it:

namespace :rails do
  
  desc "Adds a 'format.json ...' line for each 'format.xml ...' line in every controller"
  task :add_json => :get_controller_list do
    @controllers.each do |controller|
      new_lines = File.open(controller, 'r').readlines.collect do |line|
        if line =~ /^\s+format\.xml/
          [line, line.gsub(/xml/, 'json')]
        else
          line
        end
      end.flatten
        
      file = File.open(controller, 'w') do |file|
        file.puts new_lines
      end
    end
  end
  
  task :get_controller_list => :environment do
    @controllers = Dir.glob(File.join(RAILS_ROOT, 'app', 'controllers', '*_controller.rb'))
  end
end

Call the task from RAILS_ROOT like this:
rake rails:add_json

The task opens up every controller, and every time it finds a line that starts with format.xml, it makes a copy of that line just below with every instance of ‘xml’ replaced with ‘json’. It’s ~~pretty~~ very brain-dead: it doesn’t check if the json response is already there, so don’t run it twice on the same project.

Fun with Single Table Inheritance

2007-12-31T00:00:00-08:00

I’m working on the sample application for the Intermediate Ruby on Rails Workshop that Gerald Bauer and I are putting on in January. I had to remind myself of some of the foibles of Single Table Inheritance (STI). I thought others might find this useful, so here’s what I found.

The Set-up

I have two types of posts in the application I’m building: posts by people with an empty room looking for a roomie (RoomiePosts), and posts by people looking for a room (RoomPosts). They both share a lot of the same attributes, so it’s not very DRY to create two separate models.

This is where Single Table Inheritance comes in. STI is basically a way of subclassing a single model, creating subclasses that all use the same database table.

What I did is create a Post model and two subclasses: RoomiePost and RoomPost. Here’s how you do it.

Creating the Database Table

Here’s the migration for the posts database table. Note that this uses the new migrations syntax in Rails 2.0.

class CreatePosts < ActiveRecord::Migration
  def self.up
    create_table :posts do |t|
      t.integer :rent
      t.date :start_date, :expiry_date
      t.boolean :includes_utilities
      t.string :text, :type
      
      t.timestamps
    end
  end

  def self.down
    drop_table :posts
  end
end

Most of the columns are pretty straightforward. text will contain the text of the post. start_date and expiry_date will determine when the post will start and stop being shown on the site. rent and includes_utilities will only be used by the RoomiePost subclass, but still need to be created for all subclasses of the Post model.

The interesting column is type. This allows Rails to determine which subclass the row should be loaded in to. If you set type = "RoomiePost", then it’s a RoomiePost. If you set type = "RoomPost", then it’s a RoomPost.

Creating the Models

You need to create a Post, RoomiePost and RoomPost model. They look like this:

In /app/models/post.rb

class Post < ActiveRecord::Base 
  
end

In /app/models/roomie_post.rb

class RoomiePost < Post
  
end

In /app/models/room_post.rb

class RoomPost < Post
  
end

Note that RoomiePost and RoomPost are both subclasses of the Post model. Here’s where you run in to a “gotcha”.

You have to create the models in separate files. If you don’t, Rails ignores the subclassed models.

For more information, and another workaround, see this blog post. The comment by Chris at the bottom gives the ‘create the models in separate files’ solution.

Okay, so now you have the models all set up. You can do things like

>> p = RoomiePost.create(:rent => 800, :includes_utilities => true, :text => "Great room in a shared house in the Commercial Drive area")

And it will create a row in the posts table with type = "RoomiePost". If you have 5 RoomiePosts and 2 RoomPosts, then Posts.count will return 7.

>> RoomiePost.count           
=> 5
>> RoomPost.count
=> 2
>> Post.count
=> 7

Adding behaviour to STI models

I want to be able to determine if a post is active by looking at the start and expiry dates, and to get a list of all active posts, roomie_posts and room_posts. I also want to have a different icon for each post type.

The Post.active_post? and Post::active_posts methods are straight-forward:

class Post < ActiveRecord::Base
  
  def is_active?
    (start_date .. expiry_date) === DateTime.now
  end 
  
  def self.active_posts
    self.find(:all).select {|p| p.is_active? }
  end  
  
end

The only tricky part is that I used self.find(:all) instead of Post.find(:all) in the Post::active_posts method. This makes sure that RoomiePost::active_posts only returns RoomiePosts and RoomPost::active_posts only returns RoomPosts.

Getting the icon to work took a little more figuring out. If I wanted to do this in straight-up Ruby, I’d do something like

#!/usr/bin/env ruby

class Post
  ICON_PATH = "icons"
  
  def initialize
    @icon_name = 'base.gif'
  end
  
  def icon
    File.join(ICON_PATH, @icon_name)
  end
  
end

class RoomiePost < Post
  
  def initialize
    @icon_name = 'roomie.gif'
  end
  
end

class RoomPost < Post
  
  def initialize
    @icon_name = 'room.gif'
  end
  
end

Then I could do this

ruby $>irb -r subclass_test
irb(main):001:0> Post.new.icon
=> "icons/base.gif"
irb(main):002:0> RoomiePost.new.icon
=> "icons/roomie.gif"
irb(main):003:0> RoomPost.new.icon
=> "icons/room.gif"

This doesn’t quite work in the Rails version. Not to worry, you only have to make one simple change:

replace the initialize method with an after_initialize callback.

The final post.rb, roomie_post.rb and room_post.rb files look like this:

post.rb:

class Post < ActiveRecord::Base

  ICON_PATH = "/images/icons"
  
  def after_initialize
    @icon_name = 'base.gif'
  end
  
  def icon
    File.join(ICON_PATH, @icon_name)
  end  
  
  def is_active?
    (start_date .. expiry_date) === DateTime.now
  end 
  
  def self.active_posts
    self.find(:all).select {|p| p.is_active? }
  end  
  
end

roomie_post.rb:

class RoomiePost < Post
  
  def after_initialize
    @icon_name = 'roomie.gif'
  end
  
end

room_post.rb:

class RoomPost < Post
  
  def after_initialize
    @icon_name = 'room.gif'
  end
  
end

One final note: because of the way the after_initialize callback works, you have to actually define an after_initialize method. You can’t do something like this:

class RoomiePost < Post
  after_initialize :set_icon_name

  private

  def set_icon_name
    @icon_name = 'roomie.gif'
  end
  
end

If you want to learn more about STI, and a whole slew of other info on Rails, sign up for the Intermediate Rails Workshop on January 25th, 2008 in Vancouver.

Early Bird rates are available until January 9th!

RubyCamp Vancouver

2007-12-19T00:00:00-08:00

RubyCamp is a free one-day gathering for Rubyists and Railsers.

When and Where

WorkSpace in downtown Vancouver, B.C., Canada
January 26th, 2008 from 9:00 to 5:00

Who should come

Anyone who’s interested in Ruby and Rails, whether you’re just interested in learning what this Ruby thing is all about or you know Ruby inside out.

The Conference Track

A conference-style track with “classic” talks on Ruby or Rails topics. We’re looking for a few more speakers, so see here if you’re interested in giving a talk.

The Hackathon Track

An informal un-conference track focusing on hacking some Ruby code, showing off a cool feature you just added to your Rails application or demonstrating a new addition to Rails 2.0.

If you’re interested in working on some code or showing off something, get started by promoting your ideas and getting some buzz going .

A weekend of Ruby

There are two more reasons to come to Vancouver for those of you who are out of town.

On the Friday before RubyCamp, Rails Advance is giving a one day intermediate Ruby and Rails workshop. (Rails Advance is Gerald Bauer and I, by the way)

On the Sunday after RubyCamp, Peter Armstrong is giving a one day workshop on Flex and Rails.

For more information, see the RubyCamp web site

See you all at RubyCamp in Vancouver!

Amazon S3, Ruby and Rails slides

2007-12-05T00:00:00-08:00

The slides from the talk are here. (Yes, they’re hosted on S3).

There are two points in the presentation where I switched to a different window.

At the ‘S3SH DEMO’ slide, I did some live coding showing how you can work with S3 using s3sh. It basically followed the script shown in ‘s3sh demo script’ below, so read that part when you see the ‘S3SH DEMO’ slide.

At the ‘Example: S3Syncer’ slide, I switched over to textmate and showed the code for a simple script to synchronize a single directory to S3. I then demoed the script to show it working. So, when you see the ‘Example; S3Syncer’ slide, read the s3syncer code and s3syncer demo sections below.

s3sh demo script

Start up s3sh



$> s3sh

Create a bucket.
Show that you can create a bucket multiple times if you own it, but trying to create a bucket that somebody else owns raises an error.

>> Bucket.create('spatten_s3demo')
=> true
>> Bucket.create('spatten_s3demo')
=> true
>> Bucket.create('test')
AWS::S3::BucketAlreadyExists: The requested bucket name is not available. The bucket namespace is shared by all users of the system. Please select a different name and try again.
        from /usr/local/lib/ruby/gems/1.8/gems/aws-s3-0.4.0/bin/../lib/aws/s3/error.rb:38:in `raise'
        from /usr/local/lib/ruby/gems/1.8/gems/aws-s3-0.4.0/bin/../lib/aws/s3/base.rb:72:in `request'
        from /usr/local/lib/ruby/gems/1.8/gems/aws-s3-0.4.0/bin/../lib/aws/s3/base.rb:83:in `put'
        from /usr/local/lib/ruby/gems/1.8/gems/aws-s3-0.4.0/bin/../lib/aws/s3/bucket.rb:79:in `create'
        from (irb):3

You can save a bucket in a variable using Bucket.find



>> b = Bucket.find(‘spatten_s3demo’)

=> #<AWS::S3::Bucket:0×14ae7b8 @attributes={"prefix"=>nil, “name”=>"spatten_s3demo", “marker”=>nil, “max_keys”=>1000, “is_truncated”=>false, “xmlns”=>"http://s3.amazonaws.com/doc/2006-03-01/"}, @object_cache=[]>

Create a text object



>> S3Object.store(‘test.txt’, ‘This is a test’, ‘spatten_s3demo’)

=> #

>> b.objects

=> [#<AWS::S3::S3Object:0×10804170 ‘/spatten_s3demo/test.txt’>]

>> pp b.objects⁰.about

{"last-modified"=>"Wed, 05 Dec 2007 19:56:49 GMT",

 “x-amz-id-2”=>

  “JACm9T+m9CgZhmj4q6q00OSGHgSyBVAbQ1cgRWGydYZLTKdhLc/IUZ+K7b/1snOc”,

 “content-type”=>"text/plain",

 “etag”=>"\“ce114e4501d2f4e2dcea3e17b546f339\”",

 “date”=>"Wed, 05 Dec 2007 19:57:03 GMT",

 “x-amz-request-id”=>"CA170D2AA5DEB0C9",

 “server”=>"AmazonS3",

 “content-length”=>"14"}

=> nil

>> b.objects⁰.key

=> “test.txt”

>> b.objects⁰.value

=> “This is a test”

Create a binary object and show it in a browser



>> S3Object.store(‘vampire.jpg’, File.open(‘vampire.jpg’), ‘spatten_s3demo’)

=> #

Show the photo in browser

This doesn’t work, as the file is only readable by me. Make it public readable and do it again.

>> S3Object.store('vampire.jpg', File.open('vampire.jpg'), 'spatten_s3demo', 
     :access => :public_read)
=> #<AWS::S3::S3Object::Response:0x10747950 200 OK>

Show it in a browser again. It works this time.

Look at bucket.objects. We have to reload the bucket to show the new object.

>> b.objects
=> [#<AWS::S3::S3Object:0x10804170 '/spatten_s3demo/test.txt'>]
>> b.objects(:reload)
=> [#<AWS::S3::S3Object:0x10708080 '/spatten_s3demo/test.txt'>, #<AWS::S3::S3Object:0x10708070 '/spatten_s3demo/vampire.jpg'>]

Hash access to bucket objects



>> b[‘vampire.jpg’]

=> #<AWS::S3::S3Object:0×10708070 ‘/spatten_s3demo/vampire.jpg’>

>> vamp = b[‘vampire.jpg’]

=> #<AWS::S3::S3Object:0×10708070 ‘/spatten_s3demo/vampire.jpg’>

A look at metadata

>> vamp.content_type
=> "image/jpeg"
>> vamp.size
=> 10817
>> vamp.metadata
=> {}
>> vamp.metadata['subject'] = 'Claire'
=> "Claire"
>> vamp.metadata['photographer'] = 'Nadine Inkster'
=> "Nadine Inkster"
>> vamp.store
=> true

Storing the picture data in a variable



>> picdata = vamp.value

=> "\377\330\377\340\000\020JFIF\000\001\002\000…….

Downloading a picture by streaming it to an IO object.

>> File.open('vampire_downloaded.jpg', 'w') {|file| file.write(vamp.value)}
=> 10817
>> exit
s3demo $>ls
flowers.jpg             vampire.jpg
test.txt                vampire_downloaded.jpg
s3demo $>open vampire_downloaded.jpg 
s3demo $>

S3Syncer Code

Please note that this code is really only useful as an example of how to synchronize with S3.

It won’t recurse directories and it dies a horrible death if there are any symlinked files in a directory.

If you are looking for something to synchronize directories, check out s3sync.rb.

#!/usr/bin/env ruby

require 'digest/md5'
require 'aws/s3'
include AWS::S3

class S3Syncer
  attr_reader :local_files, :files_to_upload
  
  def initialize(directory, bucket_name)
    @directory = directory
    @bucket_name = bucket_name
  end
  
  def S3Syncer.sync(directory, bucket)
    syncer = S3Syncer.new(directory, bucket)
    syncer.get_local_files
    syncer.connect_to_s3
    syncer.get_bucket
    syncer.select_files_to_upload
    syncer.sync
  end
  
  # This does not recurse directories.
  def get_local_files
    @local_files = Dir.entries(@directory)
  end
  
  def connect_to_s3
    Base.establish_connection!(
        :access_key_id     => ENV['AMAZON_ACCESS_KEY_ID'],
        :secret_access_key => ENV['AMAZON_SECRET_ACCESS_KEY']
      )
  
    raise "\nERROR: Connection not made or bad access key " +
          "or bad secret access key.  Exiting" unless AWS::S3::Base.connected? 
  end  
  
  def get_bucket
    Bucket.create(@bucket_name)
    @bucket = Bucket.find(@bucket_name) 
  end
  
  # Files should be uploaded if 
  #   The file doesn't exist in the bucket
  #      OR
  #   The MD5 hashes don't match
  def select_files_to_upload
    @files_to_upload = @local_files.select do |file|                 
      case
      when File.directory?(local_name(file))
         false # Don't upload directories
      when !@bucket[file]
         true  # Upload if file does not exist on S3
      when @bucket[file].etag != Digest::MD5.hexdigest(File.read(local_name(file)))
         true  # Upload if MD5 sums don't match
      else
        false  # the MD5 matches and it exists already, so don't upload it
      end
    end
  end
  
  # This will choke on symlinked files
  def sync
    (puts "Directories are in sync"; return) if @files_to_upload.empty?

    @files_to_upload.each do |file|
      puts "#{file} ===> #{@bucket.name}:#{file}"
      S3Object.store(file, File.open(local_name(file), 'r'), @bucket_name)      
    end
  end
  
  private 
  
  def local_name(file)
    File.join(@directory, file)
  end
  
end

if __FILE__ == $0
  S3Syncer.sync('/Users/Scott/versioned/spattendesign/presentations/s3-on-rails/s3demo', 
                'spatten_syncdemo')
end

S3Syncer demo

Start with spatten_syncdemo bucket empty, and four files in the local directory.

Run the script



s3demo $>ls

flowers.jpg             vampire.jpg

test.txt                vampire_downloaded.jpg

s3demo $>s3syncer

flowers.jpg => spatten_syncdemo:flowers.jpg
test.txt => spatten_syncdemo:test.txt

vampire.jpg => spatten_syncdemo:vampire.jpg
vampire_downloaded.jpg => spatten_syncdemo:vampire_downloaded.jpg

Run it again, it says there’s no need to do anything



s3demo $>s3syncer

Directories are in sync

Change a file locally and sync again



s3demo $> vi test.txt

Make some changes using vi

s3demo $>s3syncer

test.txt ===> spatten_syncdemo:test.txt

Delete flower.jpg using the Firefox S3 Organizer and then sync again.



s3demo $>s3syncer

flowers.jpg ===> spatten_syncdemo:flowers.jpg

So there you go, a quick intro to the wonders of Amazon S3.

synch s3 host plugin update

2007-12-04T00:00:00-08:00

I’ve re-written the plugin to use s3sync.rb rather than trying to figure out which files have changed in svn. It’s a much better solution, so hopefully those of you who were having trouble will find it works for you now. Thanks, especially, to Chris in the comments for helping me figure out what was going wrong.

The plug-in also now has a permanent home in my new Projects section, here.

Amazon S3, Ruby and Rails talk

2007-11-28T00:00:00-08:00

I’m going to be giving a talk on Amazon S3, Ruby and Rails at the next Vancouver Ruby and Rails meetup. It’s this Monday, December 3rd.

Here’s the talk teaser:

Amazon’s Simple Storage Service (Amazon S3) is an online storage system. It can be used for backup, serving assets for your Rails application, storing and streaming large media files or storing customer generated assets.

The AWS/S3 gem, written by Marcel Molina, provides an elegant interface to Amazon S3. I’ll be talking about how S3 works, how to use it in your Ruby scripts and Rails applications and showing off some example code.

Hope to see you there!

Synching Your Amazon S3 Asset Host using Capistrano

2007-11-06T00:00:00-08:00

Note: This article is out of date. The latest version of this article is on the new permanent page in the projects section.

So you’ve got multiple asset hosts running in your Rails application, and you’re using Amazon’s S3 to host your assets. Now you want to make sure that your assets are kept up to date. This plugin is a Capistrano recipe that keeps the asset hosts synchronized with the public directory in your subversion repository.

Usage

After you get everything setup and do your first deploy, just run cap deploy as normal and all changed files in RAILS_ROOT/public will be uploaded to all of your asset host buckets before the final deploy:symlink task.

The following tasks are also available:

cap s3_asset_host:get_s3_revision
cap s3_asset_host:find_changed
cap s3_asset_host:list_changed
cap s3_asset_host:find_all
cap s3_asset_host:upload_changed
cap s3_asset_host:upload_all
cap s3_asset_host:upload
cap s3_asset_host:reset_and_upload
cap s3_asset_host:setup
cap s3_asset_host:create_buckets
cap s3_asset_host:delete_all
cap s3_asset_host:connect

You can get documentation on these tasks by running cap -T

Requirements

This plug-in is a Capistrano extension. It requires Capistrano 2.0.0 or greater.

You will also require the aws-s3 gem

So far, this plug-in:

assumes that you are using the ‘checkout’ method of deployment.
only works with svn.

If you are using another version control system, I think all you’ll have to change is the two methods in lib/scm.rb. If you do get something other than svn working, please let me know.

If you want to use more than one asset host, then you have to either install the multiple asset hosts plugin
or upgrade to Rails 2.0 (see setting up multiple asset hosts in Rails)

Setup

To set-up, you need to do the following

Install the plug-in
Install the AWS-S3 gem.
Set up your Rails application to use asset hosts.
Set up your asset hosts.
Configure Capistrano.

Installing the plug-in

From RAILS_ROOT, run:

script/plugin install svn://svn.spattendesign.com/svn/plugins/synch_s3_asset_host

Installing the AWS-S3 gem

You need to do this on both your local computer and the computer that is defined as the asset_host_syncher (see Capistrano Configuration, below).

$> sudo gem install aws-s3

Setting up your Rails app to use asset hosts

Single asset host

For a single asset host, simply add the following line to RAILS_ROOT/config/environments/production.rb:

config.action_controller.asset_host = "http://assets.example.com"

Multiple asset hosts

Follow the instructions in setting up multiple asset hosts in Rails

Setting up your asset hosts

Set up a CNAME entry for each asset host pointing to s3.amazonaws.com. How you do this depends on your domain host. Here’s what it looks like on easydns

You may need to wait up to 24 hours for the DNS entries for these new hosts to propagate.

Configuring Capistrano

Capistrano installation

This plugin requires Capistrano 2.0.0 or greater.

To upgrade to the latest version (currently 2.1.0):

$> gem install capistrano

Once the plug-in is installed, make sure that the recipes are seen by Capistrano

$> cap -T | grep s3_asset_host

should return a bunch of tasks. If you don’t see anything listed, then you need to update your Capfile by doing the following (this is from Jamis Buck):

In Capistrano 2.1.0 or above:



$> cd RAILS_ROOT

$> rm Capfile

$> capify .

If you do not want to delete your Capify file, or if you are using Capistrano 2.0.0, add the following line to your Capify file:
Dir['vendor/plugins/*/recipes/*.rb'].each { |plugin| load(plugin) }

Capistrano configuration

Create a new file in RAILS_ROOT/config called synch_s3_asset_host.rb. Add the following lines to it, and edit to suit:

# =============================================================================
# S3 ASSET HOST OPTIONS
# =============================================================================
set :asset_host_name, "assets%d.example.com"
set :aws_access_key, "your Amazon AWS access key" # You can also set this in your environment as AMAZON_ACCESS_KEY_ID
set :amazon_secret_access_key, "your Amazon AWS secret" # You can also set this in your environment as AMAZON_SECRET_ACCESS_KEY
# set :dry_run, false# Set to true if you want to test the asset_host uploading without doing anything on Amazon S3
before "deploy:symlink", "s3_asset_host:upload_changed"

You have to do one more thing: in RAILS_ROOT/config/deploy.rb. Specify one of your web hosts as an “asset_host_syncher”, like this:

role :web, webserver1, :asset_host_syncher => true

The first deploy

Commit all changes to your rails application and do the initial bucket setup:

$> cap s3_asset_host:setup
$> svn commit -m "Adding synch_s3_asset_host plugin"
$> cap deploy

This will do the following:

Create your Amazon S3 AWS buckets
upload everything in RAILS_ROOT/public (in your svn repository) to each bucket
Set the revision in each bucket to the latest revision in your repository.

This could take a while if you have lots of images or other big files.

You’re done!

That should do it. Now, every time you run cap deploy, your asset hosts should be updated with any changes to files in RAILS_ROOT/public.

Let me know if you have any problems, suggestions or comments.

Serving local files while using an asset host

2007-10-25T00:00:00-07:00

NOTE: The svn repository referenced here is no longer running. Hopefully you’re running an new enough version of Rails that this post isn’t necessary.

So, you’ve set up an asset host for your Rails application, but you have some files that that you don’t want to or can’t put on your asset host.

As an example, Plot-O-Matic creates images (of graphs) on the fly. If I had just blithely switched over to using asset hosting, then you wouldn’t see any changes when you updated a graph on Plot-O-Matic, as the Rails application only updates the graph image in its local public directory, and the asset host file remains unchanged.

You have a few options to fix this.

You could, of course, just use straight HTML tags to get at your local files, but that’s ~~kind of~~ just plain ugly.

You could just create the files in a directory that’s not in public. But then you can’t let people link to them easily, and you’re still not able to use asset tag helpers.

Finally, you could create a asset tag helper that created local links. That’s what I chose to do on Plot-O-Matic, and here’s how I did it.

Getting started

This will only work with the latest changes to the asset host code. There are two ways you can get this.

Upgrade to Rails 2.0

For instructions, scroll to the bottom of this post.

Install the multiple_asset_hosts plugin

script/plugin install svn://svn.spattendesign.com/svn/plugins/multiple_asset_hosts

All this plugin does is monkeypatch in the multiple asset host functionality from Rails Edge.

Creating the helper

Let’s take a look at how the standard image tag helper works. The
following code is from
vendor/rails/actionpack/lib/action_view/helpers/asset_tag_helper.rb.

def image_tag(source, options = {})
  options.symbolize_keys!
            
  options[:src] = image_path(source)
  options[:alt] ||= File.basename(options[:src], '.*').split('.').first.capitalize
       
  if options[:size]
    options[:width], options[:height] = options[:size].split("x") if options[:size] =~ %r{^\d+x\d+$}
   options.delete(:size)
  end
end

The important part here is the line

options[:src] = image_path(source)

This is where the path is calculated. Let’s take a closer look at the image_path method

def image_path(source)
  unless (source.split("/").last || source).include?(".") || source.blank?
    ActiveSupport::Deprecation.warn(
      "You've called image_path with a source that doesn't include an extension. " +
      "In Rails 2.0, that will not result in .png automatically being appended. " +
      "So you should call image_path('#{source}.png') instead", caller
    )
  end

  compute_public_path(source, 'images', 'png')
end

OK, so all that that does is call compute_public_path. compute_public_path is the part that has been updated to work with multiple asset hosts:

def compute_public_path(source, dir, ext, include_host = true)
  source += ".#{ext}" if File.extname(source).blank?

  if source =~ %r{^[-a-z]+://}
    source
  else
    source = "/#{dir}/#{source}" unless source[0] == ?/
    source = "#{@controller.request.relative_url_root}#{source}"
    rewrite_asset_path!(source)

    if include_host
      host = compute_asset_host(source)

      unless host.blank? or host =~ %r{^[-a-z]+://}
        host = "#{@controller.request.protocol}#{host}"
      end

      "#{host}#{source}"
    else
      source
    end
  end
end

Aha! So, all we need to do to get local calls is make a call to compute_public_path with include_host = false. Let’s create a local_image_tag helper that does just that. The following code should reside in RAILS_ROOT/app/helpers/application_helper.rb

#Creates an image tag for a file on the server, ignoring any asset hosts.
#Other than that, this is equivalent and mostly stolen from asset_tag_helper#image_tag
def local_image_tag(source, options = {})
  options.symbolize_keys!
          
  options[:src] = local_image_path(source)
  options[:alt] ||= File.basename(options[:src], '.*').split('.').first.capitalize
  
  if options[:size]
    options[:width], options[:height] = options[:size].split("x") if options[:size] =~ %r{^\d+x\d+$}
    options.delete(:size)
  end

  tag("img", options)
end

# Returns the path to an image on the server, ignoring any asset hosts.
# Other than that, this is equivalent to asset_tag_helper#image_path
def local_image_path(source)
unless (source.split("/").last || source).include?(".") || source.blank?
  ActiveSupport::Deprecation.warn(
    "You've called image_path with a source that doesn't include an extension. " +
    "In Rails 2.0, that will not result in .png automatically being appended. " +
    "So you should call image_path('#{source}.png') instead", caller
  )
end

  compute_public_path(source, 'images', 'png', false)
end

This is too ugly. What we really want is to be able to create an image tag like this:

def local_image_tag(source, options = {})
  image_tag(source, options.merge(:include_host => false))
end

to do this, you would change the image_tag code to something like this

def image_tag(source, options = {})
  options.symbolize_keys!
  include_host = options.has_key?(:include_host) ? options[:include_host] :  true
  options.delete(:include_host)
  options[:src] = image_path(source, include_host)
  options[:alt] ||= File.basename(options[:src], '.*').split('.').first.capitalize

  
  if options[:size]
    options[:width], options[:height] = options[:size].split("x") if options[:size] =~ %r{^\d+x\d+$}
    options.delete(:size)
  end

  tag("img", options)
end

and the image_path code to

def image_path(source, include_host = true)
  unless (source.split("/").last || source).include?(".") || source.blank?
    ActiveSupport::Deprecation.warn(
      "You've called image_path with a source that doesn't include an extension. " +
      "In Rails 2.0, that will not result in .png automatically being appended. " +
      "So you should call image_path('#{source}.png') instead", caller
    )
  end

  compute_public_path(source, 'images', 'png', include_host)
end

Setting up multiple asset hosts in Rails

2007-10-24T00:00:00-07:00

NOTE: The svn repository referenced here is no longer running. Hopefully you’re running an new enough version of Rails that this post isn’t necessary.

One of the nice goodies coming in Rails 2.0 is the ability to use multiple asset hosts. This article explains how you can use this feature now, without waiting for Rails 2.0, and why you would want to use asset hosts.

What’s an asset host, and why would I want to use it?

An asset host is another server, somewhere on the internet, where you store your static files. These can be javascripts scripts, CSS stylesheets, images, static html files and anything else that doesn’t change often. Basically, anything that lives in your public directory. You can then use a link to the asset host whenever you want to include an image, javascript or stylesheet. For example, Plot-O-Matic stores its assets at assets0.plotomatic.com — assets3.plotomatic.com. You can see the logo here: assets0.plotomatic.com/images/o-logo.png.

So, why would you want to use an asset host? It turns out that many browsers limit the number of simultaneous connections to a host. For Internet Explorer, that number is two. If you are serving a lot of small images, or you haven’t bothered to bundle your scripts or stylesheets, this can be a real bottleneck. Asset hosting allows you to increase the number of hosts a web page is loaded from, removing the bottleneck.

Asset hosting in Rails

Rails has had support for using a single asset host for at least a year. It’s achieved by setting config.action_controller.asset_host in your development environment, development.rb. For example,

config.action_controller.asset_host = "http://assets.plotomatic.com"

Once you do this, rails will pre-pend every link created with an asset tag helper with your asset host name. So, a link like this

<%= image_tag 'o-logo.png' %>

will now result in the following link

<img alt="o-logo" src="http://assets.plotomatic.com/images/o-logo.png?1193172652" />

Asset tag helpers are things like image_tag, stylesheet_link_tag or javascript_include_tag. Links made without asset tag helpers will not automatically use your asset host.

Multiple asset hosts

Now, having a single asset host is nice, but using multiple asset hosts can really speed up those page load times. Here’s how you do it:

Get Rails ready for multiple asset hosts

There are two ways you can do this.

Upgrade to Rails 2.0

For instructions, scroll to the bottom of this post.

Install the multiple_asset_hosts plugin

script/plugin install svn://svn.spattendesign.com/svn/plugins/multiple_asset_hosts

All this plugin does is monkeypatch in the multiple asset host functionality from Rails Edge.

Pick a name for your asset hosts

The change from single to multiple asset hosts is simple. You simply put a %d somewhere in your asset host name in production.rb. For Plot-O-Matic, the setting is

config.action_controller.asset_host = "http://assets%d.plotomatic.com"

Rails will replace the %d with either 0, 1, 2 or 3.

Set up your asset hosts

You need to provide four asset hosts, so I set up four asset hosts named
assets0.plotomatic.com, assets1.plotomatic.com, assets2.plotomatic.com and assets3.plotomatic.com.

I used Amazon S3 for my asset hosts, which was simple:

set up four buckets called assets0.plotomatic.com, assets1.plotomatic.com, assets2.plotomatic.com and assets3.plotomatic.com
Set up a CNAME entry for each asset host pointing to s3.amazonaws.com. How you do this depends on your domain host. Here’s what it looks like on easydns

Deploy it

Commit your changes to your version control system
Upload everything in your public directory to each of your asset hosts.
Deploy and restart your web server

running tests on deploy

2007-10-19T00:00:00-07:00

A simple Capistrano 2.0 recipe, installable as a plugin, that runs all of your tests before the final symlink test.

Synopsis

This plug-in is a Capistrano extension. It requires Capistrano 2.0.0 or greater.

Once installed, running cap deploy will run all of your tests before doing the final symlink.

If the tests fail, the symlink will not be created and your deployment will roll back.

to deploy without tests,

cap deploy:without_tests

Installation

Plugin installation

You should be able to install with the following command (from rails root):

Rails 3:
rails plugin install http://github.com/spatten/Run-Tests-on-Deploy.git

Rails 2:
script/plugin install http://github.com/spatten/Run-Tests-on-Deploy.git

If that doesn’t work, send me an e-mail at scott@scottpatten.ca

Capistrano configuration

This plugin requires Capistrano 2.0.0 or greater.
To upgrade to the latest version (currently 2.1.0)

gem install capistrano

Once the plug-in is installed, make sure that the recipes are seen by Capistrano

cap -T | grep deploy:without_tests

should return

cap deploy:without_tests # deploy without running tests

If capistrano is not seeing the deploy:without_tests task, then you need to update your Capfile.

(The following is from a post by Jamis Buck)

In Capistrano 2.1.0 or above:
you can delete your Capify file in rails root, and then, from rails root, run

capify .

If you do not want to delete your Capify file, or if you are using Capistrano 2.0.0, add the following line to your Capify file:

Dir['vendor/plugins/*/recipes/*.rb'].each { |plugin| load(plugin) }

Details

The deploy:run_tests task is executed before the symlink
task. The run_tests task does the following:

prepares your test db with rake db:test:prepare
runs all of your tests with rake test at a nice level of -10

The deploy:run_tests task won’t work if you call it by itself, as it runs from the
release_path directory, which won’t exist unless called after deploy:update_code.

to deploy without tests, cap deploy:without_tests

you can also set the run_tests option to 0 from the command line like this:

cap -S run_tests=0 deploy

this allows you to do things like deploying with migrations but without tests:

cap -S run_tests=0 deploy:migrations

Genesis

The original idea for the :run_tests task is from the testing rails blog,
which, sadly, seems to be defunct.

updating rdig on deploy

2007-10-16T00:00:00-07:00

Here’s a simple little Capistrano 2.0 recipe to update your rdig search index after every deploy.

Add the following lines to config/deploy.rb

namespace :deploy do
  after :deploy, "deploy:update_rdig"
  desc "Update the rdig database.  This has to be done after
        re-starting,  as it looks at the live web-site"
  task :update_rdig do
    run "cd #{current_path} && rdig -c config/rdig_config.rb"
  end
end

This will run automatically after every deploy. If you just want to run it manually, comment out the line that starts with after ....
You can then run it manually with cap deploy:update_rdig

Ruby unit test command line options

2007-10-11T00:00:00-07:00

Here are a few tips and tricks for unit tests

To find all of the command line options for the test::unit suite, call a unit test with --help:

ruby some_test.rb --help

The output will look something like this

~/tests $>ruby some_test.rb --help
Test::Unit automatic runner.
Usage: test/unit/graph_test.rb [options] [-- untouched arguments]

    -r, --runner=RUNNER              Use the given RUNNER.
                                     (c[onsole], f[ox], g[tk], g[tk]2, t[k])
    -n, --name=NAME                  Runs tests matching NAME.
                                     (patterns may be used).
    -t, --testcase=TESTCASE          Runs tests in TestCases matching TESTCASE.
                                     (patterns may be used).
    -v, --verbose=[LEVEL]            Set the output level (default is verbose).
                                     (s[ilent], p[rogress], n[ormal], v[erbose])        --                           Stop processing options so that the
                                     remaining options will be passed to the
                                     test.
    -h, --help                       Display this help.

Deprecated options:
        --console                    Console runner (use --runner).
        --gtk                        GTK runner (use --runner).
        --fox                        Fox runner (use --runner).

Running only some tests

The -n option is my favourite. You can use it like this to call a single test by giving it the name of a test method:

ruby some_test.rb -n test_something_very_important

Even better, you can use a regular expression and all of the matching tests will be run. For example, if I had a graph model that tested the creation of .png, .eps and .gif files with methods called test_create_png_file, test_create_eps_file and test_create_gif_file, I could run them all with the command

ruby test/unit/graph_test.rb -n /_create_.*_file/

you can also use the -n option multiple times. So, if I just wanted to test eps and png file creation, I could use

ruby test/unit/graph_test.rb -n test_create_eps_file -n test_create_png_file

The same thing applies to selecting test cases using the -t option.

Cranking up the verbosity

Sometimes you want to see what tests are failing without waiting for all of the tests to run. Use the -v or --verbose options to do this.

ruby test/unit/<some_test>_test.rb -v v

ruby test/unit/<some_test>_test.rb --verbose=verbose

Then, instead of just the boring old periods appearing as each test passes

~/rails/grapher/trunk/vendor/plugins/ploticus $>ruby test/ploticus_test.rb 
Loaded suite test/ploticus_test
Started
.................
Finished in 0.10768 seconds.

17 tests, 118 assertions, 0 failures, 0 errors

you’ll see this instead:

~/rails/grapher/trunk/vendor/plugins/ploticus $>ruby test/ploticus_test.rb -v v
Loaded suite test/ploticus_test
Started
test_data_from_columns(PloticusTest): .
test_data_from_columns_with_ragged_data(PloticusTest): .
test_data_from_hash(PloticusTest): .
test_data_from_hash_with_nils(PloticusTest): .
test_data_from_hash_with_ragged_data(PloticusTest): .
test_data_from_rows(PloticusTest): .
test_first_proc_by_name(PloticusTest): .
test_has_proc(PloticusTest): .
test_num_procs_by_name(PloticusTest): .
test_pad_to(PloticusTest): .
test_ploticus_proc_has_line(PloticusTest): .
test_ploticus_proc_line(PloticusTest): .
test_png_graph_creation(PloticusTest): .
test_proc_with_colons_in_it(PloticusTest): .
test_procs_are_enumerable(PloticusTest): .
test_procs_by_name(PloticusTest): .
test_svg_graph_creation(PloticusTest): .

Finished in 0.042398 seconds.

17 tests, 118 assertions, 0 failures, 0 errors

Obviously you don’t want to do this all of the time, but it can come in handy.

Running tests with a GUI

If you get bored with the command line, try using the “runner” option. The tk option worked for me without having to install anything:

ruby some_test.rb -r tk

You’ll see something like this:

Other valid options are console, fox, gtk, gtk2, tk.

You can mix the -n and -t options as well if you want

ruby test/unit/graph_test.rb -r t -n /downloadable/

Updating dynamic DNS on Amazon EC2

2007-10-10T00:00:00-07:00

basic idea

There are two steps to updating your dynamic DNS.

Get your IP address
Tell your dynamic DNS provider the IP address

On EC2, things are a little different from your standard hosting setup. Here’s how you get dynamic DNS working on EC2.

Getting your IP address

See this Amazon article about getting data about your instance for more information about instance data.

To make a long story short, to find out the public IP address of your EC2 instance, you make a http get request to

http://169.254.169.254/latest/meta-data/public-ipv4

Note that this has to be done from your EC2 instance.

Try it out from the command line. Log-in to your EC2 instance and type

curl http://169.254.169.254/latest/meta-data/public-ipv4

From Ruby, it’s just as simple. Cut and paste this into a script, or just run it in irb.

require 'open-uri'
ip = open('http://169.254.169.254/latest/meta-data/public-ipv4').read

Updating your DNS with your dynamic DNS provider

The following instructions assume that you want to use ez-ipupdate to update your DNS. If you don’t want to, or if ez-ipupdate doesn’t support your dynamic DNS provider, it should be relatively simple to get it working anyways.

For example, easy-dns (who I use) just requires logging in and then sending a get request formatted like this:

http://members.easydns.com/dyn/dyndns.php?hostname=example.com&myip=10.0.0.2

Installing ez-ipupdate

If you’re running Ubuntu, then install it using sudo aptitude install ez-ipupdate or sudo apt-get install ez-ipupdate. On a RedHat installation sudo yum install ez-ipupdate.

You can also install from source, of course.

When you install it using aptitude or apt-get, it will ask you for some configuration info. This won’t work on EC2, so ignore the configuration and use the script below instead.

Installing the Ruby script

Here’s the script:

#!/usr/bin/env ruby
require 'open-uri'
require 'yaml'

DYNAMIC_DNS_CONFIG_FILE = '/etc/ez-ipupdate/dynamic_dns.yml'
AMAZON_INSTANCE_DATA_ADDRESS = "http://169.254.169.254"
api = "latest"
ip = open(File.join(AMAZON_INSTANCE_DATA_ADDRESS, api, 'meta-data', 'public-ipv4')).read

dns_config = YAML::load(File.open(DYNAMIC_DNS_CONFIG_FILE))
hosts = dns_config['host'].split(',')
hosts.each do |host|
  puts "updating dynamic dns to ip = #{ip} for host #{host} using dynamic DNS service " +
       "#{dns_config['service']}"
  command = "ez-ipupdate --address #{ip} --service-type #{dns_config['service']} " +
            "--user #{dns_config['username']}:#{dns_config['password']} --host #{host}"
  puts command
  system(command)
end

If you’re running an ec2-on-rails instance, place the script in /usr/local/ec2-on-rails/startup-scripts/every-startup/update_dynamic_dns.rb

Otherwise, place it somewhere in your path and create a shell script to call it in /etc/init.d.

Make sure that both update_dynamic_dns.rb and the init script are executable.

The config file

Create a file called /etc/ez-ipupdate/dynamic_dns.yml. It should look something like this

# service should be one of the services supported by ez-ipupdate. 
# Possible values: null  ezip  pgpow  dhs  dyndns  dyndns-static 
#                         dyndns-custom ods tzo easydns easydns-partner 
#                         gnudip justlinux dyns hn zoneedit heipv6tb
# (The above list is from man ez-ipupdate)
service: service_type 
username: username_for_your_dynamic_dns_provider
password: password_for_your_dynamic_dns_provider
host: example1.com,example2.com

To update multiple hosts, put them all on the host entry, separated by commas.

Testing and debugging

Test out the script:

run it: ./update_dynamic_dns.rb.
Wait a couple of minutes.
ping your domain name from a local machine and make sure that it points at the right IP address.

Hopefully that works for you. Let me know if you have any problems, comments or suggestions.

Scott

Welcome

2007-10-09T00:00:00-07:00

I figured I needed to write a quick ‘first post’ to introduce myself and spatten design.

I’m Scott Patten. I’m a programmer, web developer and physicist.

spatten design is where I post useful stuff I’ve discovered or created. I’ll be writing about Ruby, Rails, Ajax, CSS, Amazon Web services and, as it says in the side-bar, whatever else strikes my fancy.

spatten design is also a Ruby on Rails consulting business. I provide Ruby on Rails training and Ruby on Rails web development services.

Besides the consulting work I do as spatten design, I also spend my time working on Plot-O-Matic, a scientific graphing web application written in Ruby on Rails. If you want to see what I can create, go check it out.

Thanks for dropping by, and welcome.

Scott Patten's Blog

A Truly Minimum Viable Kitchen

Truly Minimal

Why bother with a minimal kitchen anyways?

No Excuses

Getting up and running with Amber

Getting it on your machine

Creating Hello World

Hooking up a form

Further Reading

Speeding up Jekyll Generation

Opening up ports to your security group on EC2

Multiple Google Analytics Tracking Codes on one Page

How big is a Googol, anyway?

The number of grains of sand on the earth.

How many stars are there?

How many atoms are in the universe?

How many particles are in the universe?

Packing the universe with sand

What did I miss?

Booting an Alestic EC2 Image with a Larger Root Disk

SLL Subscribers Vs Time

The Myth of the Zone

Custom Validators with Validatious

Validating with Validatious: A Quick Tutorial

A quick intro to Validatious

The Gotcha

Validating on blur

Putting validations closer to the input field

Validating via Javascript

Bringing it home

My Git Aliases

How scottpatten.ca came to be

The S3 Cookbook

I'll be speaking at RailsConf

Searching Beast and WordPress from a Rails app

The Search Engine

Searching the Beast Forum

Beast setup

Searching a non-integrated Beast forum

The Topic Model

The Post Model

acts_as_ferret

scrubbing the html tags

the url method

Searching the WordPress Blog

The FullTextSearch mixin

Re-indexing the non-local models

Rubyize this: Live in Vancouver. Refactoring #1.

Rubyize This: Live in Vancouver. Refactoring #3

Rubyize This: Live in Vancouver. Refactoring #2

Adding json responses to all of your controller actions

Fun with Single Table Inheritance

The Set-up

Creating the Database Table

Creating the Models

Adding behaviour to STI models

RubyCamp Vancouver

When and Where

Who should come

The Conference Track

The Hackathon Track

A weekend of Ruby

Amazon S3, Ruby and Rails slides

s3sh demo script

S3Syncer Code

S3Syncer demo

synch s3 host plugin update

Amazon S3, Ruby and Rails talk

Synching Your Amazon S3 Asset Host using Capistrano

Usage

Requirements

Setup

Installing the plug-in

Installing the AWS-S3 gem

Setting up your Rails app to use asset hosts

Single asset host

Multiple asset hosts

Setting up your asset hosts

Configuring Capistrano