The purpose of SSL is to provide a reasonable level of protection against eavesdropping and man-in-the-middle attacks. Although SSL provides a greater level of security, it introduces a lot of overheads and hence should be used sparingly. Two of the most common places to use SSL is for payment transactions and user registration/login.
This post intentionally focuses only on the Rails application as there are numerous post on the net for SSL setup on the server. Enabling SSL in a Rails application is really trivial and there are just a few points that need your attention..

1. Enabling SSL
a. Install the ssl_requirment plugin:

./script/plugin install git://github.com/rails/ssl_requirement.git

b. Include it in your application_controller.rb:

include SslRequirement

c. Specify actions that require SSL in their respective controllers. For eg. my session controller has the following line:

ssl_required  :new, :create if Rails.env.production?

d. Add the following line in development.rb to bypass SSL in development mode:

SslRequirement.disable_ssl_check = true

2. Gotcha’s
a. Include all submit actions in requirement
Any action that processes form data from a SSL page should also be added to the requirement. In the above example, the form on the login page(new action) is processed by the create action and hence it is also included in the requirement.
b. Ajax actions
Ajax actions on a SSL page should also use SSL and must be included in the requirement. At times you do not have a body for the Ajax action and it is rendered using it’s respective RJS template. In such cases create an empty action and include it in the ssl_requirement.
c. Mixed content
A lot of browsers show you a “Mixed Content Warning” if your SSL page references non-SSL assets. IE displays a scary looking confirmation dialog while Firefox and Chrome show a exclamation in the url bar. Any relative paths(eg. using _path helpers) on the page will automatically use the https protocol but any absolute paths(eg. using _url helper or by manually specifying as a string in link_to) will need to be changed to use https.
d. Asset host issue
If you are using Rails asset hosts and do not have a SSL certificate that supports wildcard(for subdomains), then you need to disable them for the SSL pages. Just add the following code to your production.rb:

ActionController::Base.asset_host = Proc.new { |source, request|
  if request.ssl?
    "#{request.protocol}#{request.host_with_port}"
  else
    "#{request.protocol}assets%d.yourdomain.com" % (source.hash % 4)
  end
}

Replace “yourdomain” with your apps domain and “4″ with the number of asset hosts required.

The above should ensure that you have a proper SSL setup without displaying warnings to the user.

Related posts:

1. Ruby on Rails Caching And JavaScript Techniques Cross posted from darthsid While implementing caching in a recent...
2. 11 Things to Consider Before Deploying Your Ruby on Rails Application At VinSol, we have been developing and deploying Rails applications...

Related posts brought to you by Yet Another Related Posts Plugin.

“FullCalendar” provides a full-sized, drag & drop calendar. It uses AJAX to fetch events on-the-fly for each month. It also supports an intuitive interface to manage events that spans over multiple days or weeks. It is visually customizable and exposes hooks for user-triggered events (like clicking or dragging an event).

I decided to give it a try and utilize its hooks for user triggered events within our Rails application. This small effort resulted in a barebone Rails app that might provide a good base for your project which require calendar, scheduling or appointment features. I called it fullcalendar_rails and it is now available on github with a working demo at http://fullcalendar.vinsol.com.

Feel free to give your valuable feedback. I hope you will find this useful.

Update: On popular demand, I have added recurring events functionality with daily, weekly and monthly frequencies. It also allows for exceptions to recurring events including delete and edit features.

Related posts:

1. Ruby on Rails Caching And JavaScript Techniques Cross posted from darthsid While implementing caching in a recent...

Related posts brought to you by Yet Another Related Posts Plugin.

I do like to skip ri and RDoc documentation while installing gems on my development machine, because it takes more time to generate ri and RDoc then actual installation of gem.

We can skip rdoc and ri documentation while installing a gem by:

I am sure that you would not like to give –no-ri and –no-rdoc every time you install a gem. To avoid this situation, you can create/update $HOME/.gemrc file with following option:

Now every time you install a gem, it will skip installation of ri and RDoc for the gem.

My .gemrc file looks like:

Basically I want to implement something like where a user can drag drop large image to view specific portion.

I am signing up there as a beta tester. Lets see how webby-node can help rails community by its services.

Surprised, This is not a typo. Saturday is now converted to Funday. How??? Why??? When???

Geekeerie has started organizing new kind of events in India. One of those kinds of event is Ruby Funday. Which is a full day pure geek ruby event. Where geeks will show their code, tricks and funde .

The first Ruby Funday is scheduled for 22, Nov 2008 at Noida, India. For more details please visit their site(beta).

Related posts:

1. Jquery Full Calendar with Ruby on Rails Contrary to popular belief, working on a client project gives...

Related posts brought to you by Yet Another Related Posts Plugin.

Also if you have 2-3 or more apps running in production then managing such exception mails is also a big headache. In such case one have to keep track of many things like which type of error is resolved/unresolved for which project etc… .

So, here is a good news for those who don’t know about Hoptoad. It is an hosted service by thoughtbot which receives your exceptions, notify you once per error type by email and keep track(resolved/unresolved, count etc…) of your errors on project basis.

By now its a free service. I’m gonna use this as my next project goes live. What abt you??? ;-P

But, for SEO sake I had to change configuration so that wordpress can be access by http://domain.com/blog instead of http://blog.domain.com/

The problem was if I configure wordpress for http://domain.com/blog and go to this url, the request was handled by rails app because of domain.com virtualhost.

So what I did? I changed apache virtualhost configuration for http://blog.domain.com and http://domain.com as:

Also I created a symbolic link to wordpress installation directory under rails public folder(ln -s /var/www/html/wordpress /var/www/html/railsapp/blog).

I restarted apache and it worked fine. Wordpress was running at http://domain.com/blog and rails app was as http://domain.com/.

For those who are new to Ultrasphinx: Ultrasphinx is a rails plugin and client to the Sphinx(full text search engine) written by Evan Weaver. More about Ultrasphinx here.

Lets get into the situation. Consider a STI case where we are using ultrasphinx to index several fields:

Now assume we have following data in posts table:

and now in application root:

By this point we have created sphinx configuration file, indexed all records from ultrasphinx models and started sphinx search daemon.

Now open script/console and create and fire a query to find some articles:

Here when you run the query you get an exception, because it is trying to find an article with id 5, which actually a story type not article. So why it is trying to find such article?

Now have a look into config/ultrasphinx/development.conf file. Under section “source articles_main” you’ll get “SELECT (posts.id * 3 + 0) AS id, ‘’ AS article_title, posts.body AS body, ‘Article’ AS class, 0 AS class_id, posts.title AS title FROM posts WHERE posts.id >= $start AND posts.id <= $end GROUP BY posts.id" Which is a SQL to get record to index.

If you check it carefully you'll findout that it should select all articles record to index but unfortunately it is selecting all records from table and considering them as articles. To fix it you need to modify this query and add "posts.type = 'Article'" in where condition. So the query should be "SELECT (posts.id * 3 + 0) AS id, '' AS article_title, posts.body AS body, 'Article' AS class, 0 AS class_id, posts.title AS title FROM posts WHERE (posts.type = 'Article') and posts.id >= $start AND posts.id <= $end GROUP BY posts.id"

But still there is a problem. If you do this manually you have to do it again and again whenever you issue "rake ultrasphinx:configure" because this configuration file will be overwritten.

Better option is to add conditions in models as is_indexed options like:

and it worked fine.

Manik gave me an idea to write a patch for ultrasphinx to add such conditions automatically in case of STI. So may be another post related to ultrasphinx will be soon ;-)

Update: Here is the patch Git patch: Fix STI Issue. After applying this patch you need not to add conditions explicitly for such case. It will automatically check and add conditions for STI.