If you’re writing an i18n-enabled Rails application, you have to deal with the translation of your own application as well as with the translation of several Rails core strings (like validation error messages, date formats, and so on). You usually don’t need to translate the Rails core strings yourself, since there’s a big repository of user-contributed core translations you can pull into your application. However keeping this translations up to date in your application can quickly become cumbersome.

For this reason, I refactored the rails-i18n(*) repository a while ago and made it a Rails plugin. You could install the plugin to your application and it provided Rails core translations for you. Unfortunately it became annoying to keep the fork up to date with the latest translation changes since every rebase or merge was a pain because of the moved file locations.

So my latest solution is another small plugin called rails-i18n-updater, which does not contain the translations itself but downloads them from the above mentioned repository of core translations.

Just out of curiosity, I took two of my recent Rails applications today and tried them with Ruby 1.9. It was surprisingly easy to make all tests pass without any warnings or errors. Rails 2.3 already has quite good support for Ruby 1.9. The main gotcha however was about file encodings. If you have source files which contain non-ASCII characters, Ruby 1.9 now needs to know which encoding the file was saved with. If you don’t specify the encoding for a file with non-ASCII characters, you’ll get an invalid multibyte char (US-ASCII) error message.

Last week I received an interesting bug report for an application I’m working on at the office. It has a controller with an index action that displays a list of items which can be filtered by tags. A tester reported that every time he chooses to filter the list by a tag that contains a dot, the site returns an error. First this seems strange since tags with a dot worked perfectly fine in tests.

But after digging deeper, I found a surprising reason for this strange error: To make URLs look nicer, I defined an extra route like this:

1 map.things 'things/:tag', :controller => 'things', :action => 'index',
2   :tag => nil

The intention was, to have /things/foo instead of /things&tag=foo as the URL. This actually worked fine – except for cases when the tag contained a dot.

While moving some applications to Rails 2.3 recently, I stumbled across some problems with testing cookies. E.g. this blog uses cookies to remember your name, email address and web url if you leave a comment. This way, you don’t need to re-enter these information the next time you leave a comment. These cookies are set by the controller action that creates new comments (in CommentsController#create):

1 cookies['blog_visitor_name'] = { :value => @comment.name, :expires => 1.year.from_now }

This is the extended form of setting a cookie (a hash is used to not only set the cookies value but also the expiration time. There are several more hash options available which are described in the ActionController#cookies documentation. The basic form (which only sets a simple session cookie that is valid until the visitor closes the browser) uses a simple string:

On April 9th 2009, the fourth CSS Naked Day event will take place. The idea behind this event is to promote web standards (like proper semantic markup and a good hierarchy structure). On April 9th, participants are encouraged to completely remove all stylesheets from their site, stripping it entirely of its design. If your site has proper semantic markup, it’ll stay well usable and understandable even without styles. If not, you better hide and don’t take part in this event :)

To make it as easy as possible, I created a little Rails plugin that, once installed to an app, simply disables the stylesheet_link_tag helper for the duration of the event. It’s as easy as install and forget (assuming that you used stylesheet_link_tag in your layouts and didn’t add inline styles or stylesheet links manually).

So let’s see how many sites will join this funny but yet expressive event this year. It’s time to show off your <body> ;-)

There are many factors that determine how long a page takes to show up in a browser. When a page is requested by a browser, the server needs some time to compute the page contents (controller, model/database, view) and returns HTML to the browser. Besides network latency and throughput (which can be optimized by choosing a good hosting provider), there are several ways to speed up the page generation itself (like memoization, action/fragment/query-caching, using memcached, and so on).

But even after the HTML page itself has been transferred to the browser, the page is not necessarily ready to display yet. A page usually references more resources like stylesheets, javascripts and images that need to be requested and transferred. These are mostly static files (located in the public folder of a Rails application) and are directly served by the webserver without invoking the Rails application. So these files takes very little computation time on the server, but they still need some time and bandwith to transfer, which can be optimized.

Rails I18N gives you a way to translate your views and easily switch between different languages. However, you still need to set the locale for each request, i.e. you have to choose a method to select the right locale for a request. This can be done in various ways, depending on how you perfer it to behave. Here are some examples.

1. by browser-setting
2. by toplevel domain
3. by subdomain
4. by manual selection

A while ago, I wrote about different methods in JavaScript to prevent spam harvesters from recognizing an email address. These methods work by placing a scrambled version of the email address into the page source so that a spam harvester cannot recognize it as an email address. Using JavaScript, the scrambled text is unscrambled and displayed as usual to human visitors. Usually, the “scrambling” is based on replacing characters of the email address with its hex-entities (Rails’ mail_to helper does so if using :encode => :hex or :encode => :javascript). My theory was/is, that using hex-entities is not sufficient anymore nowadays, since they can be easily reversed with simple search-and-replace operations.

So I came up with the idea to use a scrambling method that cannot be easily reversed. I assumed that spam harvesters probably can decode hex-entities, but still aren’t able to execute JavaScript. However since this was just an assumption, I started a simple test over the last 6 months to find out how good or bad the different scrambling methods perform.

Since Rails 2.2 has been released last week, lots of people are beginning to translate their applications — or, at least, many people seem to prepare their applications for later translation by replacing strings in views with t() calls.

If you’re modifying your application to be i18n-ready, or if you’re creating a new multilingual application, here’s a useful guideline to remember: translate meanings, not words.

Just a quick note that this blog is now running Rails 2.2 and makes use of the new i18n framework. Although this site is completely in English, the same application is also serving another blog in German. Moving the blog application from ruby-gettext translations to Rails i18n was quite a bit of work since the i18n simple backend works differently and after all, I had to recreate all translation files.

In the end, it works very well and reliable. While working with i18n views, I had some small ideas for improving translation handling in views – maybe I can come up with some minor improvements after thinking about it some more.

The main difference between gettext and Rails i18n is the usage of translation keys. With gettext, you place full strings of the “master language” (usually English) into the views. With Rails i18n, you build up a hierarchical structure of translation symbols (like “sidebar.latest.title”) and associate them with real text in translation files. Currently there’s no tool to extract symbols used in views to assist you to keep translations complete (like rgettext does for gettext), but I guess helpers will appear soon.

For my translations, I used the i18n simple backend, which comes with Rails and stores translation strings in simply .yml files. However, the Rails i18n framework actually is much more than just translations in .yml-files. It allows translation backends to hook into the Rails core with a documented interface, which promises to be more stable and reliable than monkeypatching solutions (e.g. like ruby-gettext does). In theory, it should be possible to create an i18n backend that utilizes gettext — which certainly is still very interesting e.g. to upgrade existing applications or for large applications with many translators (since the gettext format is widely supported by a lot of tools and services).