Gregg Kellogg

Caching actions without layout can be complicated when multiple request formats are used. In particular, an HTML response may use a dynamic layout, in which case you want to use the :layout => false option. However, other formats (such as XML) don't use a layout, but the :layout => false option to _caches_action_ does not properly cache the body in this case. To solve the problem, create two caches action statements:

caches_action :show,
              :if => lambda { |c| c.request.format == :html },
              :cache_path => lambda { |c| c.cache_key },
              :layout => false   caches_action :show,
              :unless => lambda{ |c| c.request.format == :html },
              :cache_path => lambda { |c| c.cache_key },
              :layout => true

Also, relying on the accept header may not cause the action caching module to detect the appropriate format. Try this in your application_controller:

before_filter :set_explicit_request_format
def set_explicit_request_format
  # Set format explicitly from accept header, unless it's already set
  request.format = :html if request.format == :any
  params[:format] ||= request.format.to_sym.to_s
end

Rails offers three forms of caching within your controller: page, action and fragment. Page caching results in the fastest access times, as the results of the first call to an action are saved in a file so that subsequent accesses never even hit rails. However, for most applications, this isn't useful, as there may be dynamic content on a page, and this does not allow for authentication. Fragment caching is the most detailed, and allows different parts of a page to be cached and allows you to check for the presence of a cached fragment within your controller (or view), but it requires the most maintenance of cache keys. Action caching is a nice compromise between the two. It allows the controller to get into the action but takes care of cache key creation. It also allows for a dynamic layout using the :layout option. However, in some actions, the amount of work done by the controller may be non-trivial, so it would be nice to check for the presence of the cache within the body of the controller action. This can be solved by borrowing some code from within ActionController::Cachine::Actions.

def index
  cache_path = ActionCachePath.new(self, cache_key)
  return if self.read_fragment(cache_path.path)
  # body of action
end

Controlling your own cache keys is also useful, particularly for actions that may take a number of parameters:

def cache_key
  key = "#{params[:controller]}/#{params[:action]}"
  params.each_pair {|k, v| key += ":#{k}=#{v.gsub(/\s/, "_")}
  key
end

The ButtonLabels plugin described in another post has been updated for Rails 2.3.

After a fair amount of work, I'm happy to report that HTTP Digest Authentication is now a part of Rails 2.3. Although I put the finishing touches to get this into the release, it is based on work done by Dan Manges and Xavier Shay . Also, thanks to Don Parish for bug fixes and improvements after original acceptance.

Read more about HTTP Digest Authentication in Rails 2.3 Ryan's Scraps. Relevant Lighthouse entries: 1230, 1848, and 2000. The last one includes a change, not yet approved for 2.3, which allows for using the HA1 part of the digest to store a hash of the password, rather than the cleartext of the original version. Hopefully, we'll get a version of that in soon. Also, the current implementation depends on using a session secret when computing the nonce. 2000 proposes a way to avoid this so no session is required.

Hopefully. we'll see the open issues resolved and get this into a 2.3.1 update.

Gregg

EagerFinderSql allows custom SQL to be specified when doing eager loading of associations through the :include option
to find. This allows for purpose-constructed queries to be used and still result in a fully linked
object model.

Background

ActiveRecord constructs SQL to satisfy the requirements of a find request. Associations allow for customized
SQL to be specified, using the :finder_sql option, but this has not been available when performing eager loading
using the :include option. The result is that a standardized query is constructed to bring in the associated
tables using LEFT OUTER JOIN. For some queries, this can be result in expensive queries and potentially very
large result sets.

Custom SQL

EagerFinderSql addresses this problem by allowing :finder_sql to be added
to find options when the :include option is also specified.