Question

I am running into an issue when looking into fragment-level caching within my Rails 3.0.4 application with memcached. I am a bit confused with what is going on, but I think it’s something to do with the way the output is being pulled from within the caching region. I am running memcached locally in -vv mode, and can see the key for the fragment getting saved/pulled correctly, the problem is the value of the item within memcached.

Here is what I’m doing:

< ... html before ... >
<%= cache("item_#{i.id}") do %>
    <%= render :partial => 'shared/item', :locals => { :item => i, :functionality => [:set_as_default] } %>
<% end %>
< ... html after ... >

When I look at the value of the key within the cache, it has html from within the page that is in that fragment cache block, but ALSO OUTSIDE of that (from both the html before and html after areas). Here is the interesting part though, and is kind of the reason I think its related to capturing the output–it doesn’t do the whole page, only some of the html before and some after.

According to the rails fragment cacheing guide, I think I’m doing things correctly (http://guides.rubyonrails.org/caching_with_rails.html#fragment-caching). Does anyone have thoughts as to what could be going on?

Your help is much appreciated!

-Eric

Answer 1

In this case, you are using ERB incorrectly. Basically take out the = sign. What your doing is your returning the value of the block too and hence why you are seeing double output.

<% cache(“item_#{i.id}”) do %>

Also, ActiveRecord objects respond to an internally baked in #cache_key method. Try to take advantage of that. The default #cache_key for an ActiveRecord object uses the class name, the object id and the updated_at timestamp too. The cache method should be able to take multiple args or an array and it will inturn call cache_key for every object that responds to it. Using this method, it means you will cache miss when the object is updated to, pretty cool stuff. So, IIRC

<% cache(“item”,i) do %>