Missing template blogs/index on Ruby on Rails project

Question

For one of my projects I'm getting this exception every now and then:

ActionView::MissingTemplate: Missing template blogs/index with {:handlers=>[:rxml, :erb, :builder, :rjs, :haml, :rhtml], :formats=>["image/jpeg", "image/pjpeg", "image/png", "image/gif"], :locale=>[:en, :en]} in view paths "/var/www/keeponposting/releases/20110403083651/app/views"

It seems someone is requesting an image from a URL that isn't an image:

HTTP_ACCEPT "image/jpeg, image/pjpeg, image/png, image/gif"

Any ideas what to do about it? Do I have to implement a handler for one of those and return "" to get rid of this exceptions or is there a better way to handle it?

Now I'm also getting this:

ActionView::MissingTemplate: Missing template blogs/index with {:formats=>["text/*"], :handlers=>[:rjs, :haml, :rhtml, :erb, :rxml, :builder], :locale=>[:en, :en]} in view paths "/var/www/keeponposting/releases/20110415040109/app/views"

Isn't there a way to send back HTML no matter what format is requested?

Answer 1

I agree about blocking the offending robot, but if you really want to force the response format add a before_filter and set request.format = :html, like this:

before_filter :force_request_format_to_html

private

def force_request_format_to_html
  request.format = :html
end

Answer 2

I fixed this issue (a few minutes ago – so far, so good) with this new Rails 3.1 option:

config.action_dispatch.ignore_accept_header = true

As mentioned in this Rails issue. That goes in config/application.rb.

I tested it like so in a RSpec request test (using Capybara):

it "should not break with HTTP_ACCEPT image/*;w=320;h=420 from iPhone" do
  page.driver.header "Accept", "image/*;w=320;h=420"
  visit "/some/path"
  page.should have_content("Some content")
end

Answer 3

Here is a stricter response; a suggestion by purp, from the discussion on issue 4127.

class FooController
  rescue_from  ActionView::MissingTemplate, :with => :missing_template

  def missing_template
    render :nothing => true, :status => 406
  end
end

Answer 4

I would be tempted to rescue the MissingTemplate in your application controller and log the Referrer header to see what's triggering this request. You never know, it might be some obscure part of your own app!

If on the other hand you're confident this is being caused by a robot, have you considered adding the offending URL to your robots.txt file? For example:

User-Agent: YandexImages
Disallow: /your/failed/path

Replacing 'your/failed/path' with the path that the robot is stumbling over. If the robot is struggling all over the place, you could just disallow access to the whole site for that particular robot:

User-Agent: YandexImages
Disallow: /

I think this is a cleaner and lighter approach than implementing a handler specifically to suppress errors from a seemingly badly behaved bot.