rspec users - Jul 2007 - Spec for validation plugin

Hi,

I''m writing a specification for a model called Location which has a
post_code attribute which should be a valid UK postcode. I''m using the 
http://svn.designbyfront.com/rails/plugins/validates_as_uk_postcode/
validates_as_uk_postcode  plugin to handle the validation but I''m not
sure
how I should properly specify my model.

Heres the model code:

class Location < ActiveRecord::Base
    validates_as_uk_postcode :post_code
end

In my model spec I want something like this (assume @location has been setup
as a new Location instance):
 
  it "should be invalid if the post code is incorrectly formatted" do
    @location.post_code => ''Q1 1QQ''
    @location.should_not be_valid
    @location.should have(1).error_on(:post_code)   
  end

But there are many examples of invalid post codes and they are already
tested in the plugin''s own test cases. I only really want to test my
code,
rather than retest the plugin.

So my question is how should I write the specification so that it
doesn''t
duplicate the tests in the plugin but at the same time expresses the correct
behaviour of the Location model?

Thanks in advance,
Rob
-- 
View this message in context:
http://www.nabble.com/Spec-for-validation-plugin-tf4152378.html#a11812924
Sent from the rspec-users mailing list archive at Nabble.com.

On 7/26/07, rob_twf <rob.anderton at thewebfellas.com>
wrote:
>
> Hi,
>
> I''m writing a specification for a model called Location which has
a
> post_code attribute which should be a valid UK postcode. I''m using
the
> http://svn.designbyfront.com/rails/plugins/validates_as_uk_postcode/
> validates_as_uk_postcode  plugin to handle the validation but I''m
not sure
> how I should properly specify my model.
>
> Heres the model code:
>
> class Location < ActiveRecord::Base
>     validates_as_uk_postcode :post_code
> end
>
> In my model spec I want something like this (assume @location has been
setup
> as a new Location instance):
>
>   it "should be invalid if the post code is incorrectly
formatted" do
>     @location.post_code => ''Q1 1QQ''
>     @location.should_not be_valid
>     @location.should have(1).error_on(:post_code)
>   end
>
> But there are many examples of invalid post codes and they are already
> tested in the plugin''s own test cases. I only really want to test
my code,
> rather than retest the plugin.
>
> So my question is how should I write the specification so that it
doesn''t
> duplicate the tests in the plugin but at the same time expresses the
correct
> behaviour of the Location model?
Here are some interesting thoughts on that:

http://blog.jayfields.com/2006/12/rails-unit-testing-activerecord.html

Just apply the same thinking to rspec.
>
> Thanks in advance,
> Rob
> --
> View this message in context:
http://www.nabble.com/Spec-for-validation-plugin-tf4152378.html#a11812924
> Sent from the rspec-users mailing list archive at Nabble.com.
>
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users
>

David Chelimsky-2 wrote:
> 
> 
> Here are some interesting thoughts on that:
> 
> http://blog.jayfields.com/2006/12/rails-unit-testing-activerecord.html
> 
> Just apply the same thinking to rspec.
> 
> 
Thanks for the pointer, here''s what I''ve done:

In spec_helper.rb:

class << ActiveRecord::Base  
  def source_file    
   
File.expand_path("#{RAILS_ROOT}/app/models/#{self.name.underscore}.rb")
  end
end

(I preferred source_file to standard_path - seemed more descriptive)

An then my specification looks like this:

  it "should validate that the postcode is correctly formatted" do  
    Location.should_receive(:validates_as_uk_postcode).once.with(:post_code)
    load Location.source_file
  end

That works nicely.

I suppose the downside is it allows implementation to creep into the specs
but if it saves me having to re-test somebody else''s code I can live
with
it!

I also looked at 
http://spicycode.com/2007/4/2/rspec-expecation-matchers-part-ii
SpicyCode''s
RSpec extensions  - the approach here appears to be similar to the code in
my original example as it tests validation with a single invalid example and
if that returns the expected result then assumes the validation will work
for all invalid examples (I may be doing it a disservice here as I''ve
only
quickly looked at the source code). This still leaks implementation into the
specs and I think I prefer the Jay Fields approach as it, in my opinion,
more clearly specifies the expected behaviour of the model.

Thanks again,
Rob





-- 
View this message in context:
http://www.nabble.com/Spec-for-validation-plugin-tf4152378.html#a11827396
Sent from the rspec-users mailing list archive at Nabble.com.

rob_twf wrote:
> 
> 
>   it "should validate that the postcode is correctly formatted"
do
>    
> Location.should_receive(:validates_as_uk_postcode).once.with(:post_code)
>     load Location.source_file
>   end
> 
> That works nicely.
> 
> 
Looks like I was a bit eager. Loading the class in this way causes problems
with other validations. For example, I add a name attribute to my class and
validate it''s length like this:

class Location < ActiveRecord::Base 
    validates_length_of :name, :in => 1..50
    validates_as_uk_postcode :post_code 
end 

And have this additional specification:

  it "should be invalid if the name is too long" do
    @location.name => ''x'' * 51
    @location.should_not be_valid
    @location.should have(1).error_on(:name)
  end  
  (
This fails because the length validation occurs twice giving two (identical)
errors on :name. Removing the post code validation specification allows the
name length validation to work correctly. 

So load() is causing the validation to be defined twice - is there any way
to avoid this?
-- 
View this message in context:
http://www.nabble.com/Spec-for-validation-plugin-tf4152378.html#a11828149
Sent from the rspec-users mailing list archive at Nabble.com.