Forum: RSpec Troubles with route_for

Hey guys. I'm having some trouble with a route spec. In routes.rb , I
have:
    map.connect 'foods/search/:name', :controller => :foods, :action =>
:search

foods_controller_spec.rb has: http://codepad.org/dg3FERKw
Unfortunately, that fails: http://codepad.org/lck4r1S0

After reading the rspec-rails 1.2.0 upgrade notes (http://tr.im/K1zx), I
haven't been able to figure out what I'm doing wrong. Any hints?

Thanks!
Nick

PS: I'm using 1.2.9 and Rails 2.3.4 .

Nick Hoffman wrote:
> Hey guys. I'm having some trouble with a route spec. In routes.rb , I
> have:
>     map.connect 'foods/search/:name', :controller => :foods, :action =>
> :search
> 
> foods_controller_spec.rb has: http://codepad.org/dg3FERKw
> Unfortunately, that fails: http://codepad.org/lck4r1S0
> 
> After reading the rspec-rails 1.2.0 upgrade notes (http://tr.im/K1zx), I
> haven't been able to figure out what I'm doing wrong. Any hints?
> 
> Thanks!
> Nick
> 
> PS: I'm using 1.2.9 and Rails 2.3.4 .
Hi Nick,
You need to add quotes while writing in routs.rb file.
So your route will be define as :
map.connect 'foods/search/:name', :controller => 'foods', :action => 
'search'
Now try running the code again.You will see it pass.

Amit Kulkarni wrote:
> Nick Hoffman wrote:
>> Hey guys. I'm having some trouble with a route spec. In routes.rb , I
>> have:
>>     map.connect 'foods/search/:name', :controller => :foods, :action =>
>> :search
>> 
>> foods_controller_spec.rb has: http://codepad.org/dg3FERKw
>> Unfortunately, that fails: http://codepad.org/lck4r1S0
>> 
>> After reading the rspec-rails 1.2.0 upgrade notes (http://tr.im/K1zx), I
>> haven't been able to figure out what I'm doing wrong. Any hints?
>> 
>> Thanks!
>> Nick
>> 
>> PS: I'm using 1.2.9 and Rails 2.3.4 .

> Hi Nick,
> You need to add quotes while writing in routs.rb file.
> So your route will be define as :
> map.connect 'foods/search/:name', :controller => 'foods', :action => 
> 'search'
> Now try running the code again.You will see it pass.

Thanks for the suggestion, Amit. I changed the symbols to strings. 
Unfortunately, the same error is still occurring: 
http://codepad.org/lck4r1S0

On Sun, Jan 10, 2010 at 10:58 PM, Nick Hoffman <lists@ruby-forum.com> 
wrote:

>
I'd recommend using the route_to matcher that was added in 1.2.9 
instead.

http://codepad.org/fLcxyA9N
http://rspec.rubyforge.org/rspec-rails/1.2.9/classes/Spec/Rails/Matchers.html#M000029

It's more reliable, and aligns better with the rspec matchers API.

Cheers,
David

David Chelimsky wrote:
> I'd recommend using the route_to matcher that was added in 1.2.9 
> instead.
> 
> http://codepad.org/fLcxyA9N
> http://rspec.rubyforge.org/rspec-rails/1.2.9/classes/Spec/Rails/Matchers.html#M000029
> 
> It's more reliable, and aligns better with the rspec matchers API.
> 
> Cheers,
> David

Thanks for that, David. I updated my spec (http://codepad.org/F828X7Fg). 
For some reason though, it's still failing: http://codepad.org/s65Ckubc

Just in case, this is what I have in routes.rb: 
http://codepad.org/rJfIhbM6

There must be some tiny detail that I'm missing here...

On 1/11/10 9:22 PM, Nick Hoffman wrote:
>> Cheers,
>> David
>>     
> Thanks for that, David. I updated my spec (http://codepad.org/F828X7Fg). 
> For some reason though, it's still failing: http://codepad.org/s65Ckubc
>
> Just in case, this is what I have in routes.rb: 
> http://codepad.org/rJfIhbM6
>
> There must be some tiny detail that I'm missing here...
>   
Yes, there must.  I think you've tried to twice define the route, and
the test shows you which one actually is being used.  I'll suggest
commenting out one of those routes at a time, and see whether the new
test results are illuminating.

Randy

Randy Harmon wrote:
> On 1/11/10 9:22 PM, Nick Hoffman wrote:
>>> Cheers,
>>> David
>>>     
>> Thanks for that, David. I updated my spec (http://codepad.org/F828X7Fg). 
>> For some reason though, it's still failing: http://codepad.org/s65Ckubc
>>
>> Just in case, this is what I have in routes.rb: 
>> http://codepad.org/rJfIhbM6
>>
>> There must be some tiny detail that I'm missing here...
>>   
> Yes, there must.  I think you've tried to twice define the route, and
> the test shows you which one actually is being used.  I'll suggest
> commenting out one of those routes at a time, and see whether the new
> test results are illuminating.
> 
> Randy

I swapped the order of the 2 lines in routes.rb to this:

    map.connect   'foods/search/:name', :controller => 'foods', :action 
=> 'search'
    map.resources :foods, :collection => {:search => :get}

and the spec passes now.

Thanks, guys!

By the way, does this spec:
    {:get => '/path'}.should route_to(...)
make this spec redundant?:
    params_from(:get, '/path').should == {...}

They read the same, but it feels like they each check one end of the 
route's translation.

On Tue, Jan 12, 2010 at 11:35 AM, Nick Hoffman <lists@ruby-forum.com> 
wrote:

> By the way, does this spec:
>    {:get => '/path'}.should route_to(...)
> make this spec redundant?:
>    params_from(:get, '/path').should == {...}
>

Yes - route_to checks both sides of the translation.

David Chelimsky wrote:
> On Tue, Jan 12, 2010 at 11:35 AM, Nick Hoffman <lists@ruby-forum.com> 
> wrote:
> 
>> By the way, does this spec:
>>    {:get => '/path'}.should route_to(...)
>> make this spec redundant?:
>>    params_from(:get, '/path').should == {...}
>>
> 
> Yes - route_to checks both sides of the translation.

Great, thanks for that, David! And thanks again for all of your sweat 
and blood with RSpec and RSpec-Rails!

On Wed, Jan 13, 2010 at 10:20 AM, Nick Hoffman <lists@ruby-forum.com> 
wrote:

> > Yes - route_to checks both sides of the translation.
>
> Great, thanks for that, David!


Thank Randy Harmon for this one.  It was he who recognized and solved 
the
problems with params_from by adding the route_to matcher.


> And thanks again for all of your sweat
> and blood with RSpec and RSpec-Rails!
>

It's a labor of love, but you're welcome :)

David Chelimsky wrote:
> On Wed, Jan 13, 2010 at 10:20 AM, Nick Hoffman <lists@ruby-forum.com> 
> wrote:
> 
>> > Yes - route_to checks both sides of the translation.
>>
>> Great, thanks for that, David!
> 
> 
> Thank Randy Harmon for this one.  It was he who recognized and solved 
> the
> problems with params_from by adding the route_to matcher.

Whoops, you're right. Thanks, Randy!

Hey guys. I have two different paths that lead to the same controller 
and action:
  map.connect 'foods/search/:name', :controller => 'foods', :action => 
'search'
  map.food    ':name',              :controller => 'foods', :action => 
'search'

Unfortunately, the spec for the second route fails because #route_to 
finds and uses the first route when generating the path for
  :controller => 'foods', :action => 'search', :name => 'almonds'

routes.rb: http://codepad.org/CjKRynr5
specs:     http://codepad.org/838oV3mW

Is there a solution to this problem?

Thanks,
Nick

I guess the distinction is that you generate food_path()'s based on
existing food names, and search comes from users - but you don't have a
distinct landing page for each food, hence no need for resource-based
routing for foods.

You could define a :show action in your controller, which simply calls
the 'search' action.

def show
    search
end

Then they'll have different controller-actions and different paths, yet
the implementations would be identical.  And the specs can describe them
distinctly.

I'd be curious if there are other ways to do this.

Randy