gmane.comp.lang.ruby.rspec.user

Re: shared examples with a block for expects?

Justin Ko — 2012-05-24T06:33:39

On May 23, 2012, at 6:12 PM, Patrick J. Collins wrote:


Just use a Proc:

describe '#new' do
  it_behaves_like 'an access controlled action', -> { get :new }
end

Re: received unexpected message :id= with (20)

Justin Ko — 2012-05-24T06:28:26

On May 23, 2012, at 4:27 PM, Kleber Shimabuku wrote:


I'm guessing the `Post` mock is making its way to the model layer. Either way, you need to find out why `id=` is getting called.

Shot in the dark: There is a lot of "post" names there. Maybe there is a conflict?

shared examples with a block for expects?

Patrick J. Collins — 2012-05-24T00:12:54

Hi,

Basically I want to do something like this:

describe FancyThingsController do

  def do_new
    get :new
  end

  def do_create
    post :create, :fancy_thing => { :some => "params" }
  end

  shared_examples "an access controlled action" do |expectation|
    it "denies access" do
      expect(&expectation).to raise_error(CanCan::AccessDenied)
    end
  end

  describe "#new" do
    it_behaves_like "an access controlled action", proc { do_new }
  end

  describe "#create" do
    it_behaves_like "an access controlled action", proc { do_create }
  end

end

but I get:  "expected CanCan::AccessDenied, got #<NameError: undefined
local variable or method `do_create' for
RSpec::Core::ExampleGroup::Nested_1::Nested_2:Class>"

Just curious, how can I do this?

Patrick J. Collins
http://collinatorstudios.com

received unexpected message :id= with (20)

Kleber Shimabuku — 2012-05-23T22:27:45

Please help me to understand this error message:

 1) PostsController#create when logged in should create a new post and
redirect to a successful message page
     Failure/Error: post 'create', { post: {title: 'New post',
       Mock Post(id: integer, title: string, description: text,
location: string, created_at: datetime, updated_at: datetime, status:
string, tags: string, views: integer, user_id: integer) received
unexpected message :id= with (20)
     # ./spec/controllers/posts_controller_spec.rb:56:in `block (4
levels) in <top (required)>'

Finished in 0.59991 seconds
6 examples, 1 failure



I'm trying to figure out this for about a week or more :(

Attachments:
http://www.ruby-forum.com/attachment/7429/posts_controller_spec.rb
http://www.ruby-forum.com/attachment/7430/posts_controller.rb

Re: bug: rspec rails generator generates a failingspec

Justin Ko — 2012-05-19T20:09:20

On May 19, 2012, at 11:46 AM, survivorist wrote:


I believe this is fixed in RSpec 2.10

Re: bug: rspec rails generator generates a failingspec

survivorist — 2012-05-19T17:46:39

What was the specific issue with this and how did you resolve it, I'm having
the exact same error and can't find the root of the issue?  Thanks

--
View this message in context: http://ruby.11.n6.nabble.com/bug-rspec-rails-generator-generates-a-failing-spec-tp4647101p4975461.html
Sent from the rspec-users mailing list archive at Nabble.com.

Re: How does Spork help in requests specs?

Rodrigo Rosenfeld Rosas — 2012-05-18T00:45:24

Indeed, deleting the tables manually requires more effort because of the 
foreign keys that will require a specific order, but the specs got way 
faster than using truncation. Go figure it out.

And worse than that is reading this on PG documentation:

http://www.postgresql.org/docs/9.1/static/sql-truncate.html

"TRUNCATE quickly removes all rows from a set of tables. It has the same 
effect as an unqualified DELETE on each table, but*since it does not 
actually scan the tables it is faster*."

Yeah, sure, I believed on you once. Maybe this is true for large tables, 
but not so true for testing matters.

I just found that I should report this here for benefits of others that 
might consider this approach as well on PostgreSQL.

Best,
Rodrigo.

Em 17-05-2012 21:09, Rodrigo Rosenfeld Rosas escreveu:

Re: How does Spork help in requests specs?

Rodrigo Rosenfeld Rosas — 2012-05-18T00:09:24

Sorry about this, I'm further investigating the slowness issue, and I 
found the culprit: truncation of my PostgreSQL are very slow. I'll 
investigate how to make it faster. Maybe deletion could be faster than 
truncation on PG...

Cheers,
Rodrigo.

Em 17-05-2012 21:04, Rodrigo Rosenfeld Rosas escreveu:

How does Spork help in requests specs?

Rodrigo Rosenfeld Rosas — 2012-05-18T00:04:23

Even with Spork, my requests specs are very slow to start running (about 
7 seconds).

I suspect Rails is booting each time I run "rspec -X spec/requests".

Is that true? If so, is there any way I could instruct the web server to 
keep alive after the specs run so that it would be faster on next run?

Are there any resources on how to have better performance on running 
requests specs with Capybara and Webkit?

Having to wait about 7 seconds between consecutive runs is really a 
blocker for me. If I can't get it to run faster, I'll give up on 
requests specs and only have the controller's tests and client-side code 
specs without any integration tests...

But to be honest, I feel much more comfortable having some integration 
tests for the main features of my application and I would appreciate any 
hints on making them start faster.

The last time I worked with Rails, long ago, it wasn't such a painful 
experience to write such kind of tests (Capybara + Webkit). And my 
current desktop PC is *way* faster tha

Bash shim for Rspec

Peter Fitzgibbons — 2012-05-16T09:56:53

HI Folks,
The following script is *supposed* to run rspec under Jruby with the proper
Java-switch for using SWT with OSX (-XstartOnFirstThread).
The issue is that this script doesn't actually run any of the specs!
Example:
$ bin/rspec_auto_jvm spec/swt_shoes/app_spec.rb
spec/swt_shoes/app_spec.rb
args are  spec/swt_shoes/app_spec.rb
$

Can anyone point out the flaw?

Peter Fitzgibbons
(847) 859-9550
Email: peter.fitzgibbons-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org
IM GTalk: peter.fitzgibbons
IM AOL: peter.fitzgibbons-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org


#!/usr/bin/env bash
# !/usr/bin/env jruby --1.9 --debug
#
# This file was generated by Bundler.
#
# The application 'rspec' is installed as part of a gem, and
# this file is here to facilitate running it.
#

echo "$< at >" | grep 'swt'

if [[ $< at > == *swt* ]]; then
  swt_opt="-J-XstartOnFirstThread"
fi

echo 'args are ' $*

run_rspec="
# http://jira.codehaus.org/browse/JRUBY-6324
# random seed for srand is not initialized properly
# call once and throw-away nil

Re: overview of how rspec was designed

David Chelimsky — 2012-05-15T17:41:24

The RSpec Book has a fairly deep discussion on the inner workings.
Other than that, there are some things you can get from the docs, but
there is not a central "this is how RSpec works" document that I've
written or am aware of.

Re: how can I do an "or" within the context of amatcher?

Justin Ko — 2012-05-15T16:03:52

On May 15, 2012, at 12:47 AM, Patrick J. Collins wrote:


You need a custom matcher, something like this:

RSpec::Matchers.define :have_link_or_button do |locator|
  match do |page|
    page.has_button?(locator) || page.has_link?(locator)
  end
end

BTW, I believe Capybara has a method that checks for a button *or* link - look into that.

Re: how can I do an "or" within the context of amatcher?

Rodrigo Rosenfeld Rosas — 2012-05-15T14:18:05

You're looking for Custom Matchers:

https://github.com/dchelimsky/rspec/wiki/Custom-Matchers

Em 15-05-2012 03:47, Patrick J. Collins escreveu:

Bash shim for Rspec

Peter Fitzgibbons — 2012-05-15T11:32:05

HI Folks,
The following script is *supposed* to run rspec under Jruby with the proper
Java-switch for using SWT with OSX (-XstartOnFirstThread).
The issue is that this script doesn't actually run any of the specs!
Example:
$ bin/rspec_auto_jvm spec/swt_shoes/app_spec.rb
spec/swt_shoes/app_spec.rb
args are  spec/swt_shoes/app_spec.rb
$

Can anyone point out the flaw?

Peter Fitzgibbons
(847) 859-9550
Email: peter.fitzgibbons-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org
IM GTalk: peter.fitzgibbons
IM AOL: peter.fitzgibbons-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org


#!/usr/bin/env bash
# !/usr/bin/env jruby --1.9 --debug
#
# This file was generated by Bundler.
#
# The application 'rspec' is installed as part of a gem, and
# this file is here to facilitate running it.
#

echo "$< at >" | grep 'swt'

if [[ $< at > == *swt* ]]; then
  swt_opt="-J-XstartOnFirstThread"
fi

echo 'args are ' $*

run_rspec="
# http://jira.codehaus.org/browse/JRUBY-6324
# random seed for srand is not initialized properly
# call once and throw-away nil

how can I do an "or" within the context of a matcher?

Patrick J. Collins — 2012-05-15T06:47:58

Capybara has two methods:

page.has_button?

and

page.has_link?

...  I am wondering, how can I make a helper method that will be
satisfied by one or the other?  In other words, I want to be able to do:

page.should have_link_or_button("blah")

def have_link_or_button(locator)
  have_link(locator) || have_button(locator)
end

But that doesn't work...  Is there any way to do this?

Patrick J. Collins
http://collinatorstudios.com

overview of how rspec was designed

S Ahmed — 2012-05-14T19:23:31

Other than jumping into the codebase myself, I was wondering if anyone has
done a high-level (or better yet low level) write-up on how rspec works
under the covers?

Re: Stubs and contract tests

David Chelimsky — 2012-05-11T16:48:54

On Fri, May 11, 2012 at 9:24 AM, Alexandre de Oliveira <
chavedomundo-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:


See https://github.com/rspec/rspec-mocks/issues/15

Stubs and contract tests

Alexandre de Oliveira — 2012-05-11T14:24:19

Hey guys, could you help me (in)validating an idea?

As you might know, Contract tests solve a problem stubs have: if the object
interface changes, the stubbed tests will continue passing.

In RSpec, it's common to use a shared_context to describe contracts, so
both Posts and Comments tests would have PostsCommentsContract. If
Comment's interface ever changed, Posts tests would present a failure
wherever we ran it. However, it won't tell us which stubbed methods were
invalid now; it just tell us the file that have invalid stubs.

If we had, for example, #stub_contract(:method), automatically raising
a failure if its reference (double) object's interface changed, how would
it be bad?

I'm trying to think the cons of it. Could you guys help me finding it?

--Alexandre

Re: how do I test for a status code in a :js => truetest?

David Chelimsky — 2012-05-10T11:45:23

On Wed, May 9, 2012 at 11:14 PM, Patrick J. Collins
<patrick-nwXFkk9SH4z0FiqB0faro5qQE7yCjDx5< at >public.gmane.org> wrote:

Not sure. I'd post this to the Capybara list.

how do I test for a status code in a :js => true test?

Patrick J. Collins — 2012-05-10T04:14:38

I have a view with some javascript that does page redirection, and I
wanted to confirm that under certain circumstances the user will get a
500 error from CanCan::AccessDenied.

It seems that when I am not using :js => true, I can do:

page.status_code.shoud be 500

but.. when I try that in a :js => true example group, I get:
Capybara::NotSupportedByDriverError: Capybara::NotSupportedByDriverError

also the request and response objects are nil in the context of this
test...  My dirty workaround has just been to do:

page.has_content "CanCan::AccessDenied"

Which serves my purposes, but I was hoping for a better way...  Any
ideas?

Thank you.

Patrick J. Collins
http://collinatorstudios.com

Re: is there a way to turn on :js => true in the middle of a spec?

David Chelimsky — 2012-05-09T20:44:25

On Wed, May 9, 2012 at 3:09 PM, Patrick J. Collins
<patrick-nwXFkk9SH4z0FiqB0faro5qQE7yCjDx5< at >public.gmane.org> wrote:

iiuc, browser is the issue, not headless. The login
token/cookie/whatever needs to be stored in the client so it can
identify itself to the server. What your proposing switches the
client.