broken after upgrade

[Perry Smith]

Mostly I'm looking for help on how to debug this... I suppose this could be a bug capybara-webkit but I'd be surprised.  I'm not doing anything really odd.

I have one test that I run via cucumber.  I can add @webkit to the feature and run it with capybara-webkit or remove it and it runs it via selenium.

Before the upgrade, it worked both ways.

After the upgrade, I had to dork with the authentication part to make it work with selenium but that should not affect the path using webkit.

The part I changed is this step:

Given /^I am on the welcome page$/ do

  case page.mode.to_s

  when 'selenium'

    # r = page.driver.rack_server

    # visit("http://username:password@#{r.host}:#{r.port}#{welcome_path}")

    visit("http://username:password@localhost#{welcome_path}")

  when 'rack_test'

    encoded_login = ["ped...@us.ibm.com:lostgr8t"].pack("m*")

    page.driver.header 'Authorization', "Basic #{encoded_login}"

    visit(welcome_path)

  when 'webkit'

    encoded_login = ["ped...@us.ibm.com:lostgr8t"].pack("m*")

    page.driver.header 'Authorization', "Basic #{encoded_login}"

    visit(welcome_path)

  end

end

The test itself is pretty simple:

And /^I enter a defect into the swinfo form$/ do

  within('#swinfo_form') do

    fill_in('item', with: @current_adv.defect.name)

  end

end

When /^I hit the submit button$/ do

  within('#swinfo_form') do

    click_button('Submit')

  end

end

Then /^I should see the results$/ do

  within('table.upd_apar_defs tbody') do

    tr1 = find('tr:first-child')

    tds = tr1.all('td')

    tds[0].text.should == '1'

    tds[1].text.should == @current_adv.defect.name

    tds[2].text.should == "IV22222"

  end

end

If I save the page in after the fill_in, the text is not filled in.  If I save the page after the click button, the app is complaining about no route which I think is because the item is not filled in.

I know capybara 2 is more restrictive than before but it works with selenium.

Any ideas of what is going wrong or how to track it down?

Thank you,

Perry

[Matt Horan]

On Mon, Jan 14, 2013 at 05:40:29PM -0600, Perry Smith wrote:
> The part I changed is this step:
> >
> > Given /^I am on the welcome page$/ do
> > case page.mode.to_s
> > when 'selenium'
> > # r = page.driver.rack_server
> > # visit("http://username:password@#{r.host}:#{r.port}#{welcome_path}")
> > visit("http://username:password@localhost#{welcome_path}")
> >
> > when 'rack_test'
> > encoded_login = ["ped...@us.ibm.com:lostgr8t"].pack("m*")
> > page.driver.header 'Authorization', "Basic #{encoded_login}"
> > visit(welcome_path)
> >
> > when 'webkit'
> > encoded_login = ["ped...@us.ibm.com:lostgr8t"].pack("m*")
> > page.driver.header 'Authorization', "Basic #{encoded_login}"

Instead, try

page.driver.browser.authenticate('ped...@us.ibm.com', 'lostgr8t')

There was a change rolled into the 0.14.0 release which may be causing
an issue with the Authorization header, though I'm surprised it ever
actually worked that way.

--
Matt Horan ma...@matthoran.com http://matthoran.com/

[Perry Smith]

On Jan 14, 2013, at 8:02 PM, Matt Horan wrote:

> On Mon, Jan 14, 2013 at 05:40:29PM -0600, Perry Smith wrote:
>> The part I changed is this step:
>>>

>>> when 'webkit'
>>> encoded_login = ["ped...@us.ibm.com:lostgr8t"].pack("m*")
>>> page.driver.header 'Authorization', "Basic #{encoded_login}"
>
> Instead, try
>
> page.driver.browser.authenticate('ped...@us.ibm.com', 'lostgr8t')
>
> There was a change rolled into the 0.14.0 release which may be causing
> an issue with the Authorization header, though I'm surprised it ever
> actually worked that way.

That worked. Thanks.

FYI: I came from 0.13.0 to 0.14.1