I'm new to WATIR testing (and do I love it!) and have run into a wall on how to refocus my WATIR script to a newly opened window.. Here's my (simplified) script....
#!/usr/bin/ruby
require 'rubygems'
require 'watir-webdriver'
browser=Watir::Browser.new
browser.goto("http://0:3050")
browser.text_field(:name,"userkey300203830").set("OKJHNB")
browser.button(:id, "interview48").click
puts "Expected Result:"
puts "A successful display of cars"
if browser.window(:title=>"300203830").exists?
puts " Test passed. New window opened!"
else
puts " Test Failed! No window found"
end
It all works right up to the end. After the key "interview48" is clicked, a new window is opened with the title "300203830". Looks like I find it but I just don't know how to now focus on that window.
browser.window(:title => "300203830").use do
# do something
end
More information: http://watir.github.io/docs/browser-popups/
Additionally for more than 2 windows you can use the following:
browser.windows[n].use
#n is variable for which window. n will access them in order of opened or tabs from left to right
browser.windows.last.use
browser.windows.first.use
You can use the above commands if you open a new window from first browser instance and would like to toggle between the two.
There are 3 primary selectors for windows:
:title - typically the easiest
:url - often used with a Regexp value
:element - a unique element might be the least brittle (new as of Watir 6.18!)
browser.window(title: 'new window')
browser.window(url: /my_page.html/)
browser.window(element: browser.div(id: 'my-element'))
Locating by index is no longer supported
More information: Watir Browser Windows
If there is only one other window you want to use, so as of Watir 6.18, the easiest way to work with that window is with Browser#switch_window. It can only be used if there are only 2 windows, and all it does is switch to the other one, no additional locating required.
browser.switch_window
Related
I'm writing a script to cycle through some links and print to PDF.
When you click on one of the links a new window pops up of the document to be printed and on top of that new window Chrome's print preview box pops up.
I'm trying to change the printer and then click the print button.
What's happening with my code however is that watir is only looking at the window html and not the print preview html so the elements are unable to be located.
My code looks like this:
begin
browser.window(title: 'Window Title Name').use do
browser.div(:id => "navbar-scroll-container").button(:class => "destination-settings-change-button").click
sleep(1)
browser.element(:title => "CutePDF Writer").click
browser.button(:visible_text => 'Print').click
end
rescue Watir::Exception::NoMatchingWindowFoundException
retry
end
I have tried using just "browser.windows.last.use do" instead of the title name but that also doesn't work.
I know it can only see the underlying window's html because when I puts browser.html the output is that underlying windows html.
If you want to look at the html situation just Ctrl+P this page (in chrome) to see what I mean.
Of course, after posting a question I soon found a sort of answer; basically chrome opens 2 browser windows and the print preview one couldn't be accessed by title (because it had the same title as the underlying one) but could be accessed by browser.window(:index => 1).use do. not sure why browser.windows.last.use do didn't work. Also following that I would have thought browser.window(:index => 2).use do would have been the correct one as index 0 would be the originally opened one, index 1 the popup and index 2 the print preview.
I've been doing some dabbling with test automation using Ruby and Watir webdriver...I am running into an oddity where I open a new window, but when I act upon it, it performs all my actions behind the parent windowHere's how I set it up (using Firefox if that makes a difference):
# opens new window
b.button(:text => 'Upload').click
b.window(:url => 'urlname'). use do
# action performed on the new window
b.select_list(:id => 'selector').select 'Foo'
end
Note: it correctly performs actions on the select list present, but it does this behind the main browser.Is there something specific I need to write in order for it to remain in front of the main window?
This should work:
b.window(:url => 'urlname').focus
Put that line just after your .use line, and you should see the new window called to the foreground. You may actually find that you don't end up needing the .use line (or the do/end) at all.
I'm writing tests using Selenium WebDriver and rautomation to handle system popup. I tried it on irb like following:
require 'selenium-webdriver'
require 'rautomation'
driver = Selenium::WebDriver.for :firefox
driver.get "http://rubygems.org/gems/rautomation-0.9.2.gem"
window = RAutomation::Window.new :title => "Opening rautomation-0.9.2.gem"
ok_button = window.button(:text => "&OK")
ok_button.exists?
cancel_button = window.button(:text => "&Cancel")
cancel_button.exists?
ok_button.exists? and cancel_button.exists? are returning false. Hence I can't click on the buttons.
I also tried:
window.buttons.length
to find number of buttons, but it's returning 0.
Could someone please help me why the buttons aren't detected using rautomation? Please correct me if I'm doing something wrong.
Here is a popup:
For my condition, I have to send two :tab key and then send :enter to save the file. like:
driver.get "http://rubygems.org/gems/rautomation-0.9.2.gem"
window = RAutomation::Window.new :title => /Opening/i
if window.exist?
window.activate
window.send_keys :tab;
sleep 2;
window.send_keys :tab;
sleep 2;
window.send_keys :enter
end
I don't know why I can't just save the file with:
window.activate; sleep 1; window.send_keys :enter
The problem with this dialog is that it does not use native Windows controls. When you use Spy++ or AutoIt Window Info Tool then they do not show you any controls in that window either.
When using RAutomation you can check if it has native controls on it or not like this:
win = RAutomation::Window.new :title => /Opening rautomation/
p win.present?
p win.controls.length
p win.text
win.close
The output of this script will be:
true
0
""
In other words - window was present, it had zero controls of any kind and text was an empty string. Also, closing the window really closed it which you can verify visually - this means that we were interacting with the correct window and not accidentally with some other empty window (beware: this might sometimes happen too).
This all means that you cannot interact with the controls directly with AutoIt, RAutomation or many other automation tools. There might be some specific automation tools available for handling these kind of dialogs - i'm not sure.
There is however a workaround how to work with these kind of windows - send needed keystrokes to the window. In this case, sending a return/enter key would do the trick since that is the same as clicking on the "OK" button - you can try that manually.
Here is example code, which works the same as clicking on the "OK" button:
win = RAutomation::Window.new :title => /Opening rautomation/
win.activate
sleep 1
win.send_keys :enter
I'm not sure why, but for some reason you have to activate the window manually by calling Window#activate and wait a second before sending that enter key.
After doing that a new dialog will pop up, which uses native Windows controls - you can handle that as you would have expected RAutomation to work in the first place.
However, if you would use a :ms_uia adapter instead of the default :win32 then you don't need to activate and sleep.
Here is a fully working example with :ms_uia adapter:
win = RAutomation::Window.new :title => /Opening rautomation/, :adapter => :ms_uia
win.send_keys :enter
file_dialog = RAutomation::Window.new :title => /Enter name of file/
file_dialog.button(:value => "&Save").click
To click "Cancel" on the first dialog instead of "OK" you can just use Window#close as i was using to test the window above.
I would recommend you to use :ms_uia adapter instead of :win_32 since it is getting more stable every day and will be a new default one in the far future.
To set :ms_uia adapter for default one you can use environment variable RAUTOMATION_ADAPTER before loading RAutomation itself like this:
ENV["RAUTOMATION_ADAPTER"] ||= :ms_uia
require "rautomation"
I do not see any popup when I click that link. Chrome just downloads a file. :) This could help: http://watirwebdriver.com/browser-downloads/
This code worked for me:
window = RAutomation::Window.new(:title => /Opening rautomation-0.9.2.gem/i)
window.activate
p window.exists? # => true
sleep 2
window.send_keys(:down)
window.send_keys(:enter)
I have multiple features when run together throw this error. If I run the scenarios by them self there is no issue. I think the issue is that popups are generated during the test run and are not closed properly. I have code in an After hook that closes all open windows except the very first window opened.
Error: Session [2c50a228-3ad7-a544-a6ca-5d173b86bc86] has no driver. The browser window may have been closed. (Selenium::WebDriver::Error::UnknownError)
I have added a bunch of print statements in my code to get the current state:
After:Session -> #<>Capybara::Session:0x00000100f811b8>
Before:Driver -> selenium
Before:Session Object -#<>Capybara::Session:0x00000100f811b8>
Before: (start) Driver Object #<>Capybara::Selenium::Driver:0x000001028ad790>
The scenario before this particular scenario fails, this HAS to be the culprit, but why? Can anyone point me in the right direction?
After hook
#assume ONLY last window opened is to be closed
page.driver.browser.switch_to.window(page.driver.browser.window_handles.last)
page.execute_script "window.close();"
#switch back to first window opened, make it the default window now
page.driver.browser.switch_to.window(page.driver.browser.window_handles.first)
Env:
capybara (2.0.3)
cucumber (1.1.9)
selenium-webdriver (2.29.0)
ruby 1.9.3p0
require 'selenium-webdriver'
driver = Selenium::WebDriver.for :firefox
driver.get "https://www.google.co.in/"
address = driver.find_element(:link_text, "Gmail").attribute('href')
driver.execute_script( "window.open()" )
p driver.window_handles.length
p driver.window_handles.first #=> "{f17eac79-daf9-4a6c-a1ff-1b524fef9faf}"
driver.switch_to.window( driver.window_handles.last )
driver.get address
driver.execute_script "window.close()"
driver.execute_script "window.close()"
# => Window not found. The browser window may have been closed. (Selenium::WebDriver::Error::NoSuchWindowError)
In the above code,I just tried to re-generate the error,and it happened.Error is very logical as I tried to close an already closed,non existent window.
Now I want to debug this way:
p driver.browser.window_handles.length # if this is 0,then below line obvious throw error.
page.driver.browser.switch_to.window(page.driver.browser.window_handles.last)
page.execute_script "window.close();"
p driver.browser.window_handles.length # if this is 0,then below line obvious throw error.
page.driver.browser.switch_to.window(page.driver.browser.window_handles.first)
Now debug and let me know your result.
The solution for this is to maximise the window at the first place. I've not been able to find a way to maximise the window yet. In the middle of the test i manually click the maximise browser button and the test runs perfectly.
https://github.com/fahenao/_bot
I'm using Watir-webdriver and I was wondering if there was a good way to check if a new window opens. I've googled around a bit and couldn't find anything though it feels like there should be an easy answer.
I have a printer friendly link and I want to test that the link opens in a new window or tab and I would like to test this with ie, firefox, chrome and safari if possible.
Thanks!
You can check the number of windows:
browser.windows.size
or check if a specific window exists:
browser.window(:title => "foo").exists?
More examples in the specs.
You can also use index based browser window checking where you need to worry about index only and it follows zero based index ordering. So, the default window is of index: 0 and if a new window opens it will be of index: 1, the next will be of index: 2 and so on.
To check first child window if you want to test that the link opens in a new window ,
browser.window(index: 1).exists?
Or to work inside this window,
browser.window(index: 1).use do
# do scripting here
end