How to use location method of Selenium.WebDriver Package

For me, #drag_to did work, however, its powers seem to be limited.

In order to move a UI-sortable table row down, I had to create a table with three rows, then run this code (in a Cucumber step):

element = find('tbody tr:nth-child(1)')
target = find('tbody tr:nth-child(3)')

element.drag_to target

This would swap the first with the second row. My interpretation is that Capybara does not drag far enough, so I gave it a target beyond my actual target.

Note: I have configured UI-sortable with tolerance: 'pointer'.

The methods are accessible in the first file. In essence, doing the require is the same as writing it all in the same file.

By doing just that, you can see part of your problem:

require "selenium-webdriver"

def save
  driver.save_screenshot('screenshot.png')
end

def hello
  puts "hello"
end

driver = Selenium::WebDriver.for :firefox
driver.navigate.to "http://www.google.com"

hello
save

The variable driver is not available inside of the method save.

One way around this is to let save take the driver as an argument:

def save(driver)
  driver.save_screenshot('screenshot.png')
end

# Call the method like this
save(driver)

Some points:

• Writing code in the global scope like this is not advisable. Try wrapping what you want to do in a class.
• The methods you've defined aren't really class methods. As @JörgWMittag pointed out: they become private instance methods of Object

For Java Only: In the latest version of WebClient (which is wrapped by HTMLUnitDriver) client.setThrowExceptionOnScriptError(false) method is deprecated. In case of subclassing HTMLUnitDriver, you need to override modifyWebClient method:

public class MyHtmlUnitDriver extends HtmlUnitDriver {

...

 @Override
    protected WebClient modifyWebClient(WebClient client) {
        //currently does nothing, but may be changed in future versions
        WebClient modifiedClient = super.modifyWebClient(client);

        modifiedClient.getOptions().setThrowExceptionOnScriptError(false);
        return modifiedClient;
    }
}

One more solution (in Java, because you didn't tell us your language - but it works the same way in all languages with Keys class):

String selectAll = Keys.chord(Keys.CONTROL, "a");
driver.findElement(By.whatever("anything")).sendKeys(selectAll);

You can use this to select the whole text in an <input>, or on the whole page (just find the html element and send this to it).

For using Selenium Ruby bindings:

There's no chord() method in the Keys class in Ruby bindings. Therefore, as suggested by Hari Reddy, you'll have to use Selenium Advanced user interactions API, see ActionBuilder:

    driver.action.key_down(:control)
                 .send_keys("a")
                 .key_up(:control)
                 .perform

It sounds like you may not be using rspec at all, but just want the retry behavior that is provided by the rspec-retry gem. If that's the case, you could just use ruby's rescue and retry keywords to achieve the same thing.

For example, the following code will retry the navigation 1 time if the navigation throws a Net::ReadTimeout

require 'selenium-webdriver'

Selenium::WebDriver::PhantomJS.path = 'phantomjs.exe'
driver = Selenium::WebDriver.for :phantomjs
driver.manage.timeouts.page_load = 300

attempts = 0  # has to be outside the begin/rescue to avoid infinite loop
begin
  driver.navigate.to "http://google.com"
rescue Net::ReadTimeout => e
  if attempts == 0
    attempts += 1
    retry
  else
    raise
  end
end

All you have to do is just to disable the W3C when initializing the webdriver

options = webdriver.ChromeOptions()
options.add_experimental_option('w3c', False)
create_webdriver('Chrome', options=options)

Environment:

• Chrome 75
• ChromeDriver 75

The action triggered.by click_on can occur asynchronously. Therefore, if you do a visit immediately after it can cause the login request to abort and the session cookies never get set. To solve that you need to check for page text/content that indicates the login has succeeded. Something like

expect(page).to have_text 'You are now logged in'

I was not sure of the format for use with the attribute method, but after some experimenting I was able to move past this step.

@countries = Array.new

@browser.navigate.to "http://www.testingdomain{$env}.com/global"

@browser.find_elements(:xpath, "//table[@class='global-list']//a").each do |link|
text = link.attribute("text")
href = link.attribute("href")
@countries << [text, href]
end #links
`@countries.uniq! #DEBUG for false CMS content

It's HTTP mocking, if you remove fakeweb or webmock from your environment (entirely), it should all work again.

The last comment by Adam Greene DOES WORK regarding setting up Curb with: Selenium::WebDriver.for :firefox, :http_client => Selenium::WebDriver::Remote::Http::Curb

Read the thread on the Capybara group.

The problem we're having is playing back recorded http traffic using fakeweb or webmock since web driver is now Curb. So if you're goal was to fake out traffic over Capybara, you'll get browser testing to work again but you won't be able to play the traffic back over the same browser. (We're using VCR to record.)

Adding Curb support is listed as a "ticket" on the Fakeweb's Github Issues site.