[SOLVED] clicking link results in empty page

clicking link results in empty page

I have made a capybara script that either runs in headless mode or in browser mode. It tries to create a page in a wiki. It does so by clicking 'Add' in the menu bar and then 'Page' in the dropdown that is opened.

This works in browser mode. But in headless mode (webkit), after the link 'Page' is clicked, an empty page is returned. Who can tell me why this can happen?

In my code, I have this

   click_link 'Add'
    if ENV['BROWSER'] == 'headless'
        wait_for_ajax
        verbose( "headless add page", 3)
        p page.html
        lnk = all( :css, "#createPageLink").first
        p lnk
        wait_for_ajax
        lnk.click
      else
        verbose( "klik Page", 3)
        click_link 'Add a Page'
        lnk = all( :css, "#createPageLink").first
        p lnk
        lnk.click
    end
    wait_for_ajax
    p page.html

This output in headless mode

PAGEHTML</div></body></html>"
#<Capybara::Node::Element tag="a" path="/html/body[@id='com-atlassian-confluence']/div[@id='full-height-container']/div[@id='splitter']/div[@id='splitter-content']/div[@id='main']/div[@id='main-header']/div[@id='navigation']/ul/li[3]/div/ul[@id='add-menu-link-space']/li[1]/a[@id='createPageLink']">
""

So the link is found (p lnk shows Capybara::Node::Element), but the click on the links returns an empty page, whereas when the link is clicked by the browser, I get the html that is behind the link.

I hope you can tell me what I am overlooking....

Ruud

Solution

Capybara-webkit is obsolete and basically equivalent to an 8 year old browser. Most likely it no longer supports the JS & CSS being used on the page you are interacting with. If you need headless support you're going to be much better off not using capybara-webkit and instead using selenium with chrome in headless mode, or one of the direct to Chrome via CDP drivers (like apparition) for interacting with any site written/updated in the last few years.