Question

I am using Watir to log into an application, push some buttons, etc… Basically the normal stuff that a person would use Watir for.

However, my problem is that there is one particular page that I need to test. It’s actually a dynamically-generated PDF and I need to get the actual binary data from it, so that I can load it using a certain gem that we’re using. This normally works with static PDF files because we can just use:

    open("http://site.com/something.pdf")

This works for static PDFs. However, for a dynamically generated one it doesn’t work because we are using Ruby to send the HTTP request and it is not aware of the headers/cookies/session that Watir is using. So instead of getting the actual PDF we get a login page.

Another thing we tried was to use Watir to get the PDF:

    @browser.goto "http://site.com/dynamic/thepdffile"
    @browser.text
    @browser.html

We tried getting the text or html from the page, but no luck because firefox creates a DOM when loading a pdf so the text is an empty string and the html is the DOM that firefox creates when viewing a pdf page. We need the raw HTTP response and there doesn’t seem to be a way to extract that.

So we need a solution for this and in my opinion we have these options:

1. Figure out a way to use “open” or similar method in Ruby, using the session from Watir.
2. Figure out how to use watir to get the binary http response from the PDF page.
3. Disable the pdf plugin (which doesn’t seem possible) such that the “save as” dialog appears.

Or if you have some other idea please share! Thanks in advance!

Answer 1

I figured out a solution.

In the profile for firefox you can set the plugin.scan.Acrobat to “999” which will effectively disable the PDF plugin.

    profile = Selenium::WebDriver::Firefox::Profile.new
    profile['plugin.scan.Acrobat'] = "999"
    b = Watir::Browser.new :firefox, :profile => profile