Webrat Webrat → 0.5

At the moment SeleniumSession#fill_in escapes the field_identifier:

def fill_in(field_identifier, options)
  locator = "webrat=#{Regexp.escape(field_identifier)}"
  selenium.wait_for_element locator, 5
  selenium.type(locator, "#{options[:with]}")
end

The webrat location strategy tries the id, name and label location strategy. But the name location strategy does not expect a regular expression. Thus common input names like 'post[title]' are turned into 'post\[title\]' and are not recognized.

The attached patch is incomplete because it removes Regexp.escape only at this one position, even though there are other places in the file where the same problem will come up. It is meant as a hint for people running into issues with this behaviour.