Webrat Webrat → 0.5

Comments and changes to this ticket

• 

  Bryan Helmkamp February 8th, 2009 @ 11:41 PM

  • Assigned user set to “Bryan Helmkamp”
  • State changed from “new” to “open”
  • Milestone set to “0.5”

  Zach,

  I agree it should be supported. At first glance, I'm wondering if we should add within_xpath instead of overloading the within API to take either style. On the whole, I explicit to implicit in the Webrat API, and am trying to move in that direction.

  In either form, I'd like to get this into the 0.5 release.

  WDYT?

  -Bryan

• You flagged this item as spam.
  

  Zach Dennis February 9th, 2009 @ 08:33 PM

  I agree with you Bryan, being explicit seems like a better route to go. If it becomes a need to have a single consolidated interface it can always be revisited that later.

• You flagged this item as spam.
  

  tomtt March 3rd, 2009 @ 06:53 AM

  The problem with just adding within_xpath is that methods like click_link_within will still not be able to use xpath. I do not know the codebase very well, but it seems to me that by allowing the selector to be a string (assuming it is a css selector) or an XPath object and passing that selector all the way down, then all that needs to be done is not convert the selector to an XPath object if it already is one. Am I right?

• You flagged this item as spam.
  

  tomtt March 3rd, 2009 @ 08:20 AM

  Here is a duck punch that solved my immediate need. For a real patch, it should check if selector is an XPath object.

  
  # Webrat does not accept xpath selectors. This is a quick hack to make
  # it work for me (assumes xpaths start with /, everything else is css)
  
  module Webrat
    class Scope
      protected
  
      def scoped_dom
        if @selector =~ /^\.?\/\/?/
          Webrat::XML.xpath_at(@scope.dom, @selector)
        else
          Webrat::XML.css_at(@scope.dom, @selector)
        end
      end
    end
  end
  

• You flagged this item as spam.
  

  tomtt March 4th, 2009 @ 09:07 AM

  Another duckpunch which is slightly more elegant I believe. It tries to call css_at with the selector, if it raises a syntax error the selector is passed to xpath_at. If that fails then the original exception is raised.

  My knowledge of css vs xpath selectors is limited so this may be a Very Bad Idea.

  
  # Webrat does not accept xpath selectors. This is a quick hack to make
  # it work for me (assumes xpaths start with //, everything else is css)
  
  module Webrat
    class Scope
      protected
  
      def xpath_scoped_dom
        Webrat::XML.xpath_at(@scope.dom, @selector)
      end
  
      def scoped_dom
        begin
          Webrat::XML.css_at(@scope.dom, @selector)
        rescue Nokogiri::CSS::SyntaxError, Nokogiri::XML::XPath::SyntaxError => e
          begin
            # That was not a css selector, mayby it's an xpath selector?
            xpath_scoped_dom
          rescue
            # Raise original css syntax error if selector is not xpath either
            raise e
          end
        end
      end
    end
  end
  

• You flagged this item as spam.
  

  tomtt March 16th, 2009 @ 02:22 PM

  @linoj: you are right, i had fixed it in my code, but thought it not worth another ticket update. But here you go:

  
  # Webrat does not accept xpath selectors. This is a quick hack to make
  # it work for me. It tries to call css_at with the selector, if it
  # raises a syntax error the selector is passed to xpath_at. If that
  # fails then the original exception is raised.
  
  module Webrat
    class Scope
      protected
  
      def xpath_scoped_dom
        Webrat::XML.xpath_at(@scope.dom, @selector)
      end
  
      def scoped_dom
        begin
          Webrat::XML.css_at(@scope.dom, @selector)
        rescue Nokogiri::CSS::SyntaxError, Nokogiri::XML::XPath::SyntaxError => e
          begin
            # That was not a css selector, mayby it's an xpath selector?
            xpath_scoped_dom
          rescue
            # Raise original css syntax error if selector is not xpath either
            raise e
          end
        end
      end
    end
  end
  

• You flagged this item as spam.
  

  tomtt March 18th, 2009 @ 09:01 PM

  I have forked webrat on github and committed this fix for it. It has been working well for me in practice. The only caveat I can think of is if you'd want to use an xpath expression that is also a valid css selector. I can not currently think of one though. If somebody can give me an example of an html fragment and a string that can be both a valid css selector and a valid xpath but that should produce different results then I will find a way to code for that as well.

  http://github.com/tomtt/webrat/t... 61d859b2ffcb208d6b92d4d2f00b1a880b7899a8

• You flagged this item as spam.
  

  tomtt December 2nd, 2009 @ 08:26 AM

  Thanks for that, there is one minor tweak I made: just raising does not halt the exception chain. I have updated my github webrat fork accordingly:
  http://github.com/tomtt/webrat/tree/xpath_fix
  http://github.com/tomtt/webrat/commit/bbdd1c1cd6cd374afe31255c5fcbb...

• You flagged this item as spam.
  

  tomtt December 2nd, 2009 @ 08:45 AM

  In fact, that was probably not the right thing to do. Sorry, please ignore that commit, you need to reraise the original exception to not break apps that catch it. I have added a spec to check that the correct type of error is raised.

  http://github.com/tomtt/webrat/commit/2aeb68899edd7abdc77fedb7c552b...

• You flagged this item as spam.
  

  Deleted User January 5th, 2010 @ 07:14 PM

  See also
  http://objectliteral.blogspot.com/2009/09/clicking-on-links-in-webr...

  For 0.6.0 I am using
  
module Webrat
class Scope

  {mkd-extraction-90c8d48709e810f68642dca8ed82a220}

  end end
  

• You flagged this item as spam.
  

  Deleted User January 5th, 2010 @ 07:15 PM

  Sigh. Preview does not work. How about plain text:

  module Webrat
  class Scope

  protected
  
  def xpath_scoped_dom
    #HaveXpath.new(@selector).matches.first
    hs = have_xpath(@selector)
    hs.matches(@scope.dom)
    #Webrat::XML.xpath_at(@scope.dom, @selector)
  end
  
  def scoped_dom
    begin
      @scope.dom.css(@selector).first
      #Webrat::XML.css_at(@scope.dom, @selector)
    rescue Nokogiri::CSS::SyntaxError => e#, Nokogiri::XML::XPath::SyntaxError => e
      begin
        # That was not a css selector, mayby it's an xpath selector?
        xpath_scoped_dom
      rescue Exception => e2
        # Raise original css syntax error if selector is not xpath either
        raise e2
      end
    end
  end
  

  end end

• You flagged this item as spam.
  

  dannyhawkins (at me) January 12th, 2010 @ 11:49 AM

  I've just added the following:
  

  def scoped_dom
  

    if @selector =~ /^\.?\/\/?/
      @scope.dom.xpath(@selector).first
    else
      @scope.dom.css(@selector).first
    end
  end
  

  Which seems to work very well. Any reason why this shouldn't be used?