Mechanize users - Aug 2012 - preferred practice for comparing Mechanize::Page?

I''m writing rspec tests for a specialized subclass of Mechanize. 
I''d
like to test the values returned by generic.get() against
my_subclass.get().  I know that the following doesn''t work:

it ''should produce the same result'' do
   a = generic.get("http://www.google.com/")
   b = my_subclass.get("http://www.google.com/")
   a.should == b
end

...because (for among a few other problems) two equal() urls may not
compare as ==:

       Diff:
       @@ -1,5 +1,5 @@
        #<Mechanize::Page
       - {url #<URI::HTTP:0x00000107029948 URL:http://www.google.com/>}
       + {url #<URI::HTTP:0x000001047094c8 URL:http://www.google.com/>}

This may be too general a question, but what would constitute a
sensible equality test for two Mechanize::Page objects?

TIA.

On Jul 31, 2012, at 21:14, Robert Poor <rdpoor at gmail.com>
wrote:
> I''m writing rspec tests for a specialized subclass of Mechanize. 
I''d
> like to test the values returned by generic.get() against
> my_subclass.get().  I know that the following doesn''t work:
> 
> it ''should produce the same result'' do
>   a = generic.get("http://www.google.com/")
>   b = my_subclass.get("http://www.google.com/")
>   a.should == b
> end
> 
> ...because (for among a few other problems) two equal() urls may not
> compare as ==:
> 
>       Diff:
>       @@ -1,5 +1,5 @@
>        #<Mechanize::Page
>       - {url #<URI::HTTP:0x00000107029948
URL:http://www.google.com/>}
>       + {url #<URI::HTTP:0x000001047094c8
URL:http://www.google.com/>}
Red herring:

$ ruby -ruri -e ''p URI("http://google.com") ==
URI("http://google.com")''
true

(Apparently rspec is missing the super-useful object_id suppression that
unit_diff has had since its first release.  unit_diff, The Original Result
Diffing Tool? unit_diff, Accept No Substitutes?, etc.)

Neither Mechanize::Page nor Mechanize::File define #==, so two pages will never
be equal in released versions.
> This may be too general a question, but what would constitute a
> sensible equality test for two Mechanize::Page objects?
RFC 2616 section 13.3 appears to be where you want to look.  At first glance
this seems to include URI, Last-Modified header and Etag header.