Web testing with Ruby

ruby-talk
1

Does anyone have suggestions for projects/libraries to web test code
(HTMLUnit like)?

I had this project bookmarked from a month or so ago on this list:
http://www.xpenguin.biz/download/webunit/index-en.html

But about 1/2 of the listed "libraries used" are no longer functioning links.

I also looked at watir http://wtr.rubyforge.org/
which looked nice, but unfortunately requires Internet Explorer via a
COM interface, and I am looking to run this test on a linux server via
cron.

Are there some other useful web testing libraries people could point me towards?

Thanks,

Regards,
Jason
http://blog.casey-sweat.us/

2

Take a look at Michael Neumann's WWW::Mechanize. He has made it very
easy to code up an app to fetch a page, find pertinent content, follow
links, and so on.

It's on RubyForge, as part of the Wee project

http://rubyforge.org/projects/wee

James

···

On Tue, 8 Feb 2005 03:41:21 +0900, Jason Sweat <jason.sweat@gmail.com> wrote:

Does anyone have suggestions for projects/libraries to web test code
(HTMLUnit like)?

3

Selenium: http://selenium.thoughtworks.com/

Works with IE, Mozilla, Safari.

It uses client-side Javascript to control the browser. There is a Ruby
interface, or you can specify the tests as a bunch of HTML tables that have
actions and assertions. (There are also interfaces for Java, .NET and
Python.)

I haven't used the Ruby interface, but started using the table-based one
today.

···

In article <5c2874720502071041543561d3@mail.gmail.com>, Jason Sweat wrote:

Does anyone have suggestions for projects/libraries to web test code
(HTMLUnit like)?

I had this project bookmarked from a month or so ago on this list:
http://www.xpenguin.biz/download/webunit/index-en.html

But about 1/2 of the listed "libraries used" are no longer functioning links.

I also looked at watir http://wtr.rubyforge.org/
which looked nice, but unfortunately requires Internet Explorer via a
COM interface, and I am looking to run this test on a linux server via
cron.

Are there some other useful web testing libraries people could point me towards?

4

Watir has been made to work with Linux and Wine. Details:
http://advogato.org/person/superant/diary.html?start=76

···

At 12:41 PM 2/7/2005, Jason Sweat wrote:

I also looked at watir http://wtr.rubyforge.org/
which looked nice, but unfortunately requires Internet Explorer via a
COM interface, and I am looking to run this test on a linux server via
cron.

_____________________
  Bret Pettichord
  www.pettichord.com

5

Thanks, I will have a look at it.

Regards,
Jason

···

On Tue, 8 Feb 2005 04:20:35 +0900, James G. Britt <ruby.talk.list@gmail.com> wrote:

On Tue, 8 Feb 2005 03:41:21 +0900, Jason Sweat <jason.sweat@gmail.com> wrote:
> Does anyone have suggestions for projects/libraries to web test code
> (HTMLUnit like)?

Take a look at Michael Neumann's WWW::Mechanize. He has made it very
easy to code up an app to fetch a page, find pertinent content, follow
links, and so on.

It's on RubyForge, as part of the Wee project

http://rubyforge.org/projects/wee

6

This looks great. Thanks for the pointer!

···

"Tim Sutherland" <timsuth@ihug.co.nz> wrote

Selenium: http://selenium.thoughtworks.com/

7

I seem to be running into difficulty with mechanize.

$ ruby --version
ruby 1.9.0 (2005-02-08) [i686-linux]
$ gem --version
0.8.1
$ gem list --local | grep mech
mechanize (0.1.0)
$ cat test.rb
require 'rubygems'
require 'mechanize'

agent = WWW::Mechanize.new {|a| a.log = Logger.new(STDERR) }
page = agent.get('https://www.alcoadirect.com/&#39;\)
p page.forms

$ ruby test.rb
I, [2005-02-08T09:30:10.092550 #1947] INFO -- : GET:
https://www.alcoadirect.com/
warning: peer certificate won't be verified in this SSL session
D, [2005-02-08T09:30:14.305853 #1947] DEBUG -- : request-header: accept => */*
D, [2005-02-08T09:30:14.393482 #1947] DEBUG -- : header: last-modified
: Tue, 08 Feb 2005 00:02:27 GMT
D, [2005-02-08T09:30:14.393568 #1947] DEBUG -- : header: content-type
: text/html
D, [2005-02-08T09:30:14.393607 #1947] DEBUG -- : header: date : Tue,
08 Feb 2005 15:30:43 GMT
D, [2005-02-08T09:30:14.393646 #1947] DEBUG -- : header: server : JRun
Web Server/3.0
D, [2005-02-08T09:30:14.393684 #1947] DEBUG -- : header:
transfer-encoding : chunked
I, [2005-02-08T09:30:14.554417 #1947] INFO -- : status: 200
/usr/lib/ruby/gems/1.9/gems/mechanize-0.1.0/lib/mechanize.rb:115:in
`parse': undefined method `downcase' for nil:NilClass (NoMethodError)
        from /usr/lib/ruby/gems/1.9/gems/mechanize-0.1.0/lib/mechanize.rb:112:in
`call'
        from /usr/lib/ruby/gems/1.9/gems/mechanize-0.1.0/lib/mechanize/parsing.rb:14:in
`each_recursive'
        from /usr/lib/ruby/gems/1.9/gems/mechanize-0.1.0/lib/mechanize/parsing.rb:13:in
`each'
        from /usr/lib/ruby/1.9/rexml/element.rb:916:in `each'
        from /usr/lib/ruby/1.9/rexml/xpath.rb:49:in `each'
        from /usr/lib/ruby/1.9/rexml/element.rb:916:in `each'
        from /usr/lib/ruby/gems/1.9/gems/mechanize-0.1.0/lib/mechanize/parsing.rb:13:in
`each_recursive'
        from /usr/lib/ruby/gems/1.9/gems/mechanize-0.1.0/lib/mechanize/parsing.rb:15:in
`each_recursive'
         ... 136 levels...
        from /usr/lib/ruby/gems/1.9/gems/mechanize-0.1.0/lib/mechanize/parsing.rb:13:in
`each_recursive'
        from /usr/lib/ruby/gems/1.9/gems/mechanize-0.1.0/lib/mechanize.rb:212:in
`parse_html'
        from /usr/lib/ruby/gems/1.9/gems/mechanize-0.1.0/lib/mechanize.rb:164:in
`forms'
        from test.rb:6

Is the HTML parser perhaps dependant on well formed HTML? Is there a
more appropriate forum for me to raise this issue? Thanks.

Regards,
Jason

···

On Tue, 8 Feb 2005 04:20:35 +0900, James G. Britt <ruby.talk.list@gmail.com> wrote:

On Tue, 8 Feb 2005 03:41:21 +0900, Jason Sweat <jason.sweat@gmail.com> wrote:
> Does anyone have suggestions for projects/libraries to web test code
> (HTMLUnit like)?

Take a look at Michael Neumann's WWW::Mechanize. He has made it very
easy to code up an app to fetch a page, find pertinent content, follow
links, and so on.

It's on RubyForge, as part of the Wee project

http://rubyforge.org/projects/wee

8

Look at SWExplorerAutomation
(http://home.comcast.net/~furmana/SWIEAutomation.htm). It allows
visually generate test scripts.

9

I seem to be running into difficulty with mechanize.

...

agent = WWW::Mechanize.new {|a| a.log = Logger.new(STDERR) }
page = agent.get('https://www.alcoadirect.com/&#39;\)
p page.forms

I wonder if the issue is the use of "https" ?

Can you get Mechanize to work with a non-SSL URL?

James

···

On Wed, 9 Feb 2005 00:35:54 +0900, Jason Sweat <jason.sweat@gmail.com> wrote:

10

I don't think https is the issue:
After creating page I can do

page.inspect
=> "#<WWW::Page:0xb7a0cad0 @code=\"200\", @watch_for_set=nil,
@uri=#<URI::HTTPS:0x…fdbd0689c URL:https://www.alcoadirect.com/&gt;,
@body=\" \\r\\n<HTML dir=ltr>\\r\\n<HEAD>\\r\\n<TITLE>
...snip...
        </TD>\\r\\n </TR>\\r\\n \\r\\n
      </TABLE>\\r\\n
</TD></TR></TABLE></TD></TR></TBODY></TABLE>\\r\\n<hr size=1
width=760>\\r\\n</BODY>\\r\\n</HTML>\\r\\n\", @response=#<Net::HTTPOK
200 OK readbody=true>, @cookies=>"

It gives me the error when I try the page.forms method. Any other ideas?

Regards,
Jason

···

On Wed, 9 Feb 2005 02:44:10 +0900, James G. Britt <ruby.talk.list@gmail.com> wrote:

> agent = WWW::Mechanize.new {|a| a.log = Logger.new(STDERR) }
> page = agent.get('https://www.alcoadirect.com/&#39;\)
> p page.forms

I wonder if the issue is the use of "https" ?

Can you get Mechanize to work with a non-SSL URL?

11

I don't think https is the issue:
After creating page I can do

page.inspect
=> "#<WWW::Page:0xb7a0cad0 @code=\"200\", @watch_for_set=nil,
@uri=#<URI::HTTPS:0x…fdbd0689c URL:https://www.alcoadirect.com/&gt;,
@body=\" \\r\\n<HTML dir=ltr>\\r\\n<HEAD>\\r\\n<TITLE>
...snip...
        </TD>\\r\\n </TR>\\r\\n \\r\\n
      </TABLE>\\r\\n
</TD></TR></TABLE></TD></TR></TBODY></TABLE>\\r\\n<hr size=1
width=760>\\r\\n</BODY>\\r\\n</HTML>\\r\\n\", @response=#<Net::HTTPOK
200 OK readbody=true>, @cookies=>"

It gives me the error when I try the page.forms method. Any other ideas?

Now this sounds familiar Ithough don';t tink I ever told Mike of my problems )

mechanize.rb, in the parse_html method, has ths code:

: @root.each_recursive {|node|
: name = node.name.downcase
: case name

If the node does not have a name attribute, then downcase
expectorates. I think.

Try changing that to
: name = node.name.to_s.downcase

or add a nil? test someplace

James

···

On Wed, 9 Feb 2005 03:08:25 +0900, Jason Sweat <jason.sweat@gmail.com> wrote:

Jason

12

Changing all .downcase to .to_s.downcase worked for me.

Thanks!

Regards,
Jason

···

On Wed, 9 Feb 2005 05:11:05 +0900, James G. Britt <ruby.talk.list@gmail.com> wrote:

On Wed, 9 Feb 2005 03:08:25 +0900, Jason Sweat <jason.sweat@gmail.com> wrote:

Try changing that to
: name = node.name.to_s.downcase