Re: [Wtr-general] Watir.. close, but not close enough
In my opinion, Watir is absolutely what we need. A smallish tool for the job, which can be extended at your will. That is why, among other reasons, I believe Bret started this thing, a tool with no vendorscript, but with a proper programming language. The tool allows you to do the things that are not in the standard language, namely manipulate web pages. For the rest use Ruby itself. This solution is incredibly flexible. It all started for me with Michael Kelly's article on performance measurement. That article showed me how to think about writing test scripts. Now, I log onto Oracle, get a value, run a browser interaction with parameterized input values, write to a spreadsheet, mail myself a failure report and many more things. It is really hand-rolled per requirement. If you want a solution that does more or looks prettier, get a vendor tool. Watir is for testers who like to hack a bit and roll their own solutions. It is not meant to be a one size fits all solution. I just wish the Win32-GUITest library could get the same amount of attention. I've had lots of success with Perl's version of this lib., and would like to do that testing in Ruby now. My 2 cents. Walter Kruse -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Bret Pettichord Sent: 22 February 2007 06:46 AM To: wtr-general@rubyforge.org Subject: Re: [Wtr-general] Watir.. close, but not close enough John Lolis wrote: I don't feel that its a problem Watir needs to solve. If the community really did want some framework I think it would be a project separate from Watir. I'm not even sure a project like that would work. Every application that needs testing is probably going to need a slightly different approach. If someone attempts to write a 'be all end all' framework its probably going to end up very complex and hard to understand (see most commercial testing packages). To some degree I agree. However, our user's guide tells people how to use Watir with Test::Unit and so a lot of people group them together. Indeed, we've had so many complaints about the poor fit of this unit testing framework for system acceptance testing that I've included Watir::TestCase in the recent versions of Watir 1.5. Namely, this TestCase executes its test methods in the order defined and also supplies verify methods that work like assert, but don't abort a test case when they fail. These were the two most common complaints with Test::Unit itself by Watir users. But i do agree that there is a value to keeping Watir small and flexible and it seems to me that it would probably be better to migrate this new TestCase subclass to some other project. It could still be used by Watir users, but they would be free to use whatever test harness they wanted: Test::Unit or Rspec or this new thing or maybe something they built themselves. Bret ___ Wtr-general mailing list Wtr-general@rubyforge.org http://rubyforge.org/mailman/listinfo/wtr-general ** Everything in this e-mail and attachments relating to the official business of MultiChoice Africa is proprietary to the company. Any view or opinion expressed in this message may be the view of the individual and should not automatically be ascribed to the company. If you are not the intended recipient, you may not peruse, use, disseminate, distribute or copy this message. If you have received this message in error, please notify the sender immediately by email, facsimile or telephone and destroy the original message. ** ___ Wtr-general mailing list Wtr-general@rubyforge.org http://rubyforge.org/mailman/listinfo/wtr-general
Re: [Wtr-general] Watir.. close, but not close enough
Walter Kruse wrote: I just wish the Win32-GUITest library could get the same amount of attention. I've had lots of success with Perl's version of this lib., and would like to do that testing in Ruby now. Maybe you could give it some attention. Seriously, there is no one to do this stuff except us. Bret ___ Wtr-general mailing list Wtr-general@rubyforge.org http://rubyforge.org/mailman/listinfo/wtr-general
Re: [Wtr-general] Watir.. close, but not close enough
On 2/22/07, Walter Kruse [EMAIL PROTECTED] wrote: Bret, I had a look long ago at Piotr Kaluski's very detailed docs for Perl's Win32GuiTest (http://www.piotrkaluski.com/files/winguitest/docs/winguitest.html). He even lists some C code examples. I'm afraid C is a little beyond me... Wayne Vucenic got a good start on a Ruby Win32::Guitest. http://rubyforge.org/projects/guitest/ There are actually enough functions implemented that the code is probably even useful right not. I'm an administrator for the project, so if you'd like to jump in anywhere, let me know. -Chris ___ Wtr-general mailing list Wtr-general@rubyforge.org http://rubyforge.org/mailman/listinfo/wtr-general
Re: [Wtr-general] Watir.. close, but not close enough
Indeed. Watir's strength for us was that we couldn't afford a huge application and all the time and training that entails, and Watir allowed us to put together a pretty comprehensive set of tests for our few websites. Creating a general-purpose framework would probably make it as complex as just using native watir/ruby anyway! It's bad enough with the number of tricks I have to play with my simple Login() function to cater for the places where it uses fields called username, usrname or user for the same thing. In which case it soon becomes easier and clearer to just use ie.text_field(:name, username).set(myuser) ie.text_field(:name, password).set(mypass) ie.button(:name, submit).click - Posted via Jive Forums http://forums.openqa.org/thread.jspa?threadID=6532messageID=18986#18986 ___ Wtr-general mailing list Wtr-general@rubyforge.org http://rubyforge.org/mailman/listinfo/wtr-general
Re: [Wtr-general] Watir.. close, but not close enough
John Lolis wrote: I don't feel that its a problem Watir needs to solve. If the community really did want some framework I think it would be a project separate from Watir. I'm not even sure a project like that would work. Every application that needs testing is probably going to need a slightly different approach. If someone attempts to write a 'be all end all' framework its probably going to end up very complex and hard to understand (see most commercial testing packages). To some degree I agree. However, our user's guide tells people how to use Watir with Test::Unit and so a lot of people group them together. Indeed, we've had so many complaints about the poor fit of this unit testing framework for system acceptance testing that I've included Watir::TestCase in the recent versions of Watir 1.5. Namely, this TestCase executes its test methods in the order defined and also supplies verify methods that work like assert, but don't abort a test case when they fail. These were the two most common complaints with Test::Unit itself by Watir users. But i do agree that there is a value to keeping Watir small and flexible and it seems to me that it would probably be better to migrate this new TestCase subclass to some other project. It could still be used by Watir users, but they would be free to use whatever test harness they wanted: Test::Unit or Rspec or this new thing or maybe something they built themselves. Bret ___ Wtr-general mailing list Wtr-general@rubyforge.org http://rubyforge.org/mailman/listinfo/wtr-general
Re: [Wtr-general] Watir.. close, but not close enough
I think the problem for the OP is that we don't have good examples/documentation describing this approach. I don't feel that its a problem Watir needs to solve. If the community really did want some framework I think it would be a project separate from Watir. I'm not even sure a project like that would work. Every application that needs testing is probably going to need a slightly different approach. If someone attempts to write a 'be all end all' framework its probably going to end up very complex and hard to understand (see most commercial testing packages). Watir's strength, to me, is the fact you bring your knowledge of testing and coding together. Its a great tool that grows with you and rarely holds you back. I did in fact run into the problem of a 'test harness', but I asked some questions and just tried something - and it failed, i then tossed it and tried again - and it worked. I have a strong feeling if and when i try a new project it will be even better. Its kind of a like a hand saw. Anyone with a saw can cut wood, but it truly takes experience to cut wood well (too out there?). Anywho, my 2g's :) - Posted via Jive Forums http://forums.openqa.org/thread.jspa?threadID=6532messageID=18925#18925 ___ Wtr-general mailing list Wtr-general@rubyforge.org http://rubyforge.org/mailman/listinfo/wtr-general
Re: [Wtr-general] Watir.. close, but not close enough
I'm not entirely sure what you mean by being unable to reference scripts, but I split my tests into: 1 script with helper functions such as GoToURLAndExpectText for example 1 script with the header type declarations in it, URL's, usernames etc X several scripts with various parts of the app covered within, such as 1 script for basic guest browsing 1 script for loging in and out 1 script for advanced interaction in one part 1 script for advanced interaction in another part and then pull it all together. Is this the sort of thing you mean? Or are you referring to something else alltogether? require 'rubygems' require 'watir' require 'functions' # my functions $testsite = http://myurl.com; $ie = Watir::IE.new puts Starting tests require 'basictests' require 'advancedtests' require 'somemoretests' - Posted via Jive Forums http://forums.openqa.org/thread.jspa?threadID=6532messageID=18892#18892 ___ Wtr-general mailing list Wtr-general@rubyforge.org http://rubyforge.org/mailman/listinfo/wtr-general
Re: [Wtr-general] Watir.. close, but not close enough
Cylindric wrote: I'm not entirely sure what you mean by being unable to reference scripts, but I split my tests into: 1 script with helper functions such as GoToURLAndExpectText for example 1 script with the header type declarations in it, URL's, usernames etc X several scripts with various parts of the app covered within, such as 1 script for basic guest browsing 1 script for loging in and out 1 script for advanced interaction in one part 1 script for advanced interaction in another part and then pull it all together. Is this the sort of thing you mean? Or are you referring to something else alltogether? I think the problem for the OP is that we don't have good examples/documentation describing this approach. ___ Wtr-general mailing list Wtr-general@rubyforge.org http://rubyforge.org/mailman/listinfo/wtr-general
[Wtr-general] Watir.. close, but not close enough
Critical commentary follows: [i]I have taken a leap into watir and I must say that at first glance it was fun, easy and usable. However, when I needed a more complicated and managed solution it fell flat. Watir provides an easy and simple way to access DOM elements and complete user (through web UI) tasks. Thus, really great for smoke test automation. Sadly though, its support for multiple scripts to be run in a preassigned sequence is pretty much non-existant. If you do approach your testing from the perspective that you have one long user process you want to emulate through scripting, then it can work. If you have more building block orientated approach, the structure doesnt accomodate that. At all. I dont see the gains in creating massive and long process action changes embedded in a script that covers a pieces of functionality. I see it as a duplication of effort, difficult to maintain, difficult to know what you have at a glance and, in the end, a maintenance nightmare. I want a better setup that allows me to do the following: * Allows me to reference the scripts that I want to run and, * Allows me to determine the order that they run in * Gives me better and clear reporting after the tests have finished running and, * Gives me that end result in both .xls and .html format It is severely limiting not to be able to reference scripts. Currently, due to nature of the DMS it is easy to build shorthand action scripts. Action scripts cover (at the most basic level) user end to end actions. An example of this would be Login. All it requires is a username and password and the user must click on the Login button. I want to be able to self-contain that action within a script and then call it whenever I want to set up a longer process chain of actions. If I can build a better test runner to manager to manage that, I think that would probably take first place. (and possibly integrate it with a test case management tool.. something along those lines). Otherwise, the hunt continues [/i] http://seryph.wordpress.com/2007/02/12/watir-close-but-not-close-enough/ I quote this because i basically agree. Watir is only a browser driver and we leave it to users to build their own testing framework using Test::Unit and Test::Unit Reporter and the like. This in fact is how i spend much of my time at my day job. - Posted via Jive Forums http://forums.openqa.org/thread.jspa?threadID=6532messageID=18618#18618 ___ Wtr-general mailing list Wtr-general@rubyforge.org http://rubyforge.org/mailman/listinfo/wtr-general
