[Webtest] WebTest-http-request-recorder? Verifying whole HTML-pages?

Dierk Koenig webtest@lists.canoo.com
Tue, 15 Oct 2002 16:11:43 +0200


Hi Damian,

there are numerous http recorders on the market e.g.
WebArt and Mercury WinRunner.

Canoo WebTest follows the model-driven route that is substantially
different.
Specifying your tests manually has some important benefits:

- Duplication in Tests can be avoided from the very start, making
  your Test Base much easier to keep up-to-date.

- Your Test are human readable, thus serving as documentation and
  even specification (when specified before implementation).

- They can distinguish between expected and accidential behaviour
  of your system.

- They make you think about your System:
  "What am I going to accomplish here?"
  This supports finding small and elegant solutions.

You may also refer to the White Paper for this issue.

cheers
Mittie

> -----Original Message-----
> From: webtest-admin@lists.canoo.com
> [mailto:webtest-admin@lists.canoo.com]On Behalf Of Damian Glania
> Sent: Donnerstag, 10. Oktober 2002 12:46
> To: webtest@gate.canoo.com
> Subject: [Webtest] WebTest-http-request-recorder? Verifying whole
> HTML-pages?
>
>
> Hi all,
> I would like to use the WebTest to perform the functional tests
> of our web front end.
> I'm wondering if I can generate (part of) the WebTest test script
> instead of writing
> it manually? Is there any one who tried to do it before?
>
> I think, I could do it by performing the following steps:
> 1) write an "http-request-recorder" as logging extension to our
> dispatcher servlet.
>    The recorder would perform the following tasks:
>    - each incoming http-request would be recorded in "WebTest
> format" e.g.:
>    <invoke stepid="get 0x100-01" url="Main?navId=0x100-001" />
>    - for each request, a default validation tag, comparing the
> current http-response to
>      a corresponding one - which was (optically ;-) validated
> before - would be generated.
>      (a new validation tag, the "diff-tag" has to be introduced)
>
> 2) the human tester would start the recorder and perform all
> tests as described in the
>    test plan using a browser. He validates (optically) each
> response (HTML-Page) according to the description in the test
> plan. Each response page will be written to a file (WebTest functionality)
>
> 3) After completion of the manual test the generated WebTest
> script would be manually "polished", e.g. some of the default
> "diff-tags" would be replaced by specialized ones (especially for
> responses including timestamps or other volatile parts)
>
> 4) the following test could be performed automatically until the
> front end is changed. As the backend is not a part of our
> project, we would use a stub based on a database of stored
> backend-request-response-records.  This should help to make the
> tests more repeatable.
>
>
> So my questions are:
> - what do you think about that scenario? Could it work? Is it a
> sensible way to do that test?
> - is there any software available, I can use to write the
> "WebTest-http-request-recorder"?
> - is there any software available, I can use to write the
> "diff-tag" comparing the previously stored HTML-page to the
> current response?
>
>
>  Any help, ideas, comments and software samples would be greatly
> appreciated!
>
>  Damian
>
> __________________________________________________________________
> ____________
> Die clevere Geldreserve: der DiBa-Privatkredit. Funktioniert wie
> ein Dispo,
> ist aber viel gunstiger! Alle Infos: http://diba.web.de/?mc=021104
>
> _______________________________________________
> WebTest mailing list
> WebTest@lists.canoo.com
> http://lists.canoo.com/mailman/listinfo/webtest
>