RE: [OpenSTA-devel] Re: OpenSTA cookie handling

[Michael D. <md...@in...>]

> - cookie names changing.
> - more cookies introduced.
> - recording made with cookies set when clean browser really wanted.

The problem I am having is the time the server cookie issuing changed.  =
For example, the recorded SCL has the cookie get/set after URL 20, but =
in actual replay, the Set-Cookie is issued after URL 25.

You may wonder why this would ever happen.  Here is how I am using =
OpenSTA.  Instead of recording and re-recording complex script each =
time, I componentize the SCL scripts.  Basically, I record a single user =
operation (in business logic) per script.  At load test time, I combine =
the operations (scripts) I needed into the test.  Some of them =
concurrently, and many of them serially.  Here is an example:

Script 1: Login                 (get new cookie)
Script 2: navigate to screen 1
Script 3: fill out form A
Script 4: navigate to screen 2  (cookie expired, renewal is needed here)
Script 5: fill out form B
Script 6: Logout

When composing the actual test, I may want to loop Script 3 10 times and =
Script 5 20 times.  The test become:

Script 1: Login                 (get new cookie)
Script 2: navigate to screen 1
Script 3: fill out form A X 10  (cookie expired at loop 5, cookie shift =
place)=20
Script 4: navigate to screen 2
Script 5: fill out form B X 20  (cookie expired, renewal is needed =
somewhere in loop again)=20
Script 6: Logout

In the above case, there is no way for the script writer to predict when =
and where should the cookie be handled, if the play back engine doesn't =
do that automatically.

I think componentizing the script is a common way to test a enterprise =
system, and that would very likely introduce the above problem.

--Michael