| URL recording & session cookie - Trouble using WAS 6.0 test env. as opposed to Tomcat 5.0 [message #5736] |
Mon, 14 February 2005 13:24  |
Eclipse User |
|
|
|
Originally posted by: jan.nicklisch.entimo.de
When recording a URL test suite (a sequence of HTTP requests/response pairs
that serve as basis for generating a Java test class subsequently) using a
"HTTP proxy recorder" some cookie information is also recorded and added
into the generated class. Those cookie values prevent playback of the test
using the builtin WebSphere AppServer test environment 6.0 because a new
session is created for each request, but this should not be the case for
e.g. an application where a user logs in to the WebApp (the cookies look
like 0000jf7xQ704VvqxJBovHzdsJZG:-1).
In contrast, playback using Tomcat 5.0 succeeds because of different cookie
values get reported (? I am not sure on that) or because that particular app
server treats those recorded cookies differently when played back? (Cookies
in this case look like 6E460E17DDA8DAEFFBD428052669D73D, so effectively
there is no ":-1").
My question: am I supposed to change the configuration of the test target
server in order to make the tests go through (I would not suspect I should
but would be willingly to do so) or where else is the root of my
misunderstanding? Do I need to change the recording mechanism or should I
alter the generated test class somehow?
Any help or hint is very much appreciated since I would like to evaluate the
applicability of the Hyades framework (within RAD 6.0 or as an extension to
a standard Eclipse workbench, no matter) for our needs rather soon.
Thank you in advance,
Jan Nicklisch
--
I'm using Rational Application Developer 6.0 Trial version with the built in
/ shipped with Hyades framework.
|
|
|
| Re: URL recording & session cookie - Trouble using WAS 6.0 test env. as opposed to Tomcat 5.0 [message #5806 is a reply to message #5736] |
Tue, 15 February 2005 15:11  |
 David Hodges
Messages: 2
Registered: July 2009 |
Junior Member |
|
|
The recording will capture the cookies that are used during the
recording session. You should be able to view those in the URL Test
editor to see what was captured. They are also generated into the test
class and can be modified there as well. When the test is executed the
cookies will be handled dynamically, so if the server responds with a
new cookie, as it's likely to do on the first request to retain a
session id, it should be used in the subsequent requests that use that
same cookie. If you're running 1 user, all requests from that 1 user
should use the dynamic cookie value. If you're running multiple users,
then of course, each user gets his own cookie cache and will generate a
new session.
You can see the updating of the cache if you examine the
executeRequest() method, where it retrieves the current values prior to
sending the request and then retrieves any returned cookies after the
request completes.
David Hodges
Jan Nicklisch wrote:
> When recording a URL test suite (a sequence of HTTP requests/response pairs
> that serve as basis for generating a Java test class subsequently) using a
> "HTTP proxy recorder" some cookie information is also recorded and added
> into the generated class. Those cookie values prevent playback of the test
> using the builtin WebSphere AppServer test environment 6.0 because a new
> session is created for each request, but this should not be the case for
> e.g. an application where a user logs in to the WebApp (the cookies look
> like 0000jf7xQ704VvqxJBovHzdsJZG:-1).
>
> In contrast, playback using Tomcat 5.0 succeeds because of different cookie
> values get reported (? I am not sure on that) or because that particular app
> server treats those recorded cookies differently when played back? (Cookies
> in this case look like 6E460E17DDA8DAEFFBD428052669D73D, so effectively
> there is no ":-1").
>
> My question: am I supposed to change the configuration of the test target
> server in order to make the tests go through (I would not suspect I should
> but would be willingly to do so) or where else is the root of my
> misunderstanding? Do I need to change the recording mechanism or should I
> alter the generated test class somehow?
>
> Any help or hint is very much appreciated since I would like to evaluate the
> applicability of the Hyades framework (within RAD 6.0 or as an extension to
> a standard Eclipse workbench, no matter) for our needs rather soon.
>
>
> Thank you in advance,
>
> Jan Nicklisch
>
> --
>
> I'm using Rational Application Developer 6.0 Trial version with the built in
> / shipped with Hyades framework.
>
>
|
|
|
Powered by
FUDForum. Page generated in 0.25624 seconds
] 
Search
Help
Register
Login
Home
