RE: Raw Rows versus EOs
RE: Raw Rows versus EOs
- Subject: RE: Raw Rows versus EOs
- From: "Müller, Michael" <email@hidden>
- Date: Thu, 3 Feb 2005 11:40:39 +0100
Hi Anjo,
> A probably useful fix for this would be to subclass WORepetition and
> instead of adding .<iteration> (eg. .0, .1 etc) to the elementID add
> .<elementID>/<System.identityHashCode(currentObject)> (eg.
> .0/-266637,
> .1/2039933 etc.) Then, when invoking the action you could
> check if the
> identityHashCode is still the same and if not, either raise an
> exception complaining about backtracking (caught by the parent and
> display a reasonable error message) or find the matching object and
> invoke the action there.
A while back, we did something similar. But this screwed up our
load testing as the tool we used then was based on recorded user
interaction scripts. The identity hash codes changed (of course)
in every session or ultimately when the application got restarted
(based on what kind of objects they were referring to).
Anyway, the only thing I want to point out is that coding
identity hashcodes into URLs will probably screw up the repeatability
of recorded-script-based load testing tools (unless you use
cleverer tools like Rational's or the like).
cheers,--micha
--
-----------------------------------------------------
Michael Müller, Dipl.-Math. (mailto:email@hidden)
Projektleiter, Condat AG
Alt-Moabit 91 D, D-10559 Berlin
Fon: +49.30.3949-1227; Fax: +49.30.3949-1300
http://www.condat.de/
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden