Do Events buffer object data?

[Michael Pokraka]

Hi Jocelyn,
Thanks for your feedback, I have raised it with OSS. I've raised this before - this might be what you recall - but never really got a proper solution. I didn't pursue it too far as there were more pressing problems, but deleting 30+ workflow errors/day becomes tedious.
 
It makes sense from a design perspective to evaluate it at linkage time as that's where it's specified, but from a practical point of view the WF trigger time is more appropriate as we can have up to 15 minute delays in starting the WF.
 
Cheers
Mike
 
On Tue, Oct 07, 2003 at 02:12:32AM +0200, Dart, Jocelyn wrote:
> Hi Mike,
> Yes that's correct the check function is evaluated at event time.  This has been raised before but I can't see any OSS notes.  You might like to raise the issue with OSS and ask for a formal answer on how you should cope with the event delivery of a deleted object - i.e. how do you close the workflow or stop the workflow being instantiated in this situation.
> Regards,
>         Jocelyn Dart
> Consultant (SRM, EBP, Workflow)
> and co-author of the book
> "Practical Workflow for SAP"
> SAP Australia
> email: jocelyn.dart at sap.com
> phone: +61 412 390 267
> fax:   +61 2 9935 4880
>
>
>
>
> -----Original Message-----
>> From: Michael Pokraka [mailto:workflow at quirky.me.uk]
> Sent: Monday,6 October 2003 10:33 PM
> To: SAP-WUG at MITVMA.MIT.EDU
> Subject: Re: Do Events buffer object data?
>
>
> Hmmmm, I've answered my own question. The check function is evaluated at event ttime, not when it gets delivered from the queue.
> Thus an event goes onto the queue, a document gets changed which would normally invalidate the check function. The workflow still gets triggered even though it's irrelevant.
> Has anyone had to deal with this - I'd appreciate any feedback if you have. The obvious choice is to check everything in the WF. But is currently causing a headache with deleted documents causing the WF to fail (object doesn't exist).
>
> Cheers
> Mike
>
> On Fri, Oct 03, 2003 at 11:22:07AM +0100, Michael Pokraka wrote:
> > Hi all,
> > I've encountered a somewhat odd behaviour on a 4.6c system when using the event queue. A test sequence which I can reproduce is as follows:
> >
> > A CREATED event is raised and passed into the event queue.
> > The object is changed or deleted so that the start conditions should now fail.
> > I raise another CREATED event directly via SWUE
> > This one fails on start condition
> > Later, the original event gets delivered off the queue.
> > The start condition evaluates the old values and workflow is triggered
> >
> > My next step is probably to write my own check FM, but I thought I'd share that with the list, and perhaps even gain some insight from anyone who has encountered similar oddities...
> >
> > Cheers
> > Mike