Workflows starting for deleted objects - nevermind

[Michael Pokraka]

Yup, but workflow still assumes that if an object has a key it must be in=
stantiated - even for instance indep. methods. (At least that is what it =
looks like).=20
So it still fails.=20
 
On the other hand, I've been missing the obvious all along - a check in t=
he start condition does the trick...  I do feel a bit silly now.=20
 
Thanks for the input though...=20
 
Cheers
Mike
 
On Mon, May 19, 2003 at 02:45:53PM +0200, Schmidinger, Heinz (Unaxis IT B=
Z) wrote:
> the create object handle is the problem, so I thought to use an own met=
hod
> instnceless defined which is not trying to create an object which is on=
ly
> giving back a paramter.
>=20
> -----Urspr=FCngliche Nachricht-----
> Von: Michael Pokraka [mailto:workflow at quirky.me.uk]
> Gesendet am: Montag, 19. Mai 2003 14:41
> An: SAP-WUG at MITVMA.MIT.EDU
> Betreff: Re: Workflows starting for deleted objects
>=20
> Ahh, but that is exactly what LIKP.EXISTENCECHECK doe. Or at least it's
> supposed to do.=20
> But... as far as I can figure, since it has a key, it must exist, there=
fore
> workflow instantiates the object (or at least it tries to) before execu=
tiong
> the method. In the trace it fails with 'Object handle for object cannot=
 be
> created'.
>=20
> Thanks for the input,=20
> Cheers
> Mike
>=20
> On Mon, May 19, 2003 at 01:16:10PM +0200, Schmidinger, Heinz (Unaxis IT=
 BZ)
> wrote:
> > Hi Mike,
> >=20
> > just an idee:
> >=20
> > cretae an own instanceless method just reading the table if row exist=
s and
> > return a parameter (not)exits.
> >=20
> > regards
> >=20
> > Heinz
> >=20
> > -----Urspr=FCngliche Nachricht-----
> > Von: Michael Pokraka [mailto:workflow at quirky.me.uk]
> > Gesendet am: Montag, 19. Mai 2003 13:02
> > An: SAP-WUG at MITVMA.MIT.EDU
> > Betreff: Workflows starting for deleted objects
> >=20
> > Hi all,
> > Another question to ponder:
> >=20
> > Users often create deliveries and delete them straight afterwards (I'=
m too
> > scared to ask why - but I assume it's when they mess things up and wa=
nt to
> > start again).
> > Now, we have workflows that start after a LIKP.CREATED event, using t=
he
> > event queue. The Queue introduces a delay, and hence we have workflow=
s
> which
> > fail straight away (object doesn't exist). Adding a task based on
> > LIKP.EXISTENCECHECK also fails, it appears to be that it falls over w=
hen
> > instantiating the object in order to check it's existence... (duh!).
> >=20
> > Any ideas would be welcome...
> > Thank
> > Mike