juddi-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Viens" <st...@viens.net>
Subject RE: question about 'delete_business'
Date Thu, 24 Feb 2005 04:43:28 GMT
Gier, 

I just checked this out and it's a bug ... and one that's also a problem
for delete_service and delete_binding.

Although the v2 spec doesn't explicitly give you a way to determine the
problem on the client side programmatically". The following seems to say
that a key will exist *someplace* within the error text:

  "The key that caused the error will be 
   clearly indicated in the error text."

Steve

-----Original Message-----
From: Anne Thomas Manes [mailto:atmanes@gmail.com] 
Sent: Saturday, February 19, 2005 4:44 PM
To: juddi-dev@ws.apache.org
Subject: Re: question about 'delete_business'


According to the UDDI v2 spec, the SOAP fault should return the key that
was invalid:

E_invalidKeyPassed: signifies that one of the uuid_key values passed did
not match with any known businessKey values.  No partial results will be
returned - if any businessKey values passed are not valid or if the
message contained multiple instances of a uuid_key value, this error
will be returned.  The key that caused the error will be clearly
indicated in the error text.

- Anne


On Fri, 11 Feb 2005 10:23:52 -0500, Geir Magnusson Jr
<geir@4quarters.com> wrote:
> 
> On Feb 11, 2005, at 10:08 AM, Anil Saldhana wrote:
> 
> > Geir,
> >   I had thought of the same a month back.
> >
> > I thing the right thing to do is call jUDDI n times
> > for n keys to delete, because we have to keep track of partial 
> > commits.
> 
> Right - OTOH, the JavaDoc says
> 
> "partial commits are allowed"
> 
> not required.  I figure that we do want to try the one-shot first, and

> if that fails, then iterate through...  I'll add that as a JIRA in 
> Scout-land.  (Will give me a reason to go find it...)
> 
> geir
> 
> >
> > Regards,
> > Anil
> >
> >
> > --- Geir Magnusson Jr <geir@4quarters.com> wrote:
> >
> >> Additional thought...
> >>
> >> On Feb 11, 2005, at 6:58 AM, Geir Magnusson Jr.
> >> wrote:
> >>>
> >>> JAXR says that I can do a deleteOrganizations()
> >> and have a partial
> >>> commit, stopping on the first failure.  In my
> >> case, I assume that the
> >>> first good key will be deleted, and the second
> >> not.  This seems then
> >>> that to use jUDDI (or any UDDI registry - I'm not
> >> picking on jUDDI), I
> >>> must send two separate delete_business messages?
> >>
> >> ... if I wish to perform partial commits?  seems
> >> like the performant
> >> thing to do is assume that this kind of failure is
> >> rare, always try
> >> single-message multi-key deletes, returning a JAXR BulkResponse to 
> >> my callers w/ only one error in the event of a failure,
> >> letting them
> >> assume that nothing worked?
> >>
> >> How does a UDDI user figure out which key is bad
> >> since the UDDI
> >> response message doesn't give any hint of which key
> >> was the bad one...?
> >>   Seems like you then need to play delete_business wack-a-mole...
> >>
> >> geir
> >>
> >>>
> >>> --
> >>> Geir Magnusson Jr
> >> +1-203-665-6437
> >>> geirm@apache.org
> >>>
> >>>
> >> --
> >> Geir Magnusson Jr
> >> +1-203-665-6437
> >> geir@gluecode.com
> >>
> >>
> >
> >
> >
> >
> > __________________________________
> > Do you Yahoo!?
> > Read only the mail you want - Yahoo! Mail SpamGuard. 
> > http://promotions.yahoo.com/new_mail
> >
> >
> --
> Geir Magnusson Jr                                  +1-203-665-6437
> geir@gluecode.com
> 
>



Mime
View raw message