I did manage to get around the authentication problem by setting a
java.net.Authenticator  (with the CMIS username and password) for the
session containing the applet, but then I got an error from the applet
saying the content from the URL wasn't recognized.

So I think we don't have an issue with P8 - it's purely an issue with the
viewer now.  There is probably a way around it - just need to find it..:-)

Thanks for all the help.

Tim



On Thu, Aug 28, 2014 at 8:21 PM, Lucas, Mike <mike.lu...@gwl.ca> wrote:

> Just to add one salient point, HTTP cookies are part of the HTTP spec
> (RFC6265<http://tools.ietf.org/html/rfc6265> is the most current). So
> after receiving a Set-Cookie header with the LTPA token in it, the viewer
> should echo back that cookie in the Cookie header.
>
> michael lucas  |  Senior Software Developer  |  Great-West Life |
> mike.lu...@gwl.ca<mailto:mike.lu...@gwl.ca>
>
>
> From: Jay Brown [mailto:jay.br...@us.ibm.com]
> Sent: August 28, 2014 1:22 PM
> Cc: dev@chemistry.apache.org
> Subject: Re: IBM FileNet P8 CMIS URL addressability + Daeja ViewOne Viewer
>
>
> We support SSO today.  The SSO method that we support is the LTPA token in
> the http header which is handled automatically by the Websphere container
> where our CMIS service is running.
> This is a standard way of handling this and it seems like the viewer
> should support this as well.
>
> Do we know for certain that the viewer can not support this?
>
> Bottom line as we ship today. For SSO to work the client needs to include
> that standard HTTP header (cookie) with the LTPA token.  (or do basic HTTP
> auth, which again uses HTTP headers)
>
> Putting it in the url would be non-standard (at least not part of the CMIS
> spec) so I am reaching for a way to avoid this.
>
>
> Jay Brown
> Senior Engineer, ECM Development
> IBM Software Group
> jay.br...@us.ibm.com<mailto:jay.br...@us.ibm.com>
> www.linkedin.com/in/parityerror/<http://www.linkedin.com/in/parityerror/>
>
> [Inactive hide details for Tim Webster ---08/28/2014 09:36:49 AM---Hi, Do
> you mean 'Would you want the client to request the *UR]Tim Webster
> ---08/28/2014 09:36:49 AM---Hi, Do you mean 'Would you want the client to
> request the *URL* with an extra
>
> From:
>
>
> Tim Webster <tim.webs...@gmail.com<mailto:tim.webs...@gmail.com>>
>
>
> To:
>
>
> "dev@chemistry.apache.org<mailto:dev@chemistry.apache.org>" <
> dev@chemistry.apache.org<mailto:dev@chemistry.apache.org>>
>
>
> Date:
>
>
> 08/28/2014 09:36 AM
>
>
> Subject:
>
>
> Re: IBM FileNet P8 CMIS URL addressability + Daeja ViewOne Viewer
>
> ________________________________
>
>
>
> Hi,
>
> Do you mean 'Would you want the client to request the *URL* with an extra
> extension parameter..."?
>
> I'll be retrieving the URL with this (straight from your book):
>
>        if (session.getBinding().getObjectService() instanceof LinkAccess) {
>            return ((LinkAccess)
>
> session.getBinding().getObjectService()).loadContentLink(session.getRepositoryInfo()
>                    .getId(), document.getId());
>        }
>
> If that URL had the token, great - if not and we had to do something extra
> (like add it ourselves somehow), still OK.  The main thing is that the
> repository can actually use a URL with the token in it.
>
> Another thing occurred to me - would some kind of Single-sign on make all
> this 'just work' (including the viewer)?  Apologies that I keep mentioning
> the viewer, but it's the main problem here!
>
> Thanks,
>
> Tim
>
>
>
> On Thu, Aug 28, 2014 at 5:10 PM, Jay Brown <jay.br...@us.ibm.com<mailto:
> jay.br...@us.ibm.com>> wrote:
>
> > No we don't currently support this but I am just trying to think this
> > through from a security perspective if we wanted to add a feature.
> > We can't include the user's security token in all of the returned stream
> > url's by default. I would have to be requested.
> >
> > Would you want the client to request the document with an extra extension
> > parameter indicating that the stream url should be returned with the
> > current users LTPA token embedded? (includeAuthToken=true)
> >
> >
> > Jay Brown
> > Senior Engineer, ECM Development
> > IBM Software Group
> > jay.br...@us.ibm.com<mailto:jay.br...@us.ibm.com>
> > www.linkedin.com/in/parityerror/<http://www.linkedin.com/in/parityerror/
> >
> >
> > [image: Inactive hide details for Tim Webster ---08/28/2014 02:37:16
> > AM---Jay, I've re-read your email, and I think I misunderstood wha]Tim
> > Webster ---08/28/2014 02:37:16 AM---Jay, I've re-read your email, and I
> > think I misunderstood what you were saying...
> >
> >
> >
> >    From:
> >
> >
> > Tim Webster <tim.webs...@gmail.com<mailto:tim.webs...@gmail.com>>
> >
> >    To:
> >
> >
> > "dev@chemistry.apache.org<mailto:dev@chemistry.apache.org>" <
> dev@chemistry.apache.org<mailto:dev@chemistry.apache.org>>
> >
> >    Date:
> >
> >
> > 08/28/2014 02:37 AM
> >
> >    Subject:
> >
> >
> > Re: IBM FileNet P8 CMIS URL addressability + Daeja ViewOne Viewer
> > ------------------------------
> >
> >
> >
> > Jay,
> >
> > I've re-read your email, and I think I misunderstood what you were
> > saying...
> >
> > *"We currently only support LTPA tokens with the FileNet CMIS server.
>  So
> >
> > if your client adds a 'Cookie' header with a value of a valid LTPA token
> > (for the domain where the CMIS and CE server reside)  your request will
> > succeed without a challenge for credentials. "*
> >
> >
> > This is fine if we were actually constructing the HTTP request ourselves,
> > but we're not - the Daeja ViewOne applet is doing it.
> >
> > *"...would you need us to support the passing of the token as a parameter
> > in the Content stream URL?"*
> >
> >
> > This would in fact help - if the content stream URL contained the token,
> I
> > could just pass that URL to the applet and it can do the rest.  As it is,
> > does P8 support any URL of this type to retrieve content?
> >
> > Thanks,
> >
> > Tim
> >
> >
> >
> >
> > On Wed, Aug 27, 2014 at 6:26 PM, Jay Brown <jay.br...@us.ibm.com<mailto:
> jay.br...@us.ibm.com>> wrote:
> >
> > > We currently only support LTPA tokens with the FileNet CMIS server.
>  So
> > > if your client adds a 'Cookie' header with a value of a valid LTPA
> token
> > > (for the domain where the CMIS and CE server reside)  your request will
> > > succeed without a challenge for credentials.
> > >
> > > Does this help you or would you need us to support the passing of the
> > > token as a parameter in the Content stream URL?
> > >
> > >
> > > Jay Brown
> > > ECM Development, IBM
> > >
> > >
> > > [image: Inactive hide details for Tim Webster ---08/27/2014 03:36:11
> > > AM---Hi, I'm looking to stream documents to the Daeja ViewOne View]Tim
> > > Webster ---08/27/2014 03:36:11 AM---Hi, I'm looking to stream documents
> > to
> > > the Daeja ViewOne Viewer applet using a
> > >
> > >
> > >
> > >    From:
> > >
> > >
> > > Tim Webster <tim.webs...@gmail.com<mailto:tim.webs...@gmail.com>>
> > >
> > >    To:
> > >
> > >
> > > "dev@chemistry.apache.org<mailto:dev@chemistry.apache.org>" <
> dev@chemistry.apache.org<mailto:dev@chemistry.apache.org>>
> > >
> > >    Date:
> > >
> > >
> > > 08/27/2014 03:36 AM
> > >
> > >    Subject:
> >
> > >
> > >
> > > IBM FileNet P8 CMIS URL addressability + Daeja ViewOne Viewer
> > > ------------------------------
> >
> > >
> > >
> > >
> > > Hi,
> > >
> > > I'm looking to stream documents to the Daeja ViewOne Viewer applet
> using
> > a
> > > single URL.  I understand that a document can be retrieved from the
> > > repository using a URL like so:
> > >
> > >
> > http://host:port
> >
> > >
> > >
> >
> /fncmis/resources/DEV/ContentStream/idd_9B7C7D0A-A236-489C-A512-AB32B2E676D7/0/document1.xlsx
> > >
> > > However the server prompts the user for basic HTTP authentication
> > > credentials.  It looks like Alfresco has a way to deal with this:
> > >
> > > https://wiki.alfresco.com/wiki/URL_Addressability
> > >
> > > Is there anything similar we can do for FileNet?
> > >
> > > Sorry a bit off-topic, but if there are any IBM people out there, do
> you
> > > know if credentials can be supplied to the Daeja ViewOne applet? Maybe
> > this
> > > would be a way to deal with it instead...
> > >
> > > Thanks,
> > >
> > >
> > >
> >
> >
> >
>
>

Reply via email to