Identifying Content Server Documents

Hi,

We have successfully deployed the web services to do many automated processes, and they are working nicely. Currently, we are using les 9.7.1, but we plan to move to version 10 very soon.

We do, however, have two issues that we would like possible solutions for:

1: When we try updating a project node (which we have found by utilizing the DocumentManagement service), we get an error. This error also occurs when using the lltreeview client and is easily reproducible. I asked both on CW and mailed support, but neither could provide me with an answer so far. The exception thrown is: "Required features missing for updating a Project node". Any idea what might cause this behavior?

2: We have an own module handling interoperability with an application towards Content Server. This works perfectly when opening documents through the module, but when opening via web, we don't have a 100% safe way of pairing the documents towards the correct document in CS. Is there any best-practice on how this can be achieved?

Best regards,
Rolf

Find more posts tagged with

Comments

Andreas_Probst

Hi Rolf,

I forwarded your question to the developers of Content Web Services and received the following answer:

For 9.7.1 the GetNode did not populate some necessary NodeFeatures for Projects. This is done in 10.0.0, so UpdateNode can be called on Projects. For 9.7.1, the project methods in the Collaboration service handle the node features. Depending on what you are doing, you could either use Collaboration service or manually set the NodeFeatures (I believe the former will be a more reasonable solution).

Regarding the second question: I'm afraid we don't really understand it. Are you talking about an OScript module deployed in Content Server? Did you define an own document type (different from 144)? Do you have specific behaviour on your documents, which is missing in the GUI?

Kind regards,

Andreas Probst

_________________________________________________

Andreas Probst
Manager Enterprise Library Infrastructure

Open Text                         Phone:   +49 (0) 89 4629-1468
Werner-von-Siemens-Ring 20       Fax:     +49 (0) 89 4629-33-1468
D-85630 Grasbrunn                E-mail: aprobst@opentext.com

Place of Incorporation / Sitz der Gesellschaft: Open Text Software GmbH, Werner-von-Siemens-Ring 20, 85630 Grasbrunn, Germany | Phone: +49 (0) 89 4629 0 | Fax: +49 (0) 89 4629 1199 | Register Court / Registergericht: München, Germany | Trade Register Number / HRB: 179298 | VAT ID Number /USt-ID: 118 656 356| Managing Director / Geschäftsführer: Walter Köhler

Rolf-Henning Klungsøyr

Hi Andreas,

Great, then I will postpone the project updatenode problem until we are over on 10.0.0 in a couple of weeks.

Considering the second question: We have a plug-in to another program which automatically loads when the application starts. We do not have any customiztions in Content Server, but may consider that if it's necessary (we do have some competence in builder / web services creation). In general, what we need is to identify which document has been opened. Currently, we search the repository by the document name (since most of our documents are under ADN), which works ok right now. However, a more stable solution would be greatly appreciated, and we were wondering of the best practices to achieve this. Is there any documentation available on how we can handle content server documents in external applications for example? The advantage with the solution I outlined above is that it also works if an employee works offline on a document and wants to do content server operations on it (since it is not dependable on any events from the content server side etc), but maybe there are other ways to achieve this?

Best regards,
Rolf

Quoted Andreas Probst (aprobst) on 2010/11/25 11:59:

Hi Rolf,

I forwarded your question to the developers of Content Web Services and received the following answer:

Kind regards,

Andreas Probst

_________________________________________________

Andreas Probst
Manager Enterprise Library Infrastructure

Open Text                         Phone:   +49 (0) 89 4629-1468
Werner-von-Siemens-Ring 20       Fax:     +49 (0) 89 4629-33-1468
D-85630 Grasbrunn                E-mail: aprobst@opentext.com

Andreas_Probst

Rolf, I'm still trying to understand the use case or issue. I understood you have an application (Java, C#, ?), which acts as a client of Content Web Services, i.e. uses SOAP as communication protocol.
Now you want to know, which document is opened. Do you mean, you would like to know which document is "reserved"?
Would you like to know all reserved documents in Content Server, which can be a huge amount?

The users of the client application can work offline, i.e. without network connection to the Content Server. Still they would like to do Content Server operations. Which operations should this be? Would you record those operations in order to replay them later once the user is online again?

Or could it be, the use case of your application is the following?

Reserve document
Download document into a local file
Let your application or your user do something with the local file, e.g. editing
Upload the file as new version of the document
Unreserve document

Andreas

Rolf-Henning Klungsøyr

Hi,

Yes, the last point is one of the use cases. In most cases the users will work online, and content server commands are issued normally. But in cases of i.e. travel, network access is not always available (or secure enough), so a user may store a content server file locally on the laptop.

When the user returns on-site, or has an available internet connection, he may want to do actions towards Content Server on this document, but I can't know with 100% certainty which document he is working on. Is there a way of identifying this? Is there i.e. any tags on the file (properties or otherwise) that indicates the identification numbers (content server and version) of the file? Or is there any other method available that allows me to identify this file?

And currently, this problem also applies when opening a document from the Content Server web interface. The document is automatically opened in the desired application and the plug-in is loaded, but I still don't know for sure which file it is. Currently I'm using a search tool based on document name (since most documents are under ADN, and hence docnames are unique). If zero or more than one hit is found, the user is presented with a dialog helping him/her to choose. Is there any events or arguments that I could use in stead? I see for instance that in the temp internet files, there are files that have filename patterns of docname.extension?func=doc.fetch&livelinkid=xxxx&versionid=****, but I can't see this information in the application. Of course, I could parse the temp internet files folder, but I guess there are better ways of doing it.

Do you know how other applications are handling this? Or is there any documentation available on the subject from OT?

(And yes, it's an application that uses the SOAP protocol over Content Web Services written in C#.)

Best regards,
Rolf

Andreas_Probst

Normally an application would download a document via its nodeId. The application remembers the nodeId until it is used to upload the document back as a new version. Then the ID can be forgotten.

Reading the whole thread again, it sounds like there are cases, where your application doesn't load the document, but plugs into an application, that is automatically started from Windows once the document is downloaded and opened via the web GUI. In this case you just know a temporary filename, but neither the CS document nor the ID.

Is my summary now correct?

In this very case, when the user is back online: could she use the Content Server web GUI to add a new version? The user should know where to put the file.

Could you rely on the temporary filename? It contains the IDs that you need. I guess not, if the user saves the file with a proper name somewhere else.

Would it be an option to forbid web GUI access for these particular documents or below a particular path?

Andreas_Probst

Hi Rolf,

I forward another suggestion from one of the Content Web Services developers:

I’m not sure how a file alone could ever map to a node in Content Server unless you renamed it to include a unique ID (ELID or DataID or GUID). This would be easy to break and might be a vector for attacks (i.e. trick a user into unknowingly clobbering the content of something they have permissions to). Maybe only allow the upload if the node has been reserved by that user?

Andreas

sgrasley

Hi, Rolf.

I just wanted to highlight that a GUID for Nodes and Versions are available only in CS 10.0.0. These can be retrieved using separate methods on the DocumentManagement service.

Thanks,

Scott