the context of the problem is like this: we create workflows, we save it and after a while a new implementation request comes and we change an activity. After this the workflow instances that were saved cannot run anymore. We get this error:
StateMachine Error : Cannot convert object 'True' to type 'System.String'.
It seems that the new argument added brakes the serialization order?
You'll have to implement Dynamic Update in some fashion.
We are currently in the process of getting some infrastructure set up to update existing instances, and having lots of issues. Hopefully your scenario is easier to solve than ours!
Start here: https://msdn.microsoft.com/en-us/library/hh314052(v=vs.110).aspx
Word of caution: I've found various issues with Microsoft's provided code that required a lot of investigation to fix.
Related
I've been working with Stateful Sessions (KieSession) so far and have managed to get my project running as desired using Scala with a few Java wrappers. I am now trying to switch over to StatelessKieSessions. Based on the documentation I found, I've managed to run the following to insert objects/collections into the session, fire the rules on them and update the facts:
val cmd = CommandFactory.newInsert(myObject, "myObject")
val result = ksession.execute(cmd)
When I print result (which is of class org.drools.core.common.DefaultFactHandle), it shows the structure of the desired fact, updated as expected, preceded by something of the sort "fact 0:1:2050275256:1971742898:2:DEFAULT:NON_TRAIT:"
The documentation says that I should be able to write something like result.getValue("myObject") however this option doesn't seem to be available in Scala. (https://docs.jboss.org/drools/release/6.0.0.Beta1/kie-api-javadoc/org/kie/api/runtime/StatelessKieSession.html)
I understand that Scala-Drools interoperability hasn't been provided in full, however does anyone know of a way to extract updated facts from within a StatelessKieSession or a DefaultFactHandle containing it?
What you get from this execute command is the fact handle of the newly inserted fact. The object therein would still be the one you have inserted, updated or not. You'll have to investigate whether this is something you can use in Scala or not.
There is no command to retrieve all facts that have been changed during the execution of a session. You'll have to monitor this, using some of the available technique.
There's not much to be gained by running a "Stateless Session". If you can achieve what you want using a regular (stateful) session, leave it at that. The stateless session may have its advantages, but don't grapple with it from Scala.
I have issues with the RESTXQ implementation in exist-db.
I think it might be the RestXQTrigger which is not working correctly.
The problem: I deleted (via the Dashboard) a collection including RESTXQ services inside several .xqm files. However, the services are not unregistered and are still available, even after restarting eXist.
Is there any way to force this unregistring, I mean other than recreate the previous collections/files and delete each .xqm files one by one (this way, the trigger seems to work) ?
RESTXQ in eXist at the moment only implements the Document Trigger events and not the Collection Trigger events. This is just a limitation which needs to be resolved when there is time to implement it.
There is an XQuery module provided with eXist in the namespace: http://exquery.org/ns/restxq/exist. The functions in this module enable you to manually manipulate the RESTXQ Registry. You can enable it in $EXIST_HOME/conf.xml. If you then restart eXist and re-build the function documentation you should be able to see the documentation in the function browser app for these functions. In particular you most likely want the functions:
exrest:deregister-module(xs:anyURI("/db/my-module.xqm")) and exrest:register-module(xs:anyURI("/db/my-module.xqm")).
There are also functions for registering and deregistering individual functions from a module, which are called register-resource-function and deregister-resource-function they are similar to above but take a second argument which is a function signature (as a xs:string) in the form of qname#arity e.g. "fn:substring#2"
You can stop the database, and manually remove the registry file $EXIST_HOME/webapp/WEBINF/data/restxq.registry
(I am also using .NET 4.0 and VS 2010.)
I created a function import returning a complex type, as explained at http://msdn.microsoft.com/en-us/library/bb896231.aspx. The function import and new complex type appear in my .edmx file and in the Designer.cs file. However, the function does not appear when I view the service in the browser, and when I add or update a service reference in the client project, the function does not appear there either - as is to be expected, given the first result.
Creating an imported function and using it seems conceptually very simple and straightforward, and one would think it would just work, as Microsoft's step-by-step instructions appear to suggest: http://msdn.microsoft.com/en-us/library/cc716672.aspx#Y798 (which article shows the SP returning entity types - I tried this also, and it doesn't work for me either).
This blog post shows the addition of a method to the DataService class, which Microsoft's instructions omit: http://www.codegain.com/articles/wcf/miscellaneous/how-to-use-stored-procedure-in-wcf-data-service.aspx I tried adding one method returning a list of entity types and another returning a list of complex types, and still had no success. I still could not access the functions, either directly via the browser or from the client application via a service reference.
Thanks in advance for any help with this.
config.SetServiceOperationAccessRule("*", ServiceOperationRights.All);
MS would do well to add a note to the walkthroughs stating that the above bit of code must be there. (It may be better to enable each operation explicitly than to use "*".)
http://www.codegain.com/articles/wcf/miscellaneous/how-to-use-stored-procedure-in-wcf-data-service.aspx shows that line of code. Also, something it is there in the code, commented out, when one creates the WCF Data Service. Some of us like to delete commented-out code that we aren't using and that seems irrelevant - perhaps doing so a bit prematurely, sometimes.
We have a long running workflow which uses SQL tracking service (.Net WF 4.0). In the next update, we would like to introuduce a public property in one of the arguments of the workflow. Since this is breaking change, the persisted workflow inatances throw the following error on re-loading:
System.Runtime.DurableInstancing.InstancePersistenceCommandException: The execution of the InstancePersistenceCommand named .. LoadWorkflow was interrupted by an error.
InnerException: System.Runtime.Serialization.SerializationException: 'Element' '_x003C_BookmarkName_x003E_k__BackingField' from namespace '...' is not expected. Expecting element '....'
I understand this is a typical versioning issue and one of the recommendations I noticed on some of the sites is to override OnActivityExecutionContextLoad method and fill in the missing values. But I am not sure where and how to do this! OnActivityExecutionContextLoad is declared in System.Workflow.ComponentModel.Activity (.Net 3.5?) whereas what we have is a code-based top-level custom activity derived from System.Activities.NativeActivity (which receives the argument in question). Can something be done in this class to initialize the missing property of the argument?
All suggestions are welcome :)
I have two workflow foundation (.NET 3.5 SP 1) application which compiles fine. It has two declarative rule conditions in it. Both are very simple, examples:
- !this.Shutdown
- this.Runtime == null
The project builds correctly without error or warning, however when I run the application it fails with the exception: error 1342: Activity 'whileNotShutdown' validation failed: Can not find the condition "Shutdown not requested".
In the other application I am using a more complex ruleset in policy which generates the following exception: error 1398: Activity 'Defaults' validation failed: Can not find the rule set "Rule Set1". Check if rule set is defined in the rules file.
I have tried removing the .rules file and recreating it by recreating the two rules and that did not help.
The question is: What is wrong that is causing this error?
Some things I have tried:
Placing the workflow rules in the larger workflow makes no difference to the issue.
The issue is effecting all workflows in the solution now (even those that used to work).
Using a code rule condition doesn't have the same problem - the issue is related to the rules file somehow.
Rebuilding workflow didn't help
Declarative workflow rules are not checked by the compiler so there is no help there. When you reference 'this' in rule it means the workflow. So your properties Shutdown and Runtime must be declared at the workflow level. I presume the Runtime refers to the WorkflowRuntime and that is kind of unusual, and not recommended, to have inside of your workflow so there may be a design issue here.
Eventually found the cause - it was caused by spaces in the class name, which causes the rules file to have spaces in it, which seems to cause the issue.
For example this will cause the error:
alt text http://img512.imageshack.us/img512/661/clipboard01wdu.png
However this will not:
alt text http://img140.imageshack.us/img140/9366/clipboard02t.png