Digital Enterprise Suite

New and improved features:

• Terms from the Knowledge Entity Modeler can now be linked in the name of shapes in diagrams.
• A new type of emitter (Audit) is now offered to all deployment types. It keeps the last 60 days of events and can be downloaded as a log file through the REST API.
• The auditing functionality of Cloud Execution was integrated with the audit emitter and now available from a single tab.
• It is now possible to select Any as a data type to mean that it is unknown at design time.
• Errors when execution services now includes more technical details.
• The Try It form now displays the technical details of errors when executing services.
• The distributed container administration page now has an area to enter a Kubernetes template. This template can be used in the Service Library to automatically generate a YAML file to deploy the container.
• When using the Rich Text editor for data mapping, it is now possible to view the HTML source code of the mapping.
• When using the Rich Text editor for data mapping in HTML source mode, any element attribute can be replaced by a feel expression by prefixing it with data-feel-expression- (ex: href=”…” -> data-feel-expression-href=”My FEEL Expr”.
• A new parameter is available on the Public API method to retrieve modeling places members including those from groups.
• When assigning groups to modeling places, these users are now available when assigning issues.
• The mention user interface in comments and issues was improved.
• Importing an invalid BPMN XML model that does not use DataObjectReference but instead depicts DataObject will now be automatically fixed.
• In the dialog to see the previous versions of a model, the current version of the model is now shown at the top of the list.
• The base container image of the automation engine and the distributed container are now based on alpine (as it was before V10) to eliminate CVE of the latest debian base image and reduce the contianer image size.
• A new Asynchronous Event was added for WebSessions.
• It is now possible to link to terms using the underscore character (_) in Discovery Accelerator.

Bug fixes:

• The description toggle button had no effect on the search for terms.
• When the automatic language translation fails, an error message is now properly displayed.
• When setting a translatable value (name, description, …) in an alternate language, this value will not be set to the main language anymore.
• In the Operation Library, re-importing the same OData definition twice duplicates the operations and data types.
• When undoing the naming of a data object reference, the state could not be restored correctly.
• When transforming a literal expression to an iterator the space after the iterator keyword was not considered.
• Weeks is not offered anymore as an option for Days and time duration constraints.
• ServiceStartedMessage event (under the service topic) for processes used the id of the process instead of the name as serviceName.
• Message correlation filters that returns a non boolean value will not trigger an error but be considered false.
• When processing a message on the message trigger REST endpoints, the HTTP context was not propagated correctly resulting in issues when trying to generate relative URLs.
• The message trigger endpoint would wrap certain messages in an extra layer on very specific data definition.
• Assigning a group as the performer of a workflow task will now send notifications to all members of that group.
• Resuming a process that was in an error state now only offer to change the inputs to the failed task.
• Renaming a model and trying to open it could fail.
• The description is now properly displayed over linked terms.
• The styling of the tags were aligned with the other modeling applications.
• Using an HTTP GET on the REST api to resume a workflow could incorrectly return the data type of the workflow input.
• The FEEL parser was not always recognizing the instance of function.
• The Identity Management page could not create all types of HTTP and ApiKey identities.
• The Webhook emitter could not use all type of identities.

Non-Backward Compatible Changes:

• The audit log format of service execution was slightly changed to align with the Cloud Events standard. Similar information is available but clients that are parsing these audit logs should modify their implementation to align with the new format.
• Client Hosted: With the change on how audit logs are defined in Cloud Execution, existing audit configuration will need to be reconfigured after the deployment.
• Trisotech Hosted: Audit logs generated before this version will not be downloadable anymore from the Cloud Execution user interface. Our support can send you these logs if required.

Bug fixes:

• Fixed an issue that prevented the proper execution of decision test cases.

New and improved features:

• Introduced asynchronous events generated by the modeling suite and the automation suite that can be connected to different types of emitters (webhooks, file, …). The documentation is available in the Trisotech Help under System Integration/Asynchronous Events.
• Improved the handling of recursive data types in the generation of services forms and Open API documentation.
• The service library try it form now supports to collapse complex types (and send a null value).
• A new option is available to display the alternative names in the terms view.
• It is now possible to rename the automatically generated decision services.
• There is now a new interface dedicated to creating and managing identities available under /management/identities.
• Added an option to define intermediate message event triggers.
• Added an internal preference to support custom user logout URL when integrating with Single Sign-On.
• It is now possible to define the time zone of the container images using environment an environment variable.

Bug fixes:

• When exporting and re-importing a workflow model containing hidden data associations, it may result in them being regenerated automatically.
• When renaming an expanded stage diagram with the page overlay visible, the page name would become blurry.
• Dragging an element from the Digital Enterprise Graph with the file data type would result in an incorrect connection to the file data type.
• The collection marker will not always show when assigning an input data to a built it collection of type.
• When animating a model, the animator would sometime open external model linked by reuse or reference.
• Fixed an issue that made workflow services OpenAPI file invalid to a spec validation.
• Fixed an issue where intermediate timer events could cause exception under certain conditions.
• Fixed an issue that was preventing to undo a change shape on a start event.
• When automating a workflow that has inputs or outputs with new lines, the generated API definition should have ignored these characters.
• When a complex data mapping is defined, user tasks form rendering could display null instead of the actual input value.
• Under certain conditions, the current value of the variables was not loaded in workflow service resume forms.
• When executing test cases based on a very different version of a model, an error at evaluation would prevent the results from being visible.

New and improved features:

• Improved the line wrapping when editing FEEL literal expressions.

Bug fixes:

• User names are now properly displayed in the group definition.
• Groups could become out of sync after a server restart.
• Retrieving the members of an empty group could result in an error.
• Renaming a decision logic would not always rename the decision.
• OData import of a type Edm.Date was incorrectly mapped to a String.
• The autocompletion of the “for” statement would sometime not offer the attributes of a complex type.

Bug fixes:

• It was not possible to undo the change of type for a start with a trigger definition.
• After the import of a DMN 1.2 model from an XML definition, re-exporting that model to DMN 1.2 could change the FEEL namespace to match the DMN 1.3 FEEL Namespace.
• Undoing the renaming of a shape while displaying an alternative language could result in the alternative language overwriting the main language for that element.
• Script tasks FEEL autocompletion could incorrectly omit certain mapped variables from the inner data input.
• The link to get more detail on FEEL expression was not pointing to the correct URL.
• Data associations could be incorrectly deleted when deleting a corresponding inner data input.
• Some execution environment could be hidden to administrators.
• Digital Distributed Containers now properly prefix the Docker-Content-Digest header to support Docker Client versions > 20.
• Multi-instance FEEL script tasks data mapping could incorrectly not give access to input data associations values.
• A UI error was preventing group rights setting for execution environments.

New and improved features:

• The CDS Hook Sandbox now allows the usage of FHIR server requiring basic HTTP Authentication.
• Selecting patients in the CDS Hook Sandbox patient browser can now be done using a single click.
• The CDS Hook Sandbox patient browser now display attachments on FHIR elements.

Bug fixes:

• The Try It button from the Service Library could encode binary file in the wrong structure that prevented accessors (name, extension, size, type, …) from being used in certain FEEL expression used in data mapping.
• The OAuth 1 authorization flow would fail for establishing new Twitter identities.
• The CDS Hook Sandbox patient browser now properly refresh when changing FHIR Server.
• Inviting users to a group could cause duplicate entries.

Bug fixes:

• The Case Modeler did not properly export the decisionRef/processRef/caseRef pointing to decisionTask/processTask/caseTask when exporting to CMMN XML when subscribing to the Digital Enterprise Graph.
• The generation of the Open API specification could fail in rare occurrences of Wofklow services.

Bug fixes:

• Some modeling products would not properly load on some version of Internet Explorer 11.

Bug fixes:

• Fixed an issue that would prevent a proper copy of certain elements in the Workflow Modeler.
• PATCH Operations are now properly imported from Open API files.
• The now() FEEL function could generate an invalid date and type string in the API.
• The Try it form XML, JSON, Excel and PDF modes were not properly uploading the information to the automation API.
• The Try it form XML template download incorrectly downloaded the JSON template.
• The import an invalid BPMN model with certain specific characteristics around data associations would fail.

New and improved features:

• When entering a definition or a note on an element, it is now possible to reuse terms using the _ character in a similar fashion that you would use the @ character to mention someone on comments and issues.
• The collection marker of Data Objects is now synchronized with the defined data type (if one is defined).
• When assigning a message to an event, its data input or output (depending if throwing or catching) will be read-only and based on that message name and data type.
• When assigning a data type to messages, errors, escalation and signals event, we now favor reusing existing message, error, escalation or signal instead of always creating new ones.
• The Import from Graph button was renamed Reuse from Graph.
• It is now possible to use the Digital Enterprise Suite users and groups as performers in the Case Modeler.
• The XML version of the REST endpoint of services automation will now output an additional xsi:type attribute for complex objects.

Bug fixes:

• When testing a model, under some condition, the time widget would always send the value null to the service instead of the entered time.
• Message events (start/intermediate/boundary) would consider only one output data mapping.
• Process using the data type year and months duration could fail to execute correctly and report an error if it needed to wait for external events (message, human interaction, …).
• It was not possible in the Cloud Execution to view the called process details of process without instance tags.
• When publishing a process, some referenced models that were not required for execution could be sent to the automation engine.
• In a rare occasion, a case model could be generated with duplicated sentry ids.
• Sometimes the term description overlay stayed visible after losing focus in the business rule view.

Bug fixes:

• SECURITY ISSUE. Under certain conditions, automation API endpoints could be used without a proper OAuth token containing the relevant automation grants.
• Automatic layout of CMMN models could layout more than one case per diagram over each other.
• For certain models including CQL or PMML models, the documentation export would fail.
• Extensions on messages and specific types of formal expression could cause an invalid BPMN XML export.

Bug fixes:

• In rare cases, adding or removing a user from a group could incorrectly generate an error message.
• The DMN 1.1 export could use the incorrect FEEL namespace for some models.
• Read only data type with an element of type Any would incorrectly display the selector for a new type.
• The delete button for rows of a read-only data type was displayed but had no effect.
• The data output mapping literal expression context did not autocomplete the extended FEEL functions such as today and now.
• The data type selector was not always properly visible in the message, escalation and error dialogs.

Bug fixes:

• A workflow model containing an extension with a duplicated ID with a process could cause an issue when saving and exporting to XML.
• The data mapping dialog was not properly accessible from the event trigger dialog.
• The QR code generated for Ideation session could be invalid depending on the suite security settings.
• Text annotations could not be translated correctly in the Case Modeler.
• Emails sent from workflow processes could fail to be sent when the email server configuration was done through environment variables.

Bug fixes:

• CMMN graph indexing of reused elements within stages duplicated some graph properties for the element.
• Removing a deleted user from a shared place does not properly affect the persisted store.
• Submitting multipart encoded form to the REST endpoint of a service when the input data type was not fully defined could cause an internal error and prevent the proper input parsing.
• Fixed details on the translation attributes of models when exporting to the XML format.
• The language selector when viewing a decision model did not always drop down properly.
• Improved the performance when dropping a large connector into the workflow modeler.
• Semantic links are now always synchronized when reusing an element by reference.

Bug fixes:

• Fixed a condition preventing some variables from not being accessible when using expression in Timer Events.

New and improved features:

• Models created with most modeling applications can be translated in multiple languages from the view ribbon.
  • A display language selector is available in the view ribbon.
  • A translation dialog allows to automatically translate terms and import/export translations from/to excel.
  • Translation languages can be added and removed from the model.
• Process instance tags expressions can now be assigned to process.
• Process instance tags can be viewed in Cloud Execution.
• Message event correlation can now be defined in the workflow modeler.
• A new API endpoint is now available to receive messages that will be correlated to the proper instances.
• When defining timer events for intermediate events, it is now possible to enter duration and date expressions.
• A new page is available to configure the Digital Distributed Container.
• Identities can now be deleted when publishing services.
• The client hosted deployment now supports the definition of an https certificate.

Bug fixes:

• Resetting the branding logo in the admin interface incorrectly reports an error.
• Request access feature is not accessible in certain configurations.
• Attachments with missing type cause an indexing issue.
• Operations inputs are now properly exported in the documentation.
• Data input mapping autocompletion does not use the full name of the data object.
• Non-text process instance names are now supported.
• Some accelerators could not be deleted from the user interface.
• A rare pattern in a BPMN XML model could cause excessive resource consumption when importing in the Workflow Modeler.