Workflow process definitions (both released and unreleased) can be imported onto the relevant database table using the standard build database target.
These workflow process definitions must be stored in XML files in a workflow subdirectory under the relevant Cúram server component directory (e.g. ...\EJBServer\components\core\workflow for the core component or ...\EJBServer\components\Appeal\workflow for the Appeal component etc.).
Each component in the Cúram application can have a workflow directory containing the process definition XML files relevant to it. Any process definition files stored in these workflow directories will automatically be imported when the build database target is executed. If the process definition files are not valid or if the name and version of the definitions do not match those used in the filenames, the import will fail.
The workflow process definition XML files on the file system must follow a strict naming convention. This is as follows: Process Name_vProcess Version.xml where:
The same version of a process definition can exist in multiple components in the Cúram application. The version imported will always be taken from the component with the highest component order precedence. Component order precedence is configured using the COMPONENT_ORDER_PRECEDENCE environment variable.
Each process definition when imported will be assigned a new process definition identifier that is unique for the database it is imported onto. Different versions of the same process definition will be assigned the same unique identifier and only one unreleased version of a process definition can be imported. To handle invalid workflow process definitions loaded during the build database target, strict validations are in place in the workflow engine. These ensure that a workflow process definition cannot be loaded into the process definition cache and executed unless it passes all of the process validations first. These validations are described in the earlier chapters of this document.