An export with errors will leave an adapter in an Update state, and apparently also duplicates the adapter statistics:

Scheduled connector jobs can still run when disabled, as this is no longer enforced on the Queued Connector Job Executor; A decorator to check the of the scheduler-state/connector enabled-state needs to be added/wrapped around each connector job on the ConnectorJobGenerator.

Some schema providers may create fields as single value when they are later required to be multivalue, such as the CSV connector. To do this currently, the field must be removed and re-added as a multivalue field. This may be a bit of work as the validator types available may change for some types.

Attempting to create a date relational with string priority transformation results in a "Sequence contains no elements" error appearing, and redirecting the user back to the adapter page.

The Group Target on the Multivalue.GenerateDNs UI component should be a text box, as this field is the target field against which the generated DN components will be written.

Error thrown when an attempt is made to create it:

Cannot convert type 'Unify.Framework.Web.UnifyDynamicPlugInObject' to 'Unify.Framework.Collections.GroupedNameValueCollectionKey'

No other information in the logs. This would be a minor cast issue in the service controller or view.

After selecting a base connector for the group composite with string priority transformation (4.5.5), the error "The given key was not present in the dictionary." appears. No additional information in the logs.

A number of transformations have been made potentially redundant due to some newer extended transformations that have been added over the last while. Examples include the composite key transformations - these now have controls that allow you to add one to many relationships. This makes the singular versions of the transformations redundant and confusing, and could reduce the list of transformations users have to discern between.

The only issue with doing this is backwards compatibility, and whatever mechanism we decide on for loading of old config could just pull these transformations into their newer counterparts.

Progress bars should not report percentage if a job has not yet run. The default value may be wildly inaccurate which may lead to issues about jobs seemingly getting stuck at the artificial end.

Additionally, it might be worth saying that a job is over 100% if it takes longer than the last run.

The default schema provider for CSVs should mark the key as required.