The codebeamer adapter can be configured to fetch all items in one tracker or in a baseline of a tracker. Beside this all associations between fetched items can be fetched as links.
In each batch run a new file `configuration.json` is written which describes the configuration used in the batch run.
For executed queries an additional file describing the query structure including referenced other queries is written to a `<file>_structure.json` is written.
New filtering options were introduced for the EA adapter. It is now possible to filter based on the known
include element if(
CONSTRAINT )
pattern.
A detailed description can be found
here
A new validations in the configuration indicates missing artifact references for mapped links to external artifacts.
If C artifacts can not be parsed, an error is reported that interrupts the batch run. This prevents situations where batch runs finish with no error, while no data was processed.
The Doors Next adapter will now propose projects that are under global configuration during adapter configuration. Configuring such projects will no longer cause validation errors.
The integrity adapter now supports the username and password by environment variable or system property in batch mode.
A new version of the XML Path Language (XPath) Version 3.1 is included in the XML adapter and supports even more use cases while reading values from XML files.
Queries without `.collect` are handled and a better problem description is shown.
If an adapter is configured to get custom names or positions for artifacts, some rarely used control characters in this name led to missing artifacts. Now the character is written as its unicode sequence and the artifact is visible.
In the past the data accesses were disabled if the adapter failed to load data. Even if the problem was temporary, like a missing network connection, the user had to enable them manually in the configuration. Now we ask if it should be disabled and ignore them until it’s enabled manually again, or just ignore the problem and try again on next reload.
In batch mode multiple Threads are started to fetch data from doors. In the past the number of threads depends on the number of CPU cores. For a large number of cores the overall performance of the adapter decreased.
Modules in Baselines are now ignored when they are deleted. This is consistent to the behavior in doors.
There was an issue that caused a runtime error when two configurations were compared. This has been fixed.
When configuring the query based link adapter there was an issue causing an an exception when a dataaccess was not referenced by any link type.
Under very specific conditions it was possible that the connection to the Doors client was lost or too many connection attempts were tried.