As discussed there I suggest that the check for the PublicStep is not needed, because it's the ICAT admin that controls both, the PublicStep table in ICAT and the IDS settings in run.properties. Any ICAT admin that does not want access permissions to be inferred from Dataset to Datafile and thus does not create the corresponding PublicStep, may just as well leave useReaderForPerformance at the default false. There is no point in adding extra checks to control the admin.
It is only needed that any admin knows what they're doing and that the implications of the useReaderForPerformance flag on the access permissions are well documented.
So this PR drops the check for the PublicStep and add the documentation on the implications of useReaderForPerformance instead. It furthermores add a SQL script to create the test accounts needed in the tests as a convenience, because #118 adds another account reader to this set. Finally, some spurious extra white space is removed.
This intends to amend #118.
As discussed there I suggest that the check for the
PublicStepis not needed, because it's the ICAT admin that controls both, thePublicSteptable in ICAT and the IDS settings inrun.properties. Any ICAT admin that does not want access permissions to be inferred from Dataset to Datafile and thus does not create the correspondingPublicStep, may just as well leaveuseReaderForPerformanceat the default false. There is no point in adding extra checks to control the admin.It is only needed that any admin knows what they're doing and that the implications of the
useReaderForPerformanceflag on the access permissions are well documented.So this PR drops the check for the
PublicStepand add the documentation on the implications ofuseReaderForPerformanceinstead. It furthermores add a SQL script to create the test accounts needed in the tests as a convenience, because #118 adds another accountreaderto this set. Finally, some spurious extra white space is removed.