240329 align papo clinline date inputs

[gm-ebs-ext]

No description provided.

[ml-ebs-ext]

This single-line change is an improvement, but the ICF variable is not the only one that leads to the "blank timeline" problem.
It's conceivable that a thorough review of the codebase could unearth all the places that require tweaks such as the one in this PR.

I personally would favor an alternative approach, which can be summarized as:

• Map all date+time dataset columns to dates, taking care that columns used as end_date inside by the module are rounded up to the next day.

This would take care of the whole family of errors in a single place.

There is already something similar in this codebase (T_honor_map_to_flag) that maps "Y/N" columns, as specified in the module API, to logical.

This exercise would be easier in any of our other codebases, because they have a second-generation module wrapper, so I think a good prior step would be to move this codebase to CM and TC.