I encourage you to read the code comments, reproduced here:
Yes there's a duplicate for add_access_controls_to_solr_params; but
that does not appear to be causing a problem like the duplication and
order of the now removed additional :add_advanced_parse_q_to_solr,
:add_advanced_search_to_solr filters. Those existed in their current
position and at the end of the array.
When we had those duplicates, the :add_advanced_parse_q_to_solr
obliterated the join logic for files.
<2023-10-04 Wed> Oh how I wish I wrote a bit more back in the day.
But alas, I didn't. So I must rebuild that context. Looking at the
":add_advanced_parse_q_to_solr obliterated the join logic" fragment,
I'm assuming that what I meant was "I think that the
'show_works_or_works_that_contain_files' processor chain should come
after the 'add_advanced_parse_q_to_solr' and
'add_advanced_search_to_solr' processor chains.
This assumption is bolstered by the implementation of
`Hyrax::CollectionMemberSearchBuilderDecorator` which introduces the
`show_works_or_works_that_contain_files` method. See
https://github.com/samvera/hyku/blob/07fde572f9152d513b13f71cae90dd4fdfbfba6c/app/search_builders/hyrax/collection_member_search_builder_decorator.rb#L16-L34
I encourage you to read the code comments, reproduced here:
Related to: