I'm not sure this would qualify as an idea or not, but here's what I'd like to do even if it means writing separate tool(s).

Assuming I can generate the list of source files which have no coverage (because no tests defined for them), how would I go to generate a gcovr output file using the current version of gcovr's json format such that later I can simple use --json-add-tracefile with 'empty' coverage stats combined with the actual json files which have real coverage from tests in them to get a combined report showing 0 for all the missing files.