The various arguments when those were initially proposed included that road surfaces and significant chunks of parking lots (the aisles, not the parking spaces themselves) are empty 90+% of the time (true), it's surfaces that are already not natural so there are no objections of "you're covering that beautiful field with solar panels," and by using pre-fabricated units you might be able to actually put in road surface at a comparable cost in labor.
I know my initial reaction at that time was that the concept wasn't terrible - it addressed real problems. The technology might not have been there, and still might not be there, but for some carefully chosen situations they might be a viable option. The biggest obstacle that I could see is that something like that would likely need some pretty tight tolerances in the installed environment, and "road bed" and tight tolerances don't always go together so well (see "alligator cracking").
Also, regarding the criticisms that it would cost far too much to cover all the roads in the USA, just how much electricity are you expecting to consume? I feel sure that on average houses with solar have less solar panel surface area than they have driveway area and a lot of them are (hoping to) produce more power than they need for their house. Covering all roads wouldn't be necessary, most likely even covering all suitable roads wouldn't be necessary.
And regarding France doing a large experiment with this, is it a 1000km stretch or is it multiple locations in differing road conditions, up to a total of 1000km of test plots?