Cost and Wear are going to be the two things that make this very difficult to implement on a large scale. Right now, it costs about $240,000 to resurface a mile of 4 lane road, not including bed prep, grinding, etc. These glass solar cells, even assuming a very super low cost cell are going to be much more expensive just for the surface material costs alone, ignoring the power runs and inverters, etc. To cover a typical 1 square mile section of road which is 27,878,400 square feet with industry standard 200 watt $200 per panel 8 square foot panels, you would need 34,84800 panels, at a minimal cost of $696,960,000. So call it $697 million give or take to replace $0.24 million worth of paving.
Then there are the wear concerns. A typical interstate aggregate asphalt paving surface will wear almost half an inch over a 5 year period with winter driving. The concrete ones wear much less, but are more expensive to put in. Glass isn't a really good wear surface for friction, and studded tires would pretty much instantly destroy it. There is also a very high temperature differential. A large double trailer truck fully loaded driving down a road at high speed generates a huge amount of heat at the road surface and at it's tires. That coupled with a glass surface prone to temperature shock might be a bad thing, even with the built in surface heaters.
All that being said, I don't think this is necessarily a bad idea used strategically. There are millions of miles of road shoulders and sidewalks surrounding important roads, and heating a roadbed during icy conditions would flat out save lives. If they were to put something like that in strips along areas that would not experience all that much wear, it would probably work fine and be effective if the cost could be made to work. I fully expect solar panel prices to keep going down as the new printable high efficient ones start coming on market, so it's not out of the realm of possibility.
The aggregate road paving process is actually pretty green as it is to resurface a road anyways. they typically grind off the outer wear layer, take that ground up material back to be recycled into new asphalt at the plant, add some new oil and new rock to make up for wear loss, then put about 95% of it back on the road again with 5% new material. There are new materials too that are starting to be used that allow groundwater to actually penetrate the asphalt, leading to less surface water in rain conditions, and also letting rain water get back into the soil below and filter down into water tables, leading to less water loss due to surface covering. Long term that's probably going to be more important for parking lots.
The most promising electricity from road surface technology I have seen is a technique that uses piezoelectric cells to harvest electricity from the moving weight of a vehicle. Similar to this: http://www.innowattech.co.il/ There is no reason that you couldn't make that compatible with existing paving practices. Asphalt certainly can be made electrically conductive with the introduction of metal powder to the mix, and the aggregate could be replaced with piezoelectric quartz crystals, pave the whole thing, call it $2-4 million per 4 lane mile, and then yeah, maybe you have something that's affordable enough to be implemented on a larger scale.
Before the roadways all get solerified though, people really should pass laws allocating funding to require solar cells on top of the millions of government owned buildings across the country at a federal, state, county, and city level. Even ignoring the ones that would be infeasible to add cells to for historical building preservation purposes, you're still talking a huge square footage area of publically owned roof space that the government already has right of way for.