Add SpiderMonkey JSON/Regex benchmark

[posborne]

• Add spidermonkey-json benchmark using JSON.parse/stringify
• Add SpiderMonkey Regex benchmark
• Scaled all spidermonkey benchmarks to take around ~100M cycles per iteration on my test machine which is hopefully a balanced size.
• Rename base spidermonkey benchmark to markdown

--

The new benchmarks are pretty straightforward; the tricker bit here was figuring out an approach I was reasonbly happy with for bundling js inputs into the artifact. In the end, I used the same hex encoded contents in header files approach but with a top-level header pulling in the js code bits.

Adding new JS benchmarks should be fairly simple now with just adding a new directory under benchmarks/spidermnokey/js/ with main.js and other files and then providing the expected named input/stdout/stderr files based on the pattern.

I did consider switching things over to starlingmonkey but decided to leave that alone for now as it didn't appear it would be a straightforward migration and probably doesn't have too much value in practice for what insights we're looking to get from sightglass.

CC @zkat / @TartanLlama