Where the binary and html files are awesome (and the html file is even parse-able) it would be nice to print also more structured results.

I would like to implement junit-xml based report
 - small (only traces of failed tests saved)
 - many tools around to prcoess it
 - easy to generate and validate wihtout additional deps
 
I'm currently post processing index html to junit as attached below. It is moreover what I would liek jcstress to print.
Only partially sporadic is what to do for API MISMATCHes. I'm currentl marking them as paased, because they run (and cost a lot of time), instead of ignores (which then are are missing in report) - considering how hard is to get the proepr listing. Lng story short, I would liek to keep MISMATCHes as passed.