i'm hacking on that since it's actually not too hard to get the

RunTracker

as a v2 engine param and start spawning workunits to write to

Be warned that you’ll end up with a huge set of data which will crash things like the HTML reporter and the JSON uploader if you just dump all of the v2 workunits as if they’re v1 ones

oh yes! i realized that the

WorkUnitStore

was doing a lot more intense work than e.g. making

FileBackedRWBuf

instances for outputs, and it was storing that in a persistent and queryable way. i interjected because i think we're finding that workunit stream outputs (the thing i would love to make a little more useful) are orthogonal to v2 workunits, which track the progression of nodes being evaluated