Brian Ward

Does this affect https://github.com/stan-dev/stan/issues/3330 or is it unrelated?

Huh, I didn't realize lmultiply even really existed in the math library. I'm pretty sure the stan compiler still just rewrites it to `multiply_log`. They definitely should be consistent between...

> Most of this PR is if-constexpr-template magic that goes way over my head. This is, luckily, the part that I feel comfortable reviewing, so between the two of us...

If I'm reading this correctly, I think using the keyword `multiplier` for this would be a mistake. I think that would be interpreted by most users as some sort of...

The new `-f` flag has some unituitive behavior. I'd expect the following commands to all be equivalent: ```shell $ ./bin/stansummary ./src/test/interface/example_output/bern1.csv -f ./src/test/interface/example_output/bern2.csv $ ./bin/stansummary -f ./src/test/interface/example_output/bern1.csv -f ./src/test/interface/example_output/bern2.csv $...

I'm in transit today so hard to review, will try to approve tomorrow morning. It does look like there are a few makefile changes around the print utility that still...

I'm hoping the above is the last issue, but it's hard to say. I don't know if it is necessary to touch the code as much as this PR does...

Here's an example of what I mean: https://github.com/stan-dev/cmdstan/compare/stansummary-new-stats Before updating the tests, this diff is only at +59-76, and it adds the ess_bulk, ess_tail, MAD, and new R-Hat to stansummary....

> made test more strict, but differences between R and C++ are to be expected. I am not convinced that they are expected (I recently wrote a javascript version of...

What previous implementations of diagnostics seem to do is just run the code once to get values and then "freeze" those in the test. I think #3312 deleted some tests...