RFR: 7903508: JMH: Remove the Unicode dot prefix from secondary results

Tue Jul 25 21:35:05 UTC 2023

On Tue, 25 Jul 2023 17:26:05 GMT, Aleksey Shipilev <shade at openjdk.org> wrote:

> JMH secondary results, especially from profilers, are prepended with the odd Unicode symbol. This gets inconvenient when accessing the results from Java APIs (you'd need to copy-paste that symbol), or with hand parsing, etc. 
> 
> And the profilers are not even using that prefix consistently! `perfnorm`, for example, [silently/accidentally dropped the prefix](https://github.com/openjdk/jmh/commit/88496961f43b27dc94fbaed05b0060ec21d629f0#diff-c23089eb2e9fc7b7444c5157713889c389f24329f91a2ad004cdbf24a60091f1L332), and there were no bug reports about it.
> 
> The original intent for that prefix was to sort out the profiler results after the actual benchmark results. But the reporters already separate primary and secondary results, so that trick is not even needed.
> 
> This unfortunately has an effect on tools that parse the secondary results: they would now need to drop the prefix from the secondary result name. This should be a one-time minor inconvenience during the JMH upgrade.
> 
> See before/after results in JMH Samples changes in this PR.

👍 - AFAICT we primarily use the `-rf json` output and shouldn't be adversely affected much by this change.

-------------

PR Comment: https://git.openjdk.org/jmh/pull/115#issuecomment-1650590691