`Float.parse` regression in `org_enso_benchmarks_generated_Number_Parse_float_parse`

[GregoryTravis]

This was likely caused by #11135.

At the time, benchmark results looked like this:

But @Akirathan reports that it looks like this now: