Reverse-string: add a test case with a longer string

[norbs57]

I recently mentored a student on ReverseString who had started off with a naive algorithm that had quadratic complexity. It was a bit disappointing to see that in the benchmarks, the speedup was not that great. I noticed that all the sample strings in the test suite are rather short. Maybe we could add at least one longer string so that the performance gains of a better algorithm become more impressive?

[SaschaMann]

Do you know how long the string would have to be for it to be noticeable in benchmarks?

Usually that would be something that an analyzer or mentor would discuss with the student though, not part of the test suite.

[norbs57]

With the current test suite (in Go), the speedup is about factor 2. After adding a 100 character string to the test, the speedup is closer to a factor 4. (Usual disclaimers about micro-benchmarks apply.)

[norbs57]

To be more precise - the Go benchmark test for ReverseString currently takes the average of the execution time of all test cases defined in canonical-data.json. If the benchmark data consist of a single 100-character string only, then the speedup is about a factor 5.

[SaschaMann]

I think a 100 character string would definitely be reasonable to have. I'd be in favour of adding it if you want to open a PR.