Disable Unicode support for C# to match Java, Rust, C, C++

[danmoseley]

The C benchmark does not enable Unicode support (it doesn't pass PCRE_UTF* option) nor does the C++ Boost benchmark (it doesn't use u32regex) and Java doesn't either (it doesn't set Pattern.UNICODE_CHARACTER_CLASS nor add the ?U option). Rust (since #21) explicitly passes .unicode(false).

C# however is laboring with matching Unicode character classes. To be fair, this adds RegexOptions.ECMAScript. The flag has other effects, but disabling Unicode is the only effect that is relevant to this benchmark - the other ones relate to aspects of the pattern that aren't present in this benchmark's patterns.

This increases C# perf by about ~45% on 3.1. It's still some distance off Rust/C/C++ but it brings it in line with Java (at least with the improvements in .NET 5.0) which is nice.

cc @stephentoub

[mariomka]

Finally, the benchmark will go with the default configuration. I merged Rust update but I removed .unicode(false) statement.
But I have decided to create a branch allowing code and setting optimizations. Check #27, please.

[danmoseley]

Sounds good @mariomka