Since char distribution quality isn't particularly important, we can avoid including the whole <random> and just use a biased remainder formula

Since char distribution quality isn't particularly important, we can avoid including the whole <random> and just use a biased remainder formula

This seems the to be the fastest way of reading a text file into 'std::string' without invoking OS-specific methods See this StackOverflow thread: htt

Check that 'str' satisfies pattern '1[.][0...0]' (regex '1?.*0')

Check that 'str' satisfies pattern '1[.][0...0]' (regex '1?.*0')

We want to avoid modifying the state of 'std::rand()', so we use a fast & small thread-local PRNG

clang-format off

JSON '\u' escapes use UTF-16 surrogate pairs to encode codepoints outside of basic multilingual plane, see https://unicodebook.readthedocs.io/unicode_

JSON '\u' escapes use UTF-16 surrogate pairs to encode codepoints outside of basic multilingual plane, see https://unicodebook.readthedocs.io/unicode_

Copyable semantics

Copyable semantics

Copyable semantics

Copyable semantics