decoder2: Add support for surrogates

[Larsimusrex]

Will now decode utf-16 surrogates, used by some encoders for characters outside the bilingual plane.

println(decoder2.decode[string](r'"\ud83d\ude00"')!) // '😀'

[huly-for-github]

Connected to Huly®: V_0.6-24328

[spytheman]

What will be the result, if you later JSON encode the decoded string?
Can you please add a test for that too?

[Larsimusrex]

the json2 encoder currently handles these characters incorrectly see #25115. In my new implementation it outputs utf-8 by default unless specified otherwise

[spytheman]

Excellent work @Larsimusrex.

I am curious, what are some of the JSON encoders, that produce such output?
If they are easy to install on the CI, we can add round trip tests with them.

[Larsimusrex]

Python definitely does it by default. I think java and c# too.

[spytheman]

Python is already preinstalled on the CI, so we can add a V test, that invokes a python program, that generates a json encoded value with surrogates on stdout, then decodes it and asserts on the output.

We also support // vtest build: present_python? so that the V test could be skipped on environments that do not have Python.