Robust version of `Prism.parse(code, version: 'current')`

[tompng]

In IRB, it's better to parse ruby code with current ruby version for usability, but doesn't strictly require current version.
It basically work except for few edge cases: Error tokens (valid in latest) won't be rendered RED. (example: "if a\n&&b\nend")

Parsing with Prism.parse(code, version: RUBY_VERSION >= '3.3' ? '3.3' : 'current') is not enough.
Lower bound(3.3) may change in the future, RUBY_VERSION may exceed upper bound(3.4 (prism-1.6) or 4.1 (prism >= 1.7)) that may block updating ruby version in a rails app.

It would be helpful if there is an easy way to use current version robustly (like Prism.parse(code, version: 'nearest' or 'closest')), or a way to get supported version range.

[eregon]

Parsing with Prism.parse(code, version: RUBY_VERSION >= '3.3' ? '3.3' : 'current') is not enough.

To clarify, do you mean the inverted condition:

Prism.parse(code, version: RUBY_VERSION < '3.3' ? '3.3' : 'current')

So it uses syntax version 3.3 on Ruby < 3.3, and otherwise matches RUBY_VERSION?

[eregon]

or a way to get supported version range

Could be literally a Range used like this:

RUBY_VERSION.to_f.clamp(Prism::SUPPORTED_SYNTAX_VERSIONS_RANGE)
i.e.
> 4.5.clamp(3.3..4.1)
=> 4.1

Interestingly it also seems to work with strings:

> "3.4.5".clamp("3.3".."4.1")
=> "3.4.5"
> "4.4.5".clamp("3.3".."4.1")
=> "4.1"

so could be RUBY_VERSION.clamp(Prism::SUPPORTED_SYNTAX_VERSIONS_RANGE).

But version: 'nearest' or 'closest' seems quite a bit more convenient.