fix: incorrect resolution when using shared resolvers with different main_fields

[Boshen]

The modified test case

fn test() {
    let f = super::fixture().join("restrictions");

    let resolver1 = Resolver::new(ResolveOptions {
        main_fields: vec!["style".into()],
        ..ResolveOptions::default()
    });

    let resolution = resolver1.resolve(&f, "pck2").map(|r| r.full_path());
    assert_eq!(resolution, Ok(f.join("node_modules/pck2/index.css")));

    let resolver2 = resolver1.clone_with_options(ResolveOptions {
        main_fields: vec!["module".into(), "main".into()],
        ..ResolveOptions::default()
    });

    let resolution = resolver2.resolve(&f, "pck2").map(|r| r.full_path());
    assert_eq!(resolution, Ok(f.join("node_modules/pck2/module.js")));
}

failed to revolve due to package.json caching the value of main_fields: vec!["style".into()], the second resolution main_fields: vec!["module".into(), "main".into()] would yield nothing because package.json#mainFields still holds the value of "style".

This PR changes caching the options to retrieving the values dynamically. I need to continue the work on all the other fields in #135

But I need to hotfix the main_fields change first for Rspack.

[codecov]

Codecov Report

Attention: Patch coverage is 91.66667% with 1 lines in your changes are missing coverage. Please review.

Project coverage is 94.94%. Comparing base (ee22b66) to head (0224a85).
Report is 2 commits behind head on main.

Files	Patch %	Lines
src/lib.rs	50.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #134   +/-   ##
=======================================
  Coverage   94.93%   94.94%           
=======================================
  Files          11       11           
  Lines        2212     2216    +4     
=======================================
+ Hits         2100     2104    +4     
  Misses        112      112           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[codspeed-hq]

CodSpeed Performance Report

Merging #134 will not alter performance

_{Comparing dont-cache-optins (0224a85) with main (db051d5)}

Summary

✅ 2 untouched benchmarks