lib: lazyload fs to ease Electron's monkey-patching

[codebytere]

Adjusts several uses of fs in ESM to make Electron's life a little easier as we monkey-patch FS to handle our custom archive formatting.

This pattern is used in several other areas of Node.js, so it shouldn't be too out of place. This allows us to reduce our patch surface a bit.

[nodejs-github-bot]

Review requested:

• @nodejs/loaders

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 90.28%. Comparing base (d12ee6e) to head (e366a6e).
Report is 20 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #58020      +/-   ##
==========================================
+ Coverage   90.26%   90.28%   +0.02%     
==========================================
  Files         630      630              
  Lines      186170   186157      -13     
  Branches    36478    36474       -4     
==========================================
+ Hits       168042   168069      +27     
+ Misses      10982    10972      -10     
+ Partials     7146     7116      -30     

Files with missing lines	Coverage Δ
lib/internal/modules/esm/load.js	91.91% <100.00%> (-0.04%)	⬇️
lib/internal/modules/esm/resolve.js	96.17% <100.00%> (ø)
lib/internal/modules/esm/translators.js	91.57% <100.00%> (ø)

... and 34 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[aduh95]

I don't think we should do it, it goes against our goal to have (some of) Node.js internals being unaffected by userland mutation, ESM has always been on the list of "if it's affected by userland, we have a bug".
Maybe another solution would be to have some internal/fs module that all code modules would use, and Electron et al. could patch that single module, wdyt?