assert,util: improve array comparison

[BridgeAR]

Sparse arrays and arrays containing undefined are now compared faster using assert.deepStrictEqual() or util.isDeepStrictEqual().

Update: I added another commit that also improves the comparison performance for unequal numbers.

Local benchmark (more configurations and higher iterations to show that even regular arrays improve by ~1%. This should not matter for regular runs, so I kept the benchmark simpler in here):

                                                                                                                         confidence improvement accuracy (*)   (**)  (***)
assert/deepequal-simple-array-and-set.js method='deepEqual_Array' containsUndefined=0 strict=1 len=10000 n=4000                 ***      1.39 %       ±0.67% ±0.90% ±1.17%
assert/deepequal-simple-array-and-set.js method='deepEqual_Array' containsUndefined=1 strict=1 len=10000 n=4000                 ***     24.22 %       ±0.92% ±1.23% ±1.60%
assert/deepequal-simple-array-and-set.js method='deepEqual_Set' containsUndefined=0 strict=1 len=10000 n=4000                            1.41 %       ±1.71% ±2.28% ±2.98%
assert/deepequal-simple-array-and-set.js method='deepEqual_Set' containsUndefined=1 strict=1 len=10000 n=4000                            1.56 %       ±1.57% ±2.08% ±2.71%
assert/deepequal-simple-array-and-set.js method='deepEqual_sparseArray' containsUndefined=0 strict=1 len=10000 n=4000           ***     28.45 %       ±0.80% ±1.06% ±1.38%
assert/deepequal-simple-array-and-set.js method='deepEqual_sparseArray' containsUndefined=1 strict=1 len=10000 n=4000           ***     28.36 %       ±0.85% ±1.14% ±1.48%
assert/deepequal-simple-array-and-set.js method='notDeepEqual_Array' containsUndefined=0 strict=1 len=10000 n=4000                *      1.70 %       ±1.53% ±2.05% ±2.70%
assert/deepequal-simple-array-and-set.js method='notDeepEqual_Array' containsUndefined=1 strict=1 len=10000 n=4000              ***     23.54 %       ±0.95% ±1.26% ±1.64%
assert/deepequal-simple-array-and-set.js method='notDeepEqual_Set' containsUndefined=0 strict=1 len=10000 n=4000                        -1.20 %       ±1.35% ±1.80% ±2.36%
assert/deepequal-simple-array-and-set.js method='notDeepEqual_Set' containsUndefined=1 strict=1 len=10000 n=4000                         0.33 %       ±1.81% ±2.41% ±3.14%
assert/deepequal-simple-array-and-set.js method='notDeepEqual_sparseArray' containsUndefined=0 strict=1 len=10000 n=4000        ***     28.05 %       ±0.86% ±1.15% ±1.49%
assert/deepequal-simple-array-and-set.js method='notDeepEqual_sparseArray' containsUndefined=1 strict=1 len=10000 n=4000        ***     28.11 %       ±0.98% ±1.31% ±1.70%

[nodejs-github-bot]

Review requested:

• @nodejs/performance

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 90.22%. Comparing base (a4f556f) to head (f46895a).
Report is 9 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #57619      +/-   ##
==========================================
- Coverage   90.24%   90.22%   -0.02%     
==========================================
  Files         630      630              
  Lines      185129   185053      -76     
  Branches    36238    36214      -24     
==========================================
- Hits       167064   166960     -104     
- Misses      11045    11055      +10     
- Partials     7020     7038      +18     

Files with missing lines	Coverage Δ
lib/internal/assert/myers_diff.js	89.26% <100.00%> (+0.06%)	⬆️
lib/internal/util/comparisons.js	99.63% <100.00%> (-0.01%)	⬇️

... and 19 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.

[ljharb]

it's very surprising to me that checking undefined is faster, especially in the case of getters, but if the benchmarks say it is, so be it :-)