[wmundev]

Formatting an all-zero NANPA number for a call from another NANPA region now preserves its retained raw input, matching Java and the existing netstandard2.0 implementation. The modern span shortcut previously bypassed Format(NATIONAL)'s raw-input fallback, reformatted the zeros, and could append an extension that the fallback omits.

For example, parse 000-000-0000 with ParseAndKeepRawInput(..., "US"), then call FormatOutOfCountryCallingNumber(number, "US"): the old modern implementation returned 1 000 000 0000; Java and this fix return 1 000-000-0000.

Severity: P3, high confidence. This affects all-zero/invalid numbers accepted by parsing or constructed with the builder. The fix adds the missing fallback to the NANPA span shortcut and retains the current normal-number path. No public API, metadata, regex, dependency, or Java source changes.

Java parity evidence:

• Java baseline: 33be5664343676b27761213d8f5500d682568ea7. formatOutOfCountryCallingNumber calls format(number, NATIONAL) for NANPA calls; format preserves raw input for zero national numbers. The existing upstream testFormatUSNumber covers this fallback.
• Freshly compiled unchanged Java passed 20 focused checks: builder and parsed zero numbers, US/Bahamas origin, extension handling, 150/10,000-character raw strings, empty-raw and ordinary-number controls.
• Differential corpus: 1,109 cases and 112,009 result entries per implementation across formatting overloads, custom patterns, carrier/mobile/out-of-country/original/alpha formatting, type/validity/possibility, area/destination codes, and truncation. Modern C# fixes all 12 affected entries; no previously matching entry changes to a mismatch. All three hash-verified C# target assemblies agree.
• 111,963/112,009 entries agree with Java. The remaining 46 pre-existing differences are explicitly set zero-country-code presence semantics (setCountryCode(0) is present in Java, HasCountryCode is false in C#); deferred under the one-fix limit. This is not a whole-library parity claim.
• Shipped metadata versions differ (C# 9.0.42 / Java 9.0.41). Comparisons instead consume the identical synthetic fixture, blob 14fdad223881d77c7679cbf6e6e0eecc52e036f5, through the original builders. No metadata contents were manually inspected or edited.

Validation:

• Six new regressions failed before the fix on each modern TFM; full solution now passes 1,186 tests (542 library + 51 extensions per net8/net10), zero failures/skips.
• All-TFM build: zero warnings/errors, including modern trim/AOT analyzers.
• Release package validation against the published baseline passes. A clean compilation at the same external path reproduces the SHA256 of all three packaged library DLLs.
• External BenchmarkDotNet 0.15.8 harness directly exercises the affected method (in-process, three warmups/five requested iterations, one launch each): normal 743.1 → 754.7 ns / 160 B; retained-raw zero 220.1 → 28.8 ns / 56 B; zero without raw 221.7 → 244.0 ns / 56 B. Samples include outlier removal; zero-without-raw is approximately 10% slower in this sample. No general throughput or no-regression claim; CI benchmarks remain to run.
• Diff whitespace check recognizes the test file's existing CRLF convention. No Linux execution, runtime trim/AOT publishing, complete Java JUnit suite, concurrency proof, or exhaustive production-metadata comparison.

C# review baseline: 2a0bb649a22a518ee74133fdbcebcf39f8ebbbd4. Both baseline histories were unchanged since the preceding successful review; open upstream PRs were checked for duplicates.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.08%. Comparing base (2a0bb64) to head (669c81b).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #517   +/-   ##
=======================================
  Coverage   88.07%   88.08%           
=======================================
  Files          43       43           
  Lines        3891     3893    +2     
  Branches      993      994    +1     
=======================================
+ Hits         3427     3429    +2     
  Misses        264      264           
  Partials      200      200           

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

🚀 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.