[JarryShaw]

• Searched for similar pull requests
• Followed the coding style
• make test passes, and a test case covers the change
• Changelog entry: N/A — centralised in docs(changelog): shared 1.5.0 changelog — long-lived, merges last (#610, #616, #617, #618, #620) #657

What is the purpose of your pull request?

• docs — documentation only

---

Description

Part of #1103, ruling 1: test narratives say what the behaviour is, not how a change altered it. This PR covers the files directly in tests/protocols/ plus tests/protocols/schema/ and tests/protocols/misc/. It excludes the round-trip and dispatch harness files, which #2165 owns.

• Rewrites:
  • "now does" becomes the present tense.
  • "on the unfixed tree" / "before the fix" becomes "without the fix".
  • "used to" becomes what must not happen.
  • "since #N" becomes "(#N)".
  • Paths described as "then in X, now in Y" name only the current one.
  • It also drops one pcapng.py:N line anchor and one count.
• Kept:
  • "now" meaning the present moment.
  • "no longer" used to describe data drift in allowlists.
  • Change-relative string literals and test names, which are code.
• Scope: only comments and docstrings change. With docstrings stripped, every changed file has the same AST as main.
• Tests run: the changed files, by group, plus the tier guard and docstring contract (140) all pass.
• Overlap: docs(tests): fifth whole-repo pass over tests (#1103) #2159 conflicts in four files over the same prose lines, and test(protocols): keep the frame-layer-limit timing bound stable on PyPy (#2160) #2162 merges cleanly. Whichever of this PR and docs(tests): fifth whole-repo pass over tests (#1103) #2159 lands second gets a rebase.

[JarryShaw]

Rebased onto main after #2159 merged: 1f3c2ac4e → aad2f0c76.

• Conflicts: four files, prose only. In each I kept docs(tests): fifth whole-repo pass over tests (#1103) #2159's edits and laid this PR's cleanup on top. test_lookup_by_name_2003_unit.py drops out, since docs(tests): fifth whole-repo pass over tests (#1103) #2159's wording already meets the ruling.
• Also folded in: the three wording points from the first review (the OSPF example, the double negative in the fix(protocols): the default engine does not dissect LINKTYPE_RAW (101) in PCAP and PCAP-NG #1502 test, the antecedent in the TCP option generator test). Phrases the first cleanup grep had missed are fixed too: "before #NNN", "Measured before …" and "already …".
• Checks: with docstrings stripped, all 62 files keep main's AST. The changed files, the tier guard and the docstring contract pass.

The earlier GOOD TO GO was for 1f3c2ac4e. A delta review of this head is running.

[JarryShaw]

Review: GOOD TO GO at aad2f0c76 (Sonnet delta). The three earlier nits are fixed, and the conflict resolutions keep #2159's edits. Every new hunk checks out against the code. The AST check, the changed files and the tier guard plus docstring contract all pass. One pre-existing figure on main (#646's 104 octets versus 88 by the stated arithmetic) is noted as a follow-up. This PR only changed its tense.

[JarryShaw]

Ready to merge at aad2f0c76. Part of #1103 (ruling 1).

• Review: Sonnet GOOD TO GO.
• CI: all checks complete, Required checks passed SUCCESS, 0 failed.
• KTLO: 12 legs, no warnings.
• Ordering: protocols leg merged into main, 1923 tests, 0 failures.