fix(pcapng): name the Name Resolution Block in its guard messages (#1038)

[JarryShaw]

Please follow the guide below

• You will be asked some questions, please read them carefully and answer honestly

• Put an x into all the boxes [ ] relevant to your pull request (like that [x])

• Use Preview tab to see how your pull request will actually look like

• Searched for similar pull requests

• Followed the coding style (make pylint, make mypy, make isort) — isort clean on pcapng.py; pylint/mypy not run (string-only source change)

• make test passes, and a test case covers the change — tests added; test_pcapng_unit.py, test_option_roundtrip_unit.py and tests/project run locally

• Added a changelog entry under docs/source/changelog/ and regenerated CHANGELOG.md, if the change is user-visible

What is the purpose of your pull request?

Tick the commit type your subject line carries.

• fix — corrects a defect
• feat — adds a feature
• perf — changes performance, not behaviour
• refactor — changes neither behaviour nor performance
• test — tests only
• docs — documentation only
• ci — workflows or build tooling
• chore — anything else

---

Description of your pull request and other information

Closes #1038. The six ns_dns* scope guards in pcapng.py, on both the read and the make side, check Name_Resolution_Block but named the systemd(1) Journal Export Block. They now say "must be in Name Resolution Block". register_record's RegistryWarning now says "name resolution record already registered". Neither message has a :manpage: role any more.

Spelling: ns_dnsIP4addr/ns_dnsIP6addr. That is the OptionType member name and value (const/pcapng/option_type.py:170,173), and the schema docstrings and read-side messages use it too. The make-side messages, the docstrings and the data-model docstrings now match.

Second commit, same defect class: I checked every bracketed option name against its own method. _make_option_if_rxspeed's two guards said [if_txspeed], and _read_option_epb_queue's length guard said [epb_packetid].

New tests cover every corrected message, and each one fails on main. The process.rst changelog pin is re-measured at 164 after rebasing onto #1046.

[JarryShaw]

Cross-review verdict on 4f55f49e8: GOOD TO GO (ran on Sonnet; author Opus).

• Every changed message now names what its guard checks. That covers the six NRB ns_dns* read/make guards, the register_record warning (__record__ holds only nrb_record_*), both _make_option_if_rxspeed guards, and _read_option_epb_queue's length guard.

• The sweep is complete. An independent AST walk checked:

  • 160 bracketed messages against the OptionType/RecordType/SecretsType member each method maps to;
  • 136 "must be in X Block" texts against their Enum_BlockType guard.

  Both checks found 0 mismatches. As a control, the same walk on main reports exactly these 7 defects.

• The spelling ns_dnsIP4addr/ns_dnsIP6addr matches option_type.py:170,173. The remaining lowercase ns_dnsip* strings are method-name suffixes and __option__ values, which are identifiers rather than messages.

• Nothing pins the old text. The remaining "Journal Export Block" hits are the real systemd block.

• The new tests fail on main and pass on head. On main: 12 failed, counting subtests. On head: test_pcapng_unit.py 97 passed, test_option_roundtrip_unit 6 passed.

• The changelog is in step. changelog_md.py --check passes, and the pin reads 164, matching the measured 164.

Not breaking: only message text changes. Exception and warning types, raise conditions, signatures and pack paths are unchanged.

Nit: the last line of the 1.5.0 entry is not wrapped to the surrounding width.