[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, round 6: the sixth whole-repo prose pass over docs/source/changelog/, with the root CHANGELOG.md regenerated by util/changelog_md.py.

• Breaking entry for refactor(pcapng)!: rename PCAPNG.context to section_context (#2172) #2205. PCAPNG.context returned the section Context from 1.0.0 through 1.4.x (it is in v1.0.0). That is PCAPNG.section_context now, with no alias, and context is the inherited ContextRegistry. The entry has a migration line.
• Round-5 entries for user-visible merges: fix(arp): reject bool and non-bytes addresses for non-IP ptypes (#2166) #2182, fix(tcp): keep Quick-Start rate an int and a Fast Open cookie request bytes (#2167) #2185, fix(pcapng): correct flags-option length messages and IPv6 maker defaults (#2168) #2179, fix(protocols): correct 'timestmap' and 'movile' in two messages (#2171) #2189, fix(ipx): stop dispatching IPX payloads through the IP protocol registry (#2169) #2187, fix(hip): build the ECHO_RESPONSE_UNSIGNED schema in its maker (#2170) #2188, fix(corekit): pass the copy to ListField callbacks and size SchemaField templates from it (#2173) #2190, fix(corekit): export SchemaField, SwitchField, PortEnumField and nested_packet_context (#2174) #2191 and fix(foundation): validate a registrar's code before registering it (#2186) #2193. Test-only, CI-only and docs-only merges are left out, per CONTRIBUTING.md. The intro's PR range ends at refactor(pcapng)!: rename PCAPNG.context to section_context (#2172) #2205.
• Process narrative stated as behaviour: the fix(corekit): stop EnumRegistry.get's default from minting #868, pcapng: journal entry split on b'\n\n' shreds binary data that contains it #723, fix(const,vendor): bring 8 bespoke registries onto EnumRegistry, stop them minting #869, FieldBase.unpack pads a short read with rjust regardless of byte order, silently corrupting little-endian values #604 and fix(const,vendor): return the unregistered member instead of re-entering _missing_ (#866) #867 entries.
• Tests run: test_docs_*, test_conventions_doc_claims, test_changelog_md, test_rfc_anchor_fragments and test_issue_refs_in_comments give 182 passed. The other changelog-reading tests give 242 passed.

[JarryShaw]

Review: NEEDS CHANGES at c27208ece (Sonnet).

• pcapng: journal entry split on b'\n\n' shreds binary data that contains it #723 entry is false against the code. "a bad terminator ends that entry alone, so the well-formed entries behind it are kept". But at origin/main, SystemdJournalExportBlock.post_process raises ProtocolError for a binary field with a bad terminator or length. _read_block_systemd then keeps the whole block as captured (question(pcapng): should the Journal Export binary-field clamp go strict like #1325? #1406). The neighbouring Every systemd journal field after the first binary field is discarded: entry_data.read() reads to EOF where it means to skip one newline #704 entry's "ends the entry with a SchemaWarning" is stale for the same reason.
• "unfixed tree/code" at the three names used in string annotations are never imported, one of them raising NameError under get_type_hints; and MANIFEST.in now contradicts itself about what an sdist ships #642, Every systemd journal field after the first binary field is discarded: entry_data.read() reads to EOF where it means to skip one newline #704, fix(toolkit): key pyshark link-type resolution on frame.encap_type #850 and frame.py:199 seeds packet['bytesorder'] where readers expect 'byteorder', and no big-endian PCAP fixture exists to catch it #605 entries is a banned token. Each becomes "Without the fix, …", keeping its failure strings.
• Minor: the fix(const,vendor): bring 8 bespoke registries onto EnumRegistry, stop them minting #869 OptionType sentence still mixes tenses. The fix(arp): reject bool and non-bytes addresses for non-IP ptypes (#2166) #2182 entry could mention that IP address objects are also accepted.
• Verified: the refactor(pcapng)!: rename PCAPNG.context to section_context (#2172) #2205 entry (context is in v1.0.0 and v1.4.1), the fix(foundation): validate a registrar's code before registering it (#2186) #2193 registrar list, the fix(pcapng): correct flags-option length messages and IPv6 maker defaults (#2168) #2179–fix(corekit): export SchemaField, SwitchField, PortEnumField and nested_packet_context (#2174) #2191 entries, no omitted user-visible round-5 merge, and changelog_md.py --check in step.

[JarryShaw]

Review: GOOD TO GO at 0c8ec5772 (Sonnet delta review of c27208ece→0c8ec5772).

• Systemd journal entries (pcapng: journal entry split on b'\n\n' shreds binary data that contains it #723, Every systemd journal field after the first binary field is discarded: entry_data.read() reads to EOF where it means to skip one newline #704, pcapng: a journal entry without a trailing newline returns the block padding as field data #794, Every EOF-truncated PCAP-NG file raises an uncaught ValueError: pcapng_block_selector passes a negative __length__ to SchemaField #678), measured with the worker's probe. Each of these is kept as captured with a ProtocolWarning:
  • a missing newline with padding;
  • a binary field without its terminating newline;
  • a bad entry before a good one;
  • an over-long length.
    b'A=1\n\n' parses as [A=1, <empty entry>]. The PR changes no code, so measuring against main's pcapkit measures this head too.
• "Without the fix" at three names used in string annotations are never imported, one of them raising NameError under get_type_hints; and MANIFEST.in now contradicts itself about what an sdist ships #642, fix(toolkit): key pyshark link-type resolution on frame.encap_type #850, frame.py:199 seeds packet['bytesorder'] where readers expect 'byteorder', and no big-endian PCAP fixture exists to catch it #605 and Every systemd journal field after the first binary field is discarded: entry_data.read() reads to EOF where it means to skip one newline #704, with the failure strings kept.
• Behaviour: the fix(const,vendor): bring 8 bespoke registries onto EnumRegistry, stop them minting #869 OptionType sentence is stated as behaviour, and the fix(arp): reject bool and non-bytes addresses for non-IP ptypes (#2166) #2182 wording on ipaddress objects matches _make_proto_resolve.
• Generated file: changelog_md.py --check is in step, and no "unfixed" or "later review" remains.

[JarryShaw]

Ready to merge at 0c8ec5772. Part of #1103 (round 6).

• Review: Sonnet GOOD TO GO after one NEEDS CHANGES round.
• CI: all checks complete, Required checks passed SUCCESS, 0 failed.
• KTLO: 6 legs, no warnings.
• Ordering: merged into main and run. project leg: 2298 tests, 0 failures.
• Conflicts: none against the current main.