[JarryShaw]

• Searched for similar pull requests
• Followed the coding style (make pylint, make mypy, make isort)
• make test passes, and a test case covers the change
• Added a 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 changes

---

Description of your pull request and other information

The engine tables in docs/source/index.rst were out of date now that 3.15 is released.

• Support table: Engine Support by Python Version still read "3.15 remains inferred" and had two footnotes, one saying no 3.15 interpreter was available and one saying tshark was missing. The 3.15 column and the pyshark row are now measured, and they agree with main's Unit Tests run 40010064464 engine jobs. 3rdparty.rst's "pyshark on 3.13 is unchecked" now says it works.
• Speed table: this was a 2026-09-17 hand run on an M2 laptop with no 3.15 column. It is now the Docker suite's output (examples/benchmark/run.sh, 3.10–3.15, exit 0), pasted unedited. The numbers come from an x86_64 Linux host, so they are not comparable with the old figures.
• Suite fixes:
  • The released python:3.15-slim-bookworm image (3.15.1) is pinned by digest as a default row.
  • typing-extensions is now installed for Python < 3.11, since the 3.10 image failed with ModuleNotFoundError.
  • report.py underlines headings at index.rst's level.
• Tests: test_docs_epilog_links_1936_unit and test_docs_pcap_module_link_1948_unit get narrow exemptions for the generated snippets. Each failed 1 case before the exemption, and both pass now.

Noticed, not fixed: pcap_ct takes 0.27 ms per packet on 3.15 against about 0.117 on 3.12–3.14, consistently across three passes. Filed separately.

[JarryShaw]

NEEDS CHANGES at e94d74ff2 (Sonnet review). One fix is needed; it is being made on this branch.

• Required: the Engine Comparison prose in docs/source/index.rst still says the default engine "takes about 0.2 milliseconds per packet". The table a few lines below now measures 0.4879 ms, with passes ranging 0.42–0.52.
• Passes:
  • Support table: matches main's Unit Tests run 40010064464 cell for cell on 3.10–3.15 (Scapy and PCAP_CT supported throughout, PyShark unsupported from 3.14, PyPCAP not installable and PyPCAPFile unsupported from 3.12).
  • Speed tables: unedited report.py output.
  • Test exemptions: narrow; only the generated rows are exempted.
  • Harness: test_harness.py gives 116 passed, and the typing-extensions marker matches pyproject.toml.
• UNVERIFIED: the image digest against the registry, and the pcapkit and dpkt rows, which have no Engines job.

[JarryShaw]

GOOD TO GO at 970796ee6. The only change since the review's e94d74ff2 is the two-line fix it asked for (git diff e94d74ff2 970796ee6: docs/source/index.rst, 2+/2−).

• Old: "takes about 0.2 milliseconds per packet, slower than most of the third-party engines below."
• New: "is slower than every third-party engine below except PyShark, as Test Results measures."
• Checked against the table: pcapkit runs 0.43–0.52 ms per packet. That is behind pypcapfile, dpkt, pypcap, pcap_ct (0.27 even on 3.15) and scapy, and ahead of pyshark only (~31 ms).
• Tests: the docs-link tests (1936, 1948) and test_conventions_doc_claims pass. Every other point of the review stands.

[github-actions]

Coverage: 91.54% (unit tier, Python 3.15, 970796ee6, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	2	95.16%
pcapkit/const	19599	880	2618	866	92.12%
pcapkit/corekit	2565	71	812	37	96.56%
pcapkit/dumpkit	282	3	102	3	98.44%
pcapkit/foundation	2930	23	1022	25	98.79%
pcapkit/interface	138	4	52	3	96.32%
pcapkit/protocols	17414	219	4574	185	98.14%
pcapkit/toolkit	869	36	298	10	95.20%
pcapkit/utilities	473	7	136	5	98.03%
pcapkit/vendor	4432	1988	1018	167	51.87%

Per-file detail: the coverage-html artifact of this run.

[JarryShaw]

Ready to merge at 970796ee6.

• Review: GOOD TO GO, after the stale "0.2 ms" sentence was fixed.
• CI: every blocking check passed (Required checks passed SUCCESS). The 3.8–3.10 KTLO legs carry no warnings.
• Ordering: the tests/project plain-unittest ordering leg passed on a scratch merge into main: 2180 tests, 0 failures.

If #2112 merges first, the 3.15 pcap_ct cell (0.2702) goes stale. I'll re-measure it in a follow-up.