[JarryShaw]

• Searched for similar pull requests

• Followed the coding style — isort/pylint/mypy run on the 3 source files; every pylint/mypy message is on a line this PR does not touch

• make test passes, and a test case covers the change — new module (9 passed), IPv6/HOPOPT modules, run_unittest_leg.py protocols/internet (589 OK), tests/project (404 passed)

• Changelog entry — N/A — centralised in docs(changelog): shared 1.5.0 changelog — long-lived, merges last (#610, #616, #617, #618, #620) #657

• fix — corrects a defect

---

Closes #1434

A jumbogram (Payload Length 0, Hop-by-Hop Jumbo Payload option) read no payload. The schema now splits payload/trailer at the Jumbo Payload Length, so the payload is dissected and only octets past it are trailer. The option counts only in a Hop-by-Hop header right after the IPv6 header (RFC 8200 §4.1). Beside a non-zero Payload Length (an RFC 2675 error) the Payload Length still governs, as captured, with a ProtocolWarning. make() writes Payload Length 0 when the payload exceeds 65,535 octets. No info key is added or removed; raw_len now counts from the Jumbo length.

case (bytes)	main next / trailer	this PR next / trailer
jumbo = 32, exact (72; 86 in Ethernet)	NoPayload / 32	UDP / —
jumbo = 32, +6 pad (78)	NoPayload / 38	UDP / 6
jumbo = 16 < captured (72)	NoPayload / 32	UDP / 16
jumbo = 100 > captured (72)	NoPayload / 32	UDP / — (raw_len 92)
PL 32 + Jumbo option (72)	UDP / —	UDP / —, warns
Jumbo in non-leading HBH (80)	NoPayload / 40	unchanged

Every case rebuilds byte-exact before and after. 7 of the 9 new tests fail without the fix; the other 2 pin unchanged cases. Per-frame info digests of the 1616 frames in examples/captures/ match main.

[JarryShaw]

Verdict: GOOD TO GO at 2465d9642 (Sonnet cross-review; the author ran on Opus).

• Round-trip: data == raw and Ethernet.from_data(info) are byte-exact across every probe. The probes covered:
  • Jumbo exact, shorter and longer than captured, plus Jumbo 0, 5 and 70000;
  • Jumbo with a non-zero Payload Length;
  • Jumbo in a second Hop-by-Hop header;
  • Pad1/PadN, option length 5, and truncated or over-long Hop-by-Hop headers;
  • all of these bare and inside Ethernet.
• raw_len: past capture it reports the Jumbo extent. No caller under pcapkit/ slices by IPv6 raw_len.
• Re-run myself: the new module gives 9 passed, and with the three source files from main it gives 7 failed, 2 passed. Jumbo 0, 5 and 8 parse as NoPayload with raw_len 0, so nothing goes negative.
• Merged into 16c87f901: conflict-free, and the protocols/internet leg runs 592 OK. On the PR alone, tests/project gives 404 passed, and the per-frame info digests over examples/captures/ match the base.

Nits, no change needed:

• jumbo_payload_length() accepts a Jumbo value of 65535 or less, which RFC 2675 calls an error. It uses the value silently.
• make() with a payload over 65535 writes Payload Length 0 but inserts no Jumbo option. The bytes still round-trip, and the old code behaved the same.

[github-actions]

Coverage: 89.49% (unit tier, Python 3.14, 2465d9642, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	4	93.55%
pcapkit/const	18791	968	2352	869	90.90%
pcapkit/corekit	2043	89	636	25	94.77%
pcapkit/dumpkit	234	2	84	2	98.74%
pcapkit/foundation	2648	98	930	48	94.80%
pcapkit/interface	112	7	40	5	92.11%
pcapkit/protocols	16711	219	4510	183	98.07%
pcapkit/toolkit	539	74	168	3	84.58%
pcapkit/utilities	429	4	122	4	98.55%
pcapkit/vendor	4409	2342	1006	157	43.25%

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