fix(ipv6)!: a CALIPSO option loses its trailing pad and shrinks its length on rebuild (#1403)

[JarryShaw]

• Searched for similar pull requests
• Followed the coding style (make pylint, make mypy, make isort) — ran mypy on the 6 source files and isort --check on the new test; pylint not run
• make test passes, and a test case covers the change — ran the modules listed below only
• Added a changelog entry under docs/source/changelog/ and regenerated CHANGELOG.md, if the change is user-visible — N/A, added centrally

What is the purpose of your pull request?

• fix — corrects a defect

---

Description of your pull request and other information

Closes #1403

The CALIPSO pad (schema/internet/{hopopt,ipv6_opts}.py:473) was a zero-packing PaddingField with no data-class field, so make dropped it and wrote Opt Data Len as 8 + bitmap. It is now a BytesField, carried as pad in the data class, written back by _make_opt_calipso and counted in the length. Same pattern as the TS remainder in #1399.

Breaking: data.CALIPSOOption (HOPOPT and IPv6-Opts) gains a required pad: bytes field; test_ipv6_extension_unit.py builds one directly and now passes pad=b''.

Probe HOPOPT.from_data(HOPOPT(raw).info) on 3b01 070c 00000001 0005 0000 00000000: before 3b01 0708 00000001 0005 0000 01020000 (length 8, PadN appended), after byte-exact. Same for non-zero pad and cmpt_len 2, both classes.

Tests: new tests/protocols/internet/test_ipv6_calipso_pad_unit.py (zero/non-zero pad, cmpt_len 0/2, make()). With the six sources reset to origin/main it fails 14 subtests (the 2 no-pad controls pass); with the fix all pass. run_unittest_leg.py protocols/internet 554 OK; test_option_roundtrip_unit.py 6 passed; tests/project 399 passed, 1 skipped.

[github-actions]

Coverage: 89.43% (unit tier, Python 3.14, b54e4daa7, 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	2018	88	628	24	94.78%
pcapkit/dumpkit	234	2	84	2	98.74%
pcapkit/foundation	2637	98	928	48	94.78%
pcapkit/interface	112	7	40	5	92.11%
pcapkit/protocols	16329	205	4324	167	98.16%
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.

[JarryShaw]

Cross-review verdict on b54e4daa7: GOOD TO GO (Sonnet reviewed; author Opus).

• Fuzz: 6,000 random CALIPSO headers (cmpt_len 0–10, every Opt Data Len, random pad and bitmap) across HOPOPT and IPv6-Opts. 1,574 cases that were wrong on main are now byte-exact. None regresses, and the 3,506 parse rejects are identical on both trees.
• breaking is justified: CALIPSOOption requires pad. The only other constructor is the test the PR already updates.
• Checksum: make() has never computed the CRC (RFC 5570 §5.1.7), on main either. A caller-supplied checksum survives the round trip. This gap predates fix(ipv6): a CALIPSO option loses its trailing pad and shrinks its length on rebuild #1403 and is out of its scope.
• No overlap with fix(protocols): parser registrars warn only when the entry really changes (#1362) #1400: the hunks in hopopt.py and ipv6_opts.py don't touch.
• Tests: reverting the fix fails 14 subtests. Fixtures are unchanged. I re-ran protocols/internet myself at this head: 554/0.

Follow-up, not blocking: make() writes a corrupt length once the pad reaches 248 octets, and a non-bytes pad surfaces a raw struct.error. Filed as #1408.