[JarryShaw]

• Searched for similar pull requests
• Followed the coding style (pylint/mypy/isort run on the touched pcapkit/ files with the Makefile flags; no new findings)
• make test passes, and a test case covers the change (per-module only, see below; not the full suite)
• 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 #1175. Breaking: public Enum_Type values swap to Data = 0, Control = 1 (RFC 2661 §3.1). The const is regenerated from the fixed vendor table.

• make() sets L on a control message and defaults total_length to header + payload. A new length_flag keyword, which _make_data fills in, keeps a control message captured without L byte-exact. A total_length the caller passes is written as given (documented, not checked), like UDP after design(protocols): should from_data keep declared lengths on truncated captures? #1155, so a rebuild of a snaplen-truncated capture keeps the declared length.
• The offset pad is recorded as info.padding and goes back through make().

probe	before	after
int(Type.Data), int(Type.Control)	1 0	0 1
L2TPv2(tunnel_id=1, session_id=2)	800200010002 (T=1)	000200010002
make(type=Control, ns=0, nr=1, …)	0802… (L=0)	c802000c…
pad rebuild …0003010203ff03	…0003000000ff03	byte-exact

Tests: new tests/protocols/link/test_l2tpv2_type_unit.py passes 8. With the source fix reverted, 6 fail. The 2 guards were proved by mutation: dropping length_flag from _make_data fails the guard and 6 cases in test_l2tpv2_length_roundtrip_unit. test_link_unit 24 passed, after one fixture edit giving its dummy flags len=True. tests/project 389 passed, 1 skipped. run_unittest_leg.py protocols --exclude protocols/internet reports 850 tests, 0 failures.

[JarryShaw]

Cross-review verdict on 53f2d0fc1: GOOD TO GO (ran on Sonnet; author Opus). Labelled review: good-to-go. Two design points are held for the maintainer on #1175; neither blocks the merge.

• Enum: RFC 2661 §3.1 confirms Data=0 and Control=1. The vendor generator, re-run, produces a const file byte-identical to the committed one. Nothing else in the codebase uses the numeric values.

• Round trips: byte-exact for:

  • ZLB;
  • control messages with and without S/L;
  • offset pads, which main zero-fills;
  • data messages with sequence numbers;
  • an over-declared length.

  The capture sweep is vacuous because no fixture holds L2TP, so this evidence is synthetic.

• Tests: 6 of the 8 new tests fail without the fix. The ordering leg protocols (rest) ran 850 tests with 0 failures and 0 errors. The test_link_unit change is a legitimate fixture fix: the dummy lacked flags.len.

• breaking: the label is correct, and the breakage is wider than the PR body lists:

  • parsed flags.type flips for the same bytes;
  • the default make() emits T=0;
  • a control message without total_length gains L and a Length field;
  • info and dumpkit output gain a padding key;
  • make() gains length_flag and padding.

  The changelog entry will carry all of these.

• Out of scope: §3.1 also requires S=1 on control messages, and make() does not enforce it.

[github-actions]

Coverage: 88.89% (unit tier, Python 3.14, 53f2d0fc1, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	4	93.55%
pcapkit/const	18757	1034	2342	845	90.24%
pcapkit/corekit	1878	91	580	22	94.34%
pcapkit/dumpkit	142	0	42	0	100.00%
pcapkit/foundation	2442	102	852	39	94.38%
pcapkit/interface	112	7	40	5	92.11%
pcapkit/protocols	15714	188	3980	163	98.19%
pcapkit/toolkit	487	65	144	1	85.42%
pcapkit/utilities	429	4	122	4	98.55%
pcapkit/vendor	4409	2359	1006	158	42.84%

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