fix(tcp): Quick-Start Response reader reads a stray octet; maker fails below 40 kbps

[JarryShaw]

• Searched for similar pull requests
• Followed the coding style (make pylint, make mypy, make isort). Ran with the Makefile flags on tcp.py only: no new findings. pylint's W0612 unused-variable 'size' is gone.
• make test passes, and a test case covers the change. Not the full suite; the modules run are listed below.
• Added a changelog entry under docs/source/changelog/ and regenerated CHANGELOG.md, if the change is user-visible — N/A, added centrally after the wave

What is the purpose of your pull request?

• fix — corrects a defect
• feat — adds a feature
• perf — changes performance, not behaviour
• refactor — changes neither behaviour nor performance
• test — tests only
• docs — documentation only
• ci — workflows or build tooling
• chore — anything else

---

Description of your pull request and other information

Closes #1119

• Reader. _read_mode_qs called self._read_unpack(1) past the option before its length check. That line is removed.
• Maker. _make_mode_qs produced a negative rate code below 40 kbps. It now clamps to N = 0. RFC 4782 §3.1 defines the rate as 40 kbps * 2**N, with N=0 meaning a rate of zero. §4.3 allows a response to report a lower rate than requested, so rounding down is valid and nothing is refused.
• Removes the tcp-option/Quick_Start_Response EXPECTED_FAILURES entry. make_samples now captures the option (257 cases, 7 left out).

Probe (rates 0 / 80 / 640 + payload / 20 / 60 kbps):

• before: StructError: unpack: empty buffer ×4; at 20 kbps ValueError … b'0000-001'
• after: 0, 80.0, 640.0, 0, 0, each with ttl_diff=3, nonce=0x1234

Tests: new tests/protocols/transport/test_tcp_quick_start_response_unit.py covers wire parsing and round trips, with and without payload, and rate flooring. Without the fix it gives 18 failed. With only the reader fixed, the 3 sub-40 kbps subtests fail. Each module was run in its own process:

• new module: 3 passed, 18 subtests passed
• test_option_roundtrip_unit: 6 passed, 360 subtests passed
• test_tcp_udp_unit: 18 passed; test_tcp_runtime: 5 passed
• tests/project: 379 passed, 1 skipped

[JarryShaw]

Cross-review verdict on fe0471fff: GOOD TO GO (ran on Sonnet; author Opus)

• Reader: the stray _read_unpack(1) is removed. On main, make_samples.py leaves tcp-option/Quick_Start_Response out with StructError: unpack: empty buffer. With this PR it parses.
• Maker: rates under 80 kbps now clamp to N=0. This fits RFC 4782 §3.1 and §4.3: N=0 means zero, and a response may report a lower rate. Probed across 20–1,310,720 kbps; each rate packs to the expected code and parses back.
• Tests:
  • with tcp.py reverted, the new module fails 18;
  • with only the clamp reverted, 3 subtests fail;
  • test_option_roundtrip_unit passes 360.
• Known limits, not regressions:
  • Rates of 2.6 Gbps or more raise FieldValueError, because N=16 does not fit in 4 bits.
  • req_rate comes back as a float, and main does the same.

Merge order: this PR and #1131 both delete neighbouring EXPECTED_FAILURES entries in test_option_roundtrip_unit.py, so the two conflict textually there. Whichever merges second needs a trivial rebase. The merge with #1114 in tcp.py is clean.

[github-actions]

Coverage: 88.75% (unit tier, Python 3.14, 6011af7d7, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	4	93.55%
pcapkit/const	18757	1035	2342	846	90.23%
pcapkit/corekit	1873	91	578	22	94.33%
pcapkit/dumpkit	136	0	40	0	100.00%
pcapkit/foundation	2425	145	842	34	92.56%
pcapkit/interface	112	7	40	5	92.11%
pcapkit/protocols	15658	188	3944	163	98.18%
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.

[JarryShaw]

Cross-review verdict on 6011af7d7: GOOD TO GO (ran on Sonnet; author Opus, round 2)

This round reviewed the rebase onto main at dbb30c4a3, after #1131 merged.

• The tcp.py change is the same as in round 1: the stray _read_unpack(1) is removed, and the rate is clamped with max(..., 0).
• EXPECTED_FAILURES has no tcp-option/ entries left; the import was checked against the worktree.
• Nothing from fix(tcp): User Timeout maker writes length=3; RFC 5482 requires 4 #1131 was lost: _make_mode_timeout still writes length=4, and its test passes.
• test_tcp_quick_start_response_unit passes 3 and test_option_roundtrip_unit passes 360 subtests.

The round-1 findings still hold.