[JarryShaw]

• Searched for similar pull requests
• Followed the coding style (make pylint, make mypy, make isort) — Makefile flags on traceflow/tcp.py: pylint 10.00/10, mypy clean, isort clean
• make test passes, and a test case covers the change — modules listed below
• 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 #1360

A new flow whose label an earlier flow already used now gets the next free suffix -1, -2, ... (TCP._unique_label), so it opens its own file instead of truncating the earlier one. Non-colliding labels are unchanged. The bidirectional path collided too (a new SYN on torn-down endpoints at the same timestamp), and is fixed by the same change. Visible output change: a colliding flow's label and fpout gain the suffix.

Probe (frames at t=300.0), before → after:

• unidir FIN, ACK: two flows …-300.0, one file holding Frame 2 → …-300.0 (Frame 1) + …-300.0-1 (Frame 2)
• bidir SYN, FIN, FIN(rev), SYN: two flows …-300.0, one file holding Frame 4 → …-300.0 (1-3) + …-300.0-1 (4)

Tests: new tests/foundation/traceflow/test_traceflow_label_unique_unit.py 4 passed; with tcp.py from origin/main, 3 failed, 1 passed (the unchanged-label case). test_tcp 15 passed, test_traceflow_base 5, data/test_models 1, integration/test_traceflow_end_to_end 7; tests/project 389 passed, 1 skipped. util/run_unittest_leg.py foundation (under a 200 MB file-size cap, per #1372): 487 tests, 2 failures, 2 errors, all pre-existing on main — the DPKT and Scapy test_pcap_output_opens_with_the_global_header (Errno 27, #1372) and both subtests of test_runtime_engines.py::PCAPEngineTests::test_pcap_run_writes_global_header_for_file_and_stream_modes.

[JarryShaw]

Cross-review verdict on 0520b0646: GOOD TO GO (Sonnet reviewed; author Opus).

• Repro: on main, the unidirectional case (FIN then ACK at t=300.0) leaves one file holding only frame 2. The bidirectional case (connection reuse at t=300.0) leaves one file holding only frame 4, so frames 1–3 are lost. On this PR each flow gets its own file (…-300.0, …-300.0-1) and nothing is truncated.
• No suffix collisions: _unique_label checks against every label already handed out, suffixed ones included. Five unidirectional collisions in a row gave the bare label, -1, -2, -3 and -4. Four bidirectional reuses gave the bare label and -1 to -3, each with the right frames. A collision with an already-written flow left that file intact.
• Engines and fixtures: default, dpkt and scapy produce identical labels over all 23 captures in json and pcap, with bidirectional on and off: 276 configurations and 2,886 flows. Labels, indexes, fpout names, file lists and sizes are byte-identical to main in every configuration. No fixture has a colliding label.
• Consumers: nothing in pcapkit/ parses a label (interface/misc.py:223 passes fpout straight through), and no test splits one on -.
• Tests: the revert proof fails 3 of 4. tests/foundation/traceflow gives 25 passed, test_traceflow_end_to_end 7, and tests/project 389.

Nit: _labels is created lazily via __dict__.setdefault rather than in an __init__.

[github-actions]

Coverage: 88.98% (unit tier, Python 3.14, 0520b0646, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	4	93.55%
pcapkit/const	18777	1032	2348	843	90.27%
pcapkit/corekit	1949	84	606	20	94.91%
pcapkit/dumpkit	189	1	68	1	99.22%
pcapkit/foundation	2452	102	854	39	94.40%
pcapkit/interface	112	7	40	5	92.11%
pcapkit/protocols	16013	206	4178	177	98.06%
pcapkit/toolkit	539	74	168	3	84.58%
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.