[JarryShaw]

• Searched for similar pull requests
• Followed the coding style (make pylint, make mypy, make isort)
• make test passes, and a test case covers the change
• Added a changelog entry — N/A — centralised in docs(changelog): shared 1.5.0 changelog — long-lived, merges last (#610, #616, #617, #618, #620) #657

What is the purpose of your pull request?

• fix — corrects a defect

---

Description of your pull request and other information

Closes #2129.

• Cause: in the pcap_ct environment, import pcap raises OSError on macOS (dlopen(libc.so.6 …)) and on Windows (Cannot find npcap/wpcap.dll). pcap_ct already reports that through its own unsupported_reason(), but pypcap does not. So extract(engine='pypcap') raises, and _unavailable re-raised the error because the message lacks libpcap. The whole environment exited.
• Fix: a new _loader_failure(engine) in preflight, for the engines backed by _pcap_backend.DISTRIBUTIONS only, asks _pcap_backend.probe(). If the module is installed but fails to import, the engine becomes an explained row carrying the real OSError. The 'libpcap' in str(exc) wording check is gone, so an OSError reading the capture is always a real failure.
• README: one Troubleshooting sentence. This PR does not touch the parts ci(benchmark): run the engine benchmark on Actions and post the report to Discussions (#2125) #2128 owns.
• Tests: TestLoaderFailure in examples/benchmark/test_harness.py fakes each runner's loader error. All 6 tests fail against main's benchmark.py, and the file gives 122 passed with the fix.
• Not fixed here: the library itself has the same gap. PyPCAP._reason_for does not report an unusable pcap module the way PCAP_CT._reason_for does. That is filed separately.

[JarryShaw]

GOOD TO GO at eff7ff6c7 (Sonnet review).

• Root cause: confirmed from the run's logs. preflight('pypcap') died in _extract with dlopen(libc.so.6 …) on macOS and with Cannot find npcap/wpcap.dll on Windows. Neither message contains libpcap.
• _loader_failure: it probes only the _pcap_backend.DISTRIBUTIONS engines, with the same "installed but unusable" condition as PCAP_CT._reason_for. _purge leaves no pcap module behind, so pypcap → pcap_ct → pypcap all return the same reason.
• Linux: no regression. A fake pcap raising Cannot find libpcap.so library still becomes an explained row for both engines.
• Tests: all 6 new tests fail on main's benchmark.py, and the harness gives 122 passed.
• Library: the reviewer reproduced the gap. extract(engine='pypcap') raises the OSError where engine='pcap_ct' warns and falls back. I'm filing it separately.

[github-actions]

Coverage: 91.55% (unit tier, Python 3.15, eff7ff6c7, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	2	95.16%
pcapkit/const	19599	880	2618	866	92.12%
pcapkit/corekit	2589	75	816	37	96.48%
pcapkit/dumpkit	282	3	102	3	98.44%
pcapkit/foundation	2969	23	1036	25	98.80%
pcapkit/interface	138	4	52	3	96.32%
pcapkit/protocols	17418	219	4574	185	98.14%
pcapkit/toolkit	869	36	298	10	95.20%
pcapkit/utilities	498	7	144	5	98.13%
pcapkit/vendor	4432	1988	1018	167	51.87%

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

[JarryShaw]

Ready to merge at eff7ff6c7.

• Review: the Sonnet review is GOOD TO GO.
• CI: every blocking check passed (Required checks passed SUCCESS), and the 3.8–3.10 KTLO legs carry no warnings.
• Base: the head sits directly on main 8fdc2fb33.
• Ordering: no plain-unittest ordering leg applies, because the change is confined to examples/benchmark/. examples/benchmark/test_harness.py gives 122 passed.

This closes #2129.