Review WaitLens in five minutes
WaitLens is a local host tool supporting QNX 8.0 development. The working host path is available without a QNX licence. Native build/run validation is pending and is not established by the included fixtures or test doubles.
1. Run without installation
Requirements: Python 3.10+, a desktop browser and the extracted source directory.
The current verification was on Windows / CPython 3.12.5. Linux and macOS have not
been checked in this run. No login, service, upload, pip download or API key is needed.
On Windows, py -3 can replace python if that is your Python launcher.
In the directory containing pyproject.toml:
python -m waitlens compare fixtures/send.txt fixtures/reply.txt --provenance synthetic --out build/judge-comparison
python -m waitlens analyze fixtures/cycle.txt --provenance synthetic --out build/judge-cycle
Double-click build/judge-comparison.html, or use your browser's Open File command.
The JSON files beside the HTML contain the same observations. The HTML requires no
server and makes no external requests. External reference links open only on click.
An optional python -m pip install . needs local packaging tools; source execution
above is the reproducible route and avoids package downloads entirely.
2. Check these exact results
| View | Expected observation | Interpretation limit |
|---|---|---|
| Comparison banner | One SEND → REPLY change | Synthetic demonstration, not a target capture |
| Observed changes | PID:TID 101:1; SEND / 100 → REPLY / 100; line 3 in both inputs | Stable identity is assumed; PID reuse is possible on real systems |
| All parsed threads | Server PID 100 and client PID 101 present in each sample | No specific destination worker is inferred |
| Cycle report | Candidate component 200, 300 | Process cycle alone does not prove deadlock |
| Cycle report | PID 400 RECEIVE 200 contributes no process edge | 200 is a channel ID in that row |
Use Before snapshot, After snapshot and All parsed threads to inspect the source rows. Read Read the evidence for state meanings and next debugging actions.
3. Run the local checks
python -m unittest discover -s tests -v
Expected on the checked source: 30 tests, zero failures. Native-pipeline tests use fake tools and temporary synthetic files; no real target is contacted. Browser media tooling is optional and is not imported by these tests or the application.
Invalid input returns exit code 2 and a named input/line diagnostic. A successful report returns 0; a cycle candidate is not a confirmed failure and does not change that exit code. Supported input is one UTF-8 non-verbose snapshot, at most 16 MiB; verbose output, duplicate rows and concatenated captures are rejected.
4. Inspect the QNX-specific implementation
waitlens/core.py: state parsing, SEND/REPLY process edges, RECEIVE exclusion, coverage warnings and iterative cycle analysis.waitlens/report.py: observation banner, source lines, before/after views, glossary, escaped input and interpretation limits.docs/HOST_TUTORIAL.md: complete host practice and optional capture format.target/waitlens_probe.c: prepared native QNX 8.0 experiment, not compiled or runtime-validated in this submission preparation.scripts/post_sdp.py: optional owner-authorized native pipeline; requires a licensed, already prepared environment and private development target.
Target specification for that optional path: QNX OS 8.0 x86-64 VM; intended QNX SDP
8.0 qcc -Vgcc_ntox86_64. Installed update, VM version and real behavior are not yet
verified. Refer to docs/human-steps.md only if you choose to set up that path.
No proprietary QNX compiler, headers, executable, image or licence is distributed.
Access and reproducibility
The owner must make the source, media and tutorial publicly accessible free of charge through the end of judging: January 29, 2027 17:00 EST / January 30 06:00 SGT. The host path requires no judge credentials or special hardware. The public repository and YouTube links are owner publication steps, not local test results.