Initial MVP consolidated verification

Outcome: all required gates passed after the documented publication retest.

The software execution is 0.1.0-candidate-4 execution. Its product and test inputs remained stable throughout all executable suites. The first documentation attempt failed on an RST heading underline in the operating notes; the failed log is retained there. The first auxiliary retest found another short underline in this record’s draft heading. Both publication formatting failures and the subsequent passing checks remain available below.

The stable execution recorded 54 passed, 0 failed, 0 errors, 0 skipped across pytest and Node. The supplemental visual publication case is recorded separately, without inflating that stable-suite total. Initial SYS/SWR criteria are marked passed after the criterion review in Implementation criterion review and operating limits; four later occurrence-mode SWRs remain not_assessed. Stakeholder verification, release approval and project licensing are not inferred from these results.

Candidate identity

Git input: 30aaba98776ecbe648558d278a985691d81bccb4 plus the retained dirty source/test identity 9867e11555574fedc0dd1ec7f67276e1e8b17e703237c476b180ba57e6e2d85a. Wheel SHA-256: 8a31577baec118e32d697c00faef83f06b60dd9478cb0f791be64a0c867f1fdd. The tested wheel is retained in the candidate’s artifacts directory. Package contents were compared byte-for-byte with the final product sources. Accepted SYS/SWR blocks were compared with the input Git baseline; only their independent verification metadata changed. No product or executable test changed during publication retesting.

Measurements

Real host fixture: 30,000 Needs, 90,000 authored edges; build 331.95 s. Graph script 34,897,944 bytes; detail scripts 147,393,890 bytes. Three browser repetitions per deployment mode, including 200-node scenes:

Observed timings on the recorded Windows machine

Mode

Startup median / maximum

Expansion median / maximum

file

628.4 / 631.5 ms

85.0 / 378.0 ms

http

665.3 / 701.1 ms

72.5 / 222.0 ms

Raw results include collapse, first/cached detail timings and process-tree working-set measurements. File fixture copying/cleanup is preparation time, outside the browser timing intervals. Memory can double-count shared pages. No numeric latency/memory budget has been agreed; the 200-node limit is an independently passing acceptance criterion. Windows file paths must remain short enough for Firefox’s local loader; see Implementation criterion review and operating limits.

Publication attempts and evidence

  • First auxiliary retest (failed draft heading) (retained in the engineering checkout)

  • Second auxiliary retest (passed) (retained in the engineering checkout)

  • Exact publication commands, UTC times and outcomes (retained in the engineering checkout)

  • Retained consolidation/check script (retained in the engineering checkout)

  • Visual publication JUnit (retained in the engineering checkout)

  • coverage-final: exit 0; log (retained in the engineering checkout)

  • documentation-final: exit 0; log (retained in the engineering checkout)

  • visual-publication: exit 0; log (retained in the engineering checkout)

  • traceability-final: exit 0; log (retained in the engineering checkout)

  • diagram-publication-final: exit 0; log (retained in the engineering checkout)

Intermediate attempts

The earlier attempts retain their source-stability failures and passing case outcomes. They do not verify the final frozen input: