Runs

All runs on 2026-10-08 in the hunt/forage-2026-10-08 worktree, Python 3.13 from the repository's .venv (python-flint 0.9.0, rigor.BACKEND = python-flint), one core, no CI, no Lean build.

Estimate written before the probe

The parent hunt records one 128-bit disk enclosure at 0.0093 s and one 320-node continuation step under a millisecond. Timed here before the ladder: odd_ball.rouche at 217/200, 128 bits: 0.041 s; one mpmath quadrature of H_t at 30 digits: 0.17 s, at 50 digits: 0.31 s; one float grid evaluation: 0.3 ms. Budget: route A is six quadratures per Newton step times about eight steps at each of two precisions, about 10 s; route B is a few thousand grid evaluations, under 5 s; the ladder is five rungs of one mpmath centre polish (about 2 s) plus four disks, about 12 s; the second pair and lesions, about 6 s. Expected total under one minute, no checkpointing needed.

2026-10-08, prototypes in the scratchpad (not committed)

Two throwaway scripts validated the routes before the probe was written. Both landing-time routes agreed at once (t_c = 1.0876360002296 float, 1.08763600022958693221796... at 30 and 50 digits). Two instrument refusals happened and are worth recording: the quartic tracker control first enclosed all four roots (winding 4, with Y > a no circle separates the pair from the real roots), fixed by choosing a = 1.2, Y = 0.5; and a fixed contour for the second pair at t = 0.7 returned winding 1 because one real zero had already left it after the landing, fixed by bracketing from the last off-axis time with a 0.4 radius. Neither refusal produced a number that was used.

id: dh_minus_landing-2026-10-08-prototype
hunt: dh_minus_landing
started: 2026-10-08T00:00Z
finished: 2026-10-08T00:30Z
ran:
  - .venv/bin/python <scratchpad>/proto.py (route B float discriminant, route A mpmath double zero)
  - .venv/bin/python <scratchpad>/proto2.py (second pair continuation, polynomial controls, disk ladder, lesions)
artifacts:
outcome: both landing-time routes agree on t_c near 1.0876360002296 and five rational disks below it decide; two contour refusals (winding 4, winding 1) were corrected before any number was recorded

2026-10-08, probe.py, first full run (one stage null)

37.1 s wall. Every stage returned except the shared-layer kernel lesion, which looked for the perturbed pair at t = 1.08 from the true pair's position and found a real zero there: the 1% fault moves the landing time below 1.08, so the lesion stage returned pair_found: false, a null for that stage. No other number changed between this run and the next.

id: dh_minus_landing-2026-10-08-probe1
hunt: dh_minus_landing
started: 2026-10-08T01:00Z
finished: 2026-10-08T01:01Z
ran:
  - .venv/bin/python hunts/dh_minus_landing/probe.py
artifacts:
  - hunts/dh_minus_landing/results.json (overwritten by probe2)
outcome: t_c measured by both routes to 4.7e-14 agreement, ladder decided to 10876359/10^7, second pair lands at 0.63095; the shared-layer lesion stage returned null because it started below the perturbed landing time

2026-10-08, probe.py, second full run (the recorded one)

The lesion stage now tracks the perturbed pair from t = 0, and the second pair gains the 960-node resolution check. 33.1 s wall: controls 3.3 s, route B 7.7 s, route A 10.4 s, ladder 10.9 s, second pair and lesions 4.8 s. Measured against the estimate: within budget throughout.

id: dh_minus_landing-2026-10-08-probe2
hunt: dh_minus_landing
started: 2026-10-08T01:10Z
finished: 2026-10-08T01:11Z
ran:
  - .venv/bin/python hunts/dh_minus_landing/probe.py
artifacts:
  - hunts/dh_minus_landing/results.json
outcome: t_c = 1.087636000229586932217964154453984835461141904 (route A, 50 digits; route B within 4.7e-14), strict lower endpoint 10876359/10^7 decided on four Arb configurations, door worth 2.636e-3 for this pair and shut to 1.0e-7; the 1% kernel fault moves t_c by 0.067 while both routes agree to 1.4e-13 and only the Hurwitz identity (defect 1.2e-3) catches it; no kill condition fired

2026-10-08, tests

id: dh_minus_landing-2026-10-08-tests
hunt: dh_minus_landing
started: 2026-10-08T01:20Z
finished: 2026-10-08T01:25Z
ran:
  - .venv/bin/python -m pytest -q -n0 hunts/dh_minus_landing
  - .venv/bin/python -m pytest -q -n0 hunts/dh_minus_landing tests/test_docs_numbering.py tests/test_hunt_probe_discipline.py tests/test_doors.py tests/test_hunt_doors.py tests/test_huntspec.py tests/test_hunt_numbering.py
  - .venv/bin/python scripts/make_context.py --check
artifacts:
  - hunts/dh_minus_landing/test_dh_minus_landing.py
outcome: hunt subset 15 passed in 0.65 s; governance subset 71 passed and 3 xfailed (the three pre-existing KNOWN_INCOMPLETE doors sections) in 142.24 s; CONTEXT.md up to date; no em dash, reserved word, whitespace or secret finding