Commit graph

Select branches

Hide pull requests

claude/beads-project-intro-fjiez3

main

Mono Color

• d04e585d38
  Coldstart runner: time runs with a clock that stops when the machine does claude/beads-project-intro-fjiez3 Claude 2026-09-09 07:00:30 +00:00
• 3ccbba7284
  Coldstart runner: checkpoint each run, so losing the box costs hours not days Claude 2026-09-06 20:58:05 +00:00
• 58fe56eb55
  bd: record 39.26 progress on homemaker-py-dpt Claude 2026-09-06 18:11:10 +00:00
• 99c85ec4ee
  Remove two dead paths from the objective Claude 2026-09-06 18:11:09 +00:00
• 73324deff1
  bd: save session memories and bring homemaker-py-bk9 current Claude 2026-09-06 16:31:15 +00:00
• 8a1b547285
  bd: save session memories and re-scope bk9 for the post-39.25 objective Claude 2026-09-06 15:48:18 +00:00
• 4931070602
  Outdoor space: enforce the per-level rule, drop the fraction Claude 2026-09-06 15:41:22 +00:00
• f96ec0f316
  bd: close homemaker-py-hxi and homemaker-py-ssz Claude 2026-09-06 15:29:20 +00:00
• 774747e482
  bd: close homemaker-py-hxi; file homemaker-py-dpt (scoring-model sweep) Claude 2026-09-06 15:28:34 +00:00
• a7ef9e2fa0
  Remove ratio_circulation: the score is already a ratio Claude 2026-09-06 15:28:33 +00:00
• f07b3865ef
  No size cap on circulation: twice the corridor is twice as bad, and no worse Claude 2026-09-06 15:01:12 +00:00
• 426203b154
  bd: retitle homemaker-py-hxi and correct its premise Claude 2026-09-06 12:56:40 +00:00
• 9a8852a82b
  A corridor may be corridor-shaped: no aspect cap on circulation Claude 2026-09-06 12:56:39 +00:00
• 32973cfe95
  bd: close homemaker-py-118; file homemaker-py-bk9 (re-baseline) Claude 2026-09-06 07:56:24 +00:00
• d0567d7a74
  Remove the Perl oracle Claude 2026-09-06 07:56:24 +00:00
• f47dfe60af
  Record the ecx A/B: directionally positive, not resolvable at N=6 Claude 2026-09-06 07:38:23 +00:00
• 7047a782e5
  bd: file homemaker-py-118 Claude 2026-09-05 19:25:19 +00:00
• 47079997cf
  The native-vs-Perl parity tests have never run, and were self-comparing Claude 2026-09-05 19:25:18 +00:00
• 109849e438
  A terrace is no longer worth more per m2 than a real internal room Claude 2026-09-05 19:25:02 +00:00
• a76ed3b9b8
  Quality aggregation: divide out how many questions a leaf was asked Claude 2026-09-05 18:57:00 +00:00
• fb70193074
  bd: close homemaker-py-773; file ecx Claude 2026-09-05 17:46:53 +00:00
• f5612ce06c
  Pre-flight: check the ground floor's frontage, not the per-storey average Claude 2026-09-05 17:46:52 +00:00
• 6268748fcb
  bd: file homemaker-py-773; record the 39.16 correction on u5q Claude 2026-09-05 16:04:45 +00:00
• e92a96ac99
  Correct 39.14/39.15: the crinkliness constant is Alexander 159 Claude 2026-09-05 16:04:44 +00:00
• 94d112aa02
  bd: close homemaker-py-9gj and homemaker-py-u5q; file k54 Claude 2026-09-05 15:11:19 +00:00
• 84907694d0
  Record what the crinkliness examination found: 39.13, 39.14, 39.15 Claude 2026-09-05 15:11:11 +00:00
• 970e07235f
  Audit: ask whether a room spec is feasible AS DECLARED, not merely somewhere Claude 2026-09-05 07:46:43 +00:00
• 366a047a60
  Make the crinkliness factor one-sided: stop billing the daylit wall twice Claude 2026-09-05 07:27:02 +00:00
• 6d387bbe3b
  Ignore the A/B harnesses' per-process shard files Claude 2026-09-05 06:29:15 +00:00
• cd392e77c5
  Rescale the underflowing crinkliness tail so the failing region has an ordering Claude 2026-09-05 06:29:02 +00:00
• cc39a38317
  bd: close homemaker-py-ut5 Claude 2026-09-04 17:17:03 +00:00
• 4180a9eb88
  Replace the stale 15-fail acceptance target with the cold-start baseline Claude 2026-09-04 17:16:10 +00:00
• d492b7c271 coldstart maple-court seed 2 @ 500000: 55 fails (11h/44s) Bruno Postle 2026-09-03 10:06:06 +01:00
• 2426760d40 coldstart health-centre seed 2 @ 500000: 5 fails (3h/2s) Bruno Postle 2026-09-03 03:44:11 +01:00
• 761167f6a3 coldstart programme-house seed 2 @ 500000: 1 fails (0h/1s) Bruno Postle 2026-09-02 18:10:06 +01:00
• 51782047f3 coldstart harbor-house seed 2 @ 500000: 42 fails (12h/30s) Bruno Postle 2026-09-02 17:53:41 +01:00
• b885bd3dbd coldstart maple-court seed 1 @ 500000: 73 fails (17h/56s) Bruno Postle 2026-09-02 10:52:55 +01:00
• d7ac57d131 coldstart health-centre seed 1 @ 500000: 9 fails (5h/4s) Bruno Postle 2026-09-01 20:28:37 +01:00
• b39b467662 coldstart maple-court seed 0 @ 500000: 54 fails (18h/36s) Bruno Postle 2026-09-01 03:21:09 +01:00
• 9d99cf34f7 coldstart programme-house seed 1 @ 500000: 1 fails (1h/0s) Bruno Postle 2026-08-31 21:50:30 +01:00
• 2415ad6504 coldstart harbor-house seed 1 @ 500000: 43 fails (7h/36s) Bruno Postle 2026-08-31 16:50:54 +01:00
• d5d99dd4f6 coldstart harbor-house seed 0 @ 500000: 33 fails (8h/25s) Bruno Postle 2026-08-31 09:46:46 +01:00
• 6d1eb26088 coldstart health-centre seed 0 @ 500000: 4 fails (2h/2s) Bruno Postle 2026-08-31 01:20:49 +01:00
• 481e83cdbf
  Shape-curve warm-start cannot pay off as wired; no A/B run Claude 2026-08-30 07:41:16 +00:00
• 4c95505907
  Shape-curve DP models leaf-sharing, so it can fire on real runs Claude 2026-08-29 21:36:45 +00:00
• 104ba12237
  A/Bs now report what their sample could resolve Claude 2026-08-29 19:51:47 +00:00
• c747e55abe
  coldstart runner: commit only its own artefacts, not the whole tree Claude 2026-08-29 19:40:00 +00:00
• f973d5c020
  Fix the tco reference and its mangled description Claude 2026-08-29 19:35:08 +00:00
• caf480aa32
  Harbor A/Bs at n=3 could never resolve their own margins Claude 2026-08-29 19:34:30 +00:00
• 57d8391c52
  Recover most of the cpsat test-suite cost; keep the part that is real Claude 2026-08-29 18:55:52 +00:00
• ac5913151a coldstart programme-house seed 0 @ 500000: 1 fails (0h/1s) Bruno Postle 2026-08-29 18:27:37 +01:00
• 6e8bcce8f3
  Correct a wrong bead ID in 38.20 (2xk -> 7t1) Claude 2026-08-29 14:47:53 +00:00
• 03c1c5edcf
  CP-SAT seeding re-measured deterministically: it loses Claude 2026-08-29 14:47:36 +00:00
• e9e7060fb4
  Re-validate collapse_insearch's default under the current objective Claude 2026-08-29 13:57:45 +00:00
• 50fb293fa5
  Re-verify 1ph: the iio bug could never have touched it Claude 2026-08-29 13:21:29 +00:00
• f6eeee7191
  n_workers is an algorithm parameter, not noise Claude 2026-08-29 12:52:33 +00:00
• 225e673ae3
  Staged harness re-scored under a different objective than it searched Claude 2026-08-29 12:11:38 +00:00
• 148988df14
  constructive_topology was ordered by memory address on the cpsat path Claude 2026-08-29 11:45:19 +00:00
• ee050532d6
  Declare toilet-to-sleeping adjacency where the brief supports it Claude 2026-08-29 10:57:44 +00:00
• 98d403b966
  health-centre: enlarge the plot for a courtyard typology Claude 2026-08-29 10:20:01 +00:00
• df84bfe1e6
  Missing-space cascade no longer weighted by YAML verbosity Claude 2026-08-29 09:33:40 +00:00
• f5286dde3d
  Make both failing tests assert their intent, not stale artefacts Claude 2026-08-29 07:11:25 +00:00
• e5eb397b52
  Checkpoint long searches; the cold-start runs were lost to a reclaimed box Claude 2026-08-29 05:43:58 +00:00
• e13b7b1313
  coldstart: track the in-progress run logs Claude 2026-08-28 23:30:01 +00:00
• 4d7a3e65b9
  Remove cold-start smoke-test artefacts (budget 1000, pipeline check only) Claude 2026-08-28 23:29:17 +00:00
• e2d5838433
  coldstart programme-house seed 0 @ 1000: 7 fails (3h/4s) Claude 2026-08-28 23:29:01 +00:00
• 45eff37f96
  Records Room ruled a store; stays exempt Claude 2026-08-28 23:22:02 +00:00
• c9b9eafef6
  Owner's daylight ruling: corridors need it; retract the 61% headline Claude 2026-08-28 23:14:21 +00:00
• 31d50ba72b
  beads: correct a mangled ssz comment Claude 2026-08-28 22:42:12 +00:00
• 335dd9c1ff
  beads: close bdf, attach ut5's failing test, file 1ue Claude 2026-08-28 22:41:42 +00:00
• 055d7101cd
  ssz: crinkliness is declared per space; there is no daylight attribute Claude 2026-08-28 22:38:54 +00:00
• 0d4ae7ad54 Add missing test files Bruno Postle 2026-08-28 09:46:11 +01:00
• bb27a5e548
  ssz: 61% of reported crinkliness fails are not defects; two corrections Claude 2026-08-26 18:07:01 +00:00
• 6405ca1a59
  ssz: record the A/B result honestly -- not a pass at n=3 Claude 2026-08-26 17:57:11 +00:00
• f1831687c1
  ssz A/B: results snapshot (6/30 arms; run in progress) Claude 2026-08-26 16:50:28 +00:00
• 02967d5de7
  ssz: use the module-level programme import in needs_daylight Claude 2026-08-26 16:46:59 +00:00
• 189efdbfc7
  ssz: daylight is required of rooms that do not need it Claude 2026-08-26 16:45:10 +00:00
• 8522576467
  Frontage feasibility as a pre-flight check; retract the "3x short" claim Claude 2026-08-26 16:20:52 +00:00
• 024311a2c2
  §39.10: preserving constructed connectivity is NULL — and it reframes §39.9 Claude 2026-08-26 15:11:46 +00:00
• 7c41069226
  §39.9: level-not-connected is destroyed by the resize, not by the search Claude 2026-08-26 14:39:07 +00:00
• 3aee813ccd
  §39.8: homemaker-py-2v1 connectivity weighting — MEASURED NULL, premise retracted Claude 2026-08-26 14:15:22 +00:00
• 03aadec1ba
  §39.7: access requirements become a declared usage: attribute (homemaker-py-sel) Claude 2026-08-26 13:39:41 +00:00
• 72673a2ebf
  homemaker-py-sel: usage is a plain space attribute — drop the mapping table Claude 2026-08-26 12:56:12 +00:00
• d5e10172c5
  homemaker-py-sel: all 52 corpus codes assigned; utility distinct; open vocabulary Claude 2026-08-26 12:34:03 +00:00
• 6bed7cc535
  homemaker-py-sel: propose usage for all 52 corpus codes; scope was understated Claude 2026-08-26 10:47:11 +00:00
• f476a5849e
  beads: revise homemaker-py-sel design — mandatory code-level usage, no fallback Claude 2026-08-26 10:32:14 +00:00
• a25dc2cb59
  §39.4 completion + §39.5 retraction + §39.6: the usage namespace is NOT clean Claude 2026-08-26 10:09:14 +00:00
• 7ec4e5d121
  §39.4: tighten generic-type matching, reverting the harbor rename Claude 2026-08-26 09:45:28 +00:00
• 697914fac8
  ju3: reject programme codes colliding with the c/o/s generic type prefixes Claude 2026-08-26 09:06:08 +00:00
• fd9802e499
  DESIGN.md §39: config audit — programme codes collide with the c/o/s type namespace Claude 2026-08-26 08:27:49 +00:00
• 1b3a0adc2e
  beads: file homemaker-py-2v1 (connectivity mispricing, P0) Claude 2026-08-26 07:40:37 +00:00
• 50b5cfd476
  §38.2 refinement: connectivity is under-priced ~3x, not just a crinkliness bug Claude 2026-08-26 07:40:37 +00:00
• 753356e47e
  beads: file §38 objective-gradient findings Claude 2026-08-25 23:18:47 +00:00
• c0c47a6e2d
  DESIGN.md §38: the plateau is an objective-gradient problem, not a search problem Claude 2026-08-25 23:18:20 +00:00
• 7046c7bd76
  Ignore beads auto-import tracking marker Claude 2026-08-25 23:01:40 +00:00
• 25e0503ddc bd: sync issues.jsonl export after 8oq/57z creation main Bruno Postle 2026-08-06 08:08:07 +01:00
• 91f7626a77 bd: close homemaker-py-pek Bruno Postle 2026-08-05 16:59:25 +01:00
• 6dd4ae6470 homemaker-py-pek: delete dead first process_storey definition Bruno Postle 2026-08-05 16:57:20 +01:00
• 5f5e640906 DESIGN.md sec 13.11: correct stale 7ua note, cross-ref follow-up 4ok Bruno Postle 2026-08-05 13:54:42 +01:00
• 88f36177bb homemaker-py-7ua: pin collapse_insearch=True in run_staged_search.py rescore Bruno Postle 2026-08-05 10:08:04 +01:00