FA-94136 / Shift rostering labor rules / Open access
Leave taken before the accrual cap is enforced · case 01
Workers at the cap keep accrual they should have forfeited because leave was deducted first.
ROOT CAUSE
Usage is deducted before the balance is capped.
VERIFIED REPAIR
Cap the balance immediately after accrual, then deduct approved leave.
Unsuccessful approach: Netting the request into the cap test still forfeits the wrong amount.
Case contract
Per pay period [worked minutes, leave minutes requested]. Accrual is worked * rate_bp / 10000 minutes with the fractional remainder (in 1/10000 minutes) carried to later periods. After accrual the balance is capped (excess forfeited), then the request is approved up to the balance; unapproved minutes accumulate. Return [balance, forfeited, unapproved].
Why this case matters
Leave accrual ledgers attached to rosters must carry fractions and apply caps in the right order.
1 / The failure
Exit 1"""Failure Map reference implementation. Python standard library only."""
import json
N = 1
observations = []
def solve(periods, rate_bp, cap):
bal = carry = forfeit = unappr = 0
for worked, used in periods:
num = worked * rate_bp + carry
bal += num // 10000
carry = num % 10000
take = min(used, bal)
unappr += used - take
bal -= take
if bal > cap:
forfeit += bal - cap
bal = cap
return [bal, forfeit, unappr]
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
fixtures = [[('regression: cap before usage 1', [[[2400, 0], [2400, 500]], 1207, 400], [0, 179, 100]),
('regression variant: cap before usage 2',
[[[0, 480], [0, 0], [2280, 240], [2400, 0], [2280, 0], [2700, 900]], 833, 400], [0, 215, 1031]),
('partial repair guard 3',
[[[1200, 0], [2280, 0], [1200, 240], [2400, 0], [2700, 0], [2700, 2000]], 1207, 600], [0, 666, 1400]),
('boundary control 4', [[[99, 0], [99, 0], [2, 0]], 5000, 100], [100, 0, 0]),
('boundary control 5', [[[2400, 0], [2400, 0], [2400, 0]], 1207, 11200], [869, 0, 0]),
('normal control 6',
[[[2700, 240], [2298, 0], [2400, 240], [2700, 0], [2280, 0], [319, 0], [1200, 0]], 1207, 1200],
[1197, 0, 0]),
('normal control 7', [[[2400, 900], [2400, 480], [419, 240], [2400, 900], [0, 0]], 833, 1200],
[0, 0, 1886]),
('normal control 8',
[[[2280, 0], [1990, 0], [1908, 480], [2700, 240], [1835, 2000], [2700, 240]], 833, 11200], [0, 0, 1843])],
[('regression: cap before usage 1', [[[0, 2000], [0, 0], [2597, 0], [2280, 2000], [2400, 0]], 1207, 400],
[290, 188, 3600]),
('regression variant: cap before usage 2',
[[[2700, 0], [0, 0], [2280, 480], [2019, 900], [2400, 480], [2700, 900], [2055, 0], [2169, 2000]], 1207,
400],
[0, 311, 3101]),
('partial repair guard 3',
[[[2700, 0], [2280, 2000], [1200, 0], [2700, 240], [0, 240], [2700, 0]], 1207, 400], [326, 271, 1680]),
('boundary control 4', [[[2400, 400], [0, 400]], 1207, 11200], [0, 0, 511]),
('boundary control 5', [[[99, 0], [99, 0], [2, 0]], 5000, 100], [100, 0, 0]),
('normal control 6', [[[0, 0], [1200, 0]], 769, 600], [92, 0, 0]),
('normal control 7',
[[[0, 480], [2280, 0], [0, 480], [2280, 240], [1200, 0], [1200, 0], [2700, 240]], 1207, 11200],
[410, 0, 685]),
('normal control 8', [[[2094, 0], [2400, 2000]], 769, 1200], [0, 0, 1655])],
[('regression: cap before usage 1',
[[[0, 0], [2400, 900], [2700, 0], [2280, 900], [2700, 2000], [1200, 0], [1200, 0], [2280, 0]], 1207, 600],
[565, 1, 2585]),
('regression variant: cap before usage 2',
[[[2280, 900], [2700, 0], [2400, 0], [1034, 2000], [2700, 2000], [1200, 0]], 1207, 400],
[145, 340, 3899]),
('partial repair guard 3',
[[[2700, 0], [0, 0], [2280, 480], [2019, 900], [2400, 480], [2700, 900], [2055, 0], [2169, 2000]], 1207,
400],
[0, 311, 3101]),
('boundary control 4', [[[2400, 0], [2400, 500]], 1207, 400], [0, 179, 100]),
('boundary control 5', [[[2400, 400], [0, 400]], 1207, 11200], [0, 0, 511]),
('normal control 6', [[[1200, 2000], [2700, 0], [2280, 240], [0, 0], [2700, 0]], 833, 600], [400, 0, 1901]),
('normal control 7', [[[2400, 0], [1200, 0], [0, 0], [0, 900], [1200, 0]], 1000, 600], [120, 0, 540]),
('normal control 8', [[[2400, 2000], [0, 900], [2700, 480]], 769, 1200], [0, 0, 2988])],
[('regression: cap before usage 1',
[[[2400, 240], [2280, 0], [1200, 0], [2700, 0], [2400, 0], [887, 900]], 769, 600], [0, 128, 356]),
('regression variant: cap before usage 2',
[[[1200, 0], [1754, 0], [2700, 0], [1200, 0], [1200, 0], [2280, 480], [2400, 240], [2700, 0]], 1000, 600],
[390, 433, 0]),
('partial repair guard 3', [[[2280, 0], [2700, 0], [1200, 240], [2400, 240]], 1000, 600], [360, 18, 0]),
('boundary control 4', [[[2400, 0], [2400, 0], [2400, 0]], 1207, 11200], [869, 0, 0]),
('boundary control 5', [[[2400, 0], [2400, 500]], 1207, 400], [0, 179, 100]),
('normal control 6', [[[2400, 0], [1200, 0], [0, 0], [0, 900], [1200, 0]], 1000, 600], [120, 0, 540]),
('normal control 7', [[[0, 240], [2280, 2000], [2400, 900], [2700, 480]], 1207, 400], [0, 0, 2730]),
('normal control 8', [[[1200, 900], [2700, 2000], [2700, 900]], 769, 11200], [0, 0, 3293])],
[('regression: cap before usage 1',
[[[2479, 2000], [2700, 0], [1200, 0], [1200, 0], [2400, 480], [2400, 0]], 1000, 400], [240, 350, 1833]),
('regression variant: cap before usage 2',
[[[2400, 480], [1434, 2000], [1200, 480], [0, 900], [2400, 0], [0, 0], [2400, 0], [2400, 480]], 1207, 400],
[0, 469, 3333]),
('partial repair guard 3',
[[[0, 0], [2400, 480], [2400, 240], [2400, 0], [1200, 240], [0, 480], [2280, 2000], [1200, 900]], 1207,
400],
[0, 84, 2991]),
('boundary control 4', [[[99, 0], [99, 0], [2, 0]], 5000, 100], [100, 0, 0]),
('boundary control 5', [[[2400, 0], [2400, 0], [2400, 0]], 1207, 11200], [869, 0, 0]),
('normal control 6', [[[2280, 240], [2400, 900], [2928, 240]], 769, 600], [0, 0, 795]),
('normal control 7', [[[2700, 240], [2280, 240], [2700, 900], [2400, 480], [2400, 2000]], 1000, 600],
[0, 0, 2612]),
('normal control 8', [[[0, 480], [2400, 0], [0, 0], [2700, 0], [0, 240], [1200, 0]], 769, 600],
[244, 0, 480])]]
for label, args, expected in fixtures[N-1]:
check(label, solve(*args), expected)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
| Boundary fixture | Actual | Expected | Outcome |
|---|---|---|---|
| regression: cap before usage 1 | [79, 0, 0] | [0, 179, 100] | Failed |
| regression variant: cap before usage 2 | [0, 0, 816] | [0, 215, 1031] | Failed |
| partial repair guard 3 | [0, 340, 1074] | [0, 666, 1400] | Failed |
| boundary control 4 | [100, 0, 0] | [100, 0, 0] | Passed |
| boundary control 5 | [869, 0, 0] | [869, 0, 0] | Passed |
| normal control 6 | [1197, 0, 0] | [1197, 0, 0] | Passed |
| normal control 7 | [0, 0, 1886] | [0, 0, 1886] | Passed |
| normal control 8 | [0, 0, 1843] | [0, 0, 1843] | Passed |
SHA-256 / 8c956cca19b9f474bedf438663060bc815defa24b03d8ff754694e51d475ef0c
2 / The unsuccessful fix
Exit 1"""Failure Map reference implementation. Python standard library only."""
import json
N = 1
observations = []
def solve(periods, rate_bp, cap):
bal = carry = forfeit = unappr = 0
for worked, used in periods:
num = worked * rate_bp + carry
bal += num // 10000
carry = num % 10000
if bal - used > cap:
forfeit += bal - cap
bal = cap
take = min(used, bal)
unappr += used - take
bal -= take
return [bal, forfeit, unappr]
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
fixtures = [[('regression: cap before usage 1', [[[2400, 0], [2400, 500]], 1207, 400], [0, 179, 100]),
('regression variant: cap before usage 2',
[[[0, 480], [0, 0], [2280, 240], [2400, 0], [2280, 0], [2700, 900]], 833, 400], [0, 215, 1031]),
('partial repair guard 3',
[[[1200, 0], [2280, 0], [1200, 240], [2400, 0], [2700, 0], [2700, 2000]], 1207, 600], [0, 666, 1400]),
('boundary control 4', [[[99, 0], [99, 0], [2, 0]], 5000, 100], [100, 0, 0]),
('boundary control 5', [[[2400, 0], [2400, 0], [2400, 0]], 1207, 11200], [869, 0, 0]),
('normal control 6',
[[[2700, 240], [2298, 0], [2400, 240], [2700, 0], [2280, 0], [319, 0], [1200, 0]], 1207, 1200],
[1197, 0, 0]),
('normal control 7', [[[2400, 900], [2400, 480], [419, 240], [2400, 900], [0, 0]], 833, 1200],
[0, 0, 1886]),
('normal control 8',
[[[2280, 0], [1990, 0], [1908, 480], [2700, 240], [1835, 2000], [2700, 240]], 833, 11200], [0, 0, 1843])],
[('regression: cap before usage 1', [[[0, 2000], [0, 0], [2597, 0], [2280, 2000], [2400, 0]], 1207, 400],
[290, 188, 3600]),
('regression variant: cap before usage 2',
[[[2700, 0], [0, 0], [2280, 480], [2019, 900], [2400, 480], [2700, 900], [2055, 0], [2169, 2000]], 1207,
400],
[0, 311, 3101]),
('partial repair guard 3',
[[[2700, 0], [2280, 2000], [1200, 0], [2700, 240], [0, 240], [2700, 0]], 1207, 400], [326, 271, 1680]),
('boundary control 4', [[[2400, 400], [0, 400]], 1207, 11200], [0, 0, 511]),
('boundary control 5', [[[99, 0], [99, 0], [2, 0]], 5000, 100], [100, 0, 0]),
('normal control 6', [[[0, 0], [1200, 0]], 769, 600], [92, 0, 0]),
('normal control 7',
[[[0, 480], [2280, 0], [0, 480], [2280, 240], [1200, 0], [1200, 0], [2700, 240]], 1207, 11200],
[410, 0, 685]),
('normal control 8', [[[2094, 0], [2400, 2000]], 769, 1200], [0, 0, 1655])],
[('regression: cap before usage 1',
[[[0, 0], [2400, 900], [2700, 0], [2280, 900], [2700, 2000], [1200, 0], [1200, 0], [2280, 0]], 1207, 600],
[565, 1, 2585]),
('regression variant: cap before usage 2',
[[[2280, 900], [2700, 0], [2400, 0], [1034, 2000], [2700, 2000], [1200, 0]], 1207, 400],
[145, 340, 3899]),
('partial repair guard 3',
[[[2700, 0], [0, 0], [2280, 480], [2019, 900], [2400, 480], [2700, 900], [2055, 0], [2169, 2000]], 1207,
400],
[0, 311, 3101]),
('boundary control 4', [[[2400, 0], [2400, 500]], 1207, 400], [0, 179, 100]),
('boundary control 5', [[[2400, 400], [0, 400]], 1207, 11200], [0, 0, 511]),
('normal control 6', [[[1200, 2000], [2700, 0], [2280, 240], [0, 0], [2700, 0]], 833, 600], [400, 0, 1901]),
('normal control 7', [[[2400, 0], [1200, 0], [0, 0], [0, 900], [1200, 0]], 1000, 600], [120, 0, 540]),
('normal control 8', [[[2400, 2000], [0, 900], [2700, 480]], 769, 1200], [0, 0, 2988])],
[('regression: cap before usage 1',
[[[2400, 240], [2280, 0], [1200, 0], [2700, 0], [2400, 0], [887, 900]], 769, 600], [0, 128, 356]),
('regression variant: cap before usage 2',
[[[1200, 0], [1754, 0], [2700, 0], [1200, 0], [1200, 0], [2280, 480], [2400, 240], [2700, 0]], 1000, 600],
[390, 433, 0]),
('partial repair guard 3', [[[2280, 0], [2700, 0], [1200, 240], [2400, 240]], 1000, 600], [360, 18, 0]),
('boundary control 4', [[[2400, 0], [2400, 0], [2400, 0]], 1207, 11200], [869, 0, 0]),
('boundary control 5', [[[2400, 0], [2400, 500]], 1207, 400], [0, 179, 100]),
('normal control 6', [[[2400, 0], [1200, 0], [0, 0], [0, 900], [1200, 0]], 1000, 600], [120, 0, 540]),
('normal control 7', [[[0, 240], [2280, 2000], [2400, 900], [2700, 480]], 1207, 400], [0, 0, 2730]),
('normal control 8', [[[1200, 900], [2700, 2000], [2700, 900]], 769, 11200], [0, 0, 3293])],
[('regression: cap before usage 1',
[[[2479, 2000], [2700, 0], [1200, 0], [1200, 0], [2400, 480], [2400, 0]], 1000, 400], [240, 350, 1833]),
('regression variant: cap before usage 2',
[[[2400, 480], [1434, 2000], [1200, 480], [0, 900], [2400, 0], [0, 0], [2400, 0], [2400, 480]], 1207, 400],
[0, 469, 3333]),
('partial repair guard 3',
[[[0, 0], [2400, 480], [2400, 240], [2400, 0], [1200, 240], [0, 480], [2280, 2000], [1200, 900]], 1207,
400],
[0, 84, 2991]),
('boundary control 4', [[[99, 0], [99, 0], [2, 0]], 5000, 100], [100, 0, 0]),
('boundary control 5', [[[2400, 0], [2400, 0], [2400, 0]], 1207, 11200], [869, 0, 0]),
('normal control 6', [[[2280, 240], [2400, 900], [2928, 240]], 769, 600], [0, 0, 795]),
('normal control 7', [[[2700, 240], [2280, 240], [2700, 900], [2400, 480], [2400, 2000]], 1000, 600],
[0, 0, 2612]),
('normal control 8', [[[0, 480], [2400, 0], [0, 0], [2700, 0], [0, 240], [1200, 0]], 769, 600],
[244, 0, 480])]]
for label, args, expected in fixtures[N-1]:
check(label, solve(*args), expected)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
| Boundary fixture | Actual | Expected | Outcome |
|---|---|---|---|
| regression: cap before usage 1 | [79, 0, 0] | [0, 179, 100] | Failed |
| regression variant: cap before usage 2 | [0, 0, 816] | [0, 215, 1031] | Failed |
| partial repair guard 3 | [0, 340, 1074] | [0, 666, 1400] | Failed |
| boundary control 4 | [100, 0, 0] | [100, 0, 0] | Passed |
| boundary control 5 | [869, 0, 0] | [869, 0, 0] | Passed |
| normal control 6 | [1197, 0, 0] | [1197, 0, 0] | Passed |
| normal control 7 | [0, 0, 1886] | [0, 0, 1886] | Passed |
| normal control 8 | [0, 0, 1843] | [0, 0, 1843] | Passed |
SHA-256 / c58211feb4fc0c7fcb44508edf91aae96b7120f4d24fc2f1713d1a3d2d8777fb
3 / The verified repair
Exit 0"""Failure Map reference implementation. Python standard library only."""
import json
N = 1
observations = []
def solve(periods, rate_bp, cap):
bal = carry = forfeit = unappr = 0
for worked, used in periods:
num = worked * rate_bp + carry
bal += num // 10000
carry = num % 10000
if bal > cap:
forfeit += bal - cap
bal = cap
take = min(used, bal)
unappr += used - take
bal -= take
return [bal, forfeit, unappr]
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
fixtures = [[('regression: cap before usage 1', [[[2400, 0], [2400, 500]], 1207, 400], [0, 179, 100]),
('regression variant: cap before usage 2',
[[[0, 480], [0, 0], [2280, 240], [2400, 0], [2280, 0], [2700, 900]], 833, 400], [0, 215, 1031]),
('partial repair guard 3',
[[[1200, 0], [2280, 0], [1200, 240], [2400, 0], [2700, 0], [2700, 2000]], 1207, 600], [0, 666, 1400]),
('boundary control 4', [[[99, 0], [99, 0], [2, 0]], 5000, 100], [100, 0, 0]),
('boundary control 5', [[[2400, 0], [2400, 0], [2400, 0]], 1207, 11200], [869, 0, 0]),
('normal control 6',
[[[2700, 240], [2298, 0], [2400, 240], [2700, 0], [2280, 0], [319, 0], [1200, 0]], 1207, 1200],
[1197, 0, 0]),
('normal control 7', [[[2400, 900], [2400, 480], [419, 240], [2400, 900], [0, 0]], 833, 1200],
[0, 0, 1886]),
('normal control 8',
[[[2280, 0], [1990, 0], [1908, 480], [2700, 240], [1835, 2000], [2700, 240]], 833, 11200], [0, 0, 1843])],
[('regression: cap before usage 1', [[[0, 2000], [0, 0], [2597, 0], [2280, 2000], [2400, 0]], 1207, 400],
[290, 188, 3600]),
('regression variant: cap before usage 2',
[[[2700, 0], [0, 0], [2280, 480], [2019, 900], [2400, 480], [2700, 900], [2055, 0], [2169, 2000]], 1207,
400],
[0, 311, 3101]),
('partial repair guard 3',
[[[2700, 0], [2280, 2000], [1200, 0], [2700, 240], [0, 240], [2700, 0]], 1207, 400], [326, 271, 1680]),
('boundary control 4', [[[2400, 400], [0, 400]], 1207, 11200], [0, 0, 511]),
('boundary control 5', [[[99, 0], [99, 0], [2, 0]], 5000, 100], [100, 0, 0]),
('normal control 6', [[[0, 0], [1200, 0]], 769, 600], [92, 0, 0]),
('normal control 7',
[[[0, 480], [2280, 0], [0, 480], [2280, 240], [1200, 0], [1200, 0], [2700, 240]], 1207, 11200],
[410, 0, 685]),
('normal control 8', [[[2094, 0], [2400, 2000]], 769, 1200], [0, 0, 1655])],
[('regression: cap before usage 1',
[[[0, 0], [2400, 900], [2700, 0], [2280, 900], [2700, 2000], [1200, 0], [1200, 0], [2280, 0]], 1207, 600],
[565, 1, 2585]),
('regression variant: cap before usage 2',
[[[2280, 900], [2700, 0], [2400, 0], [1034, 2000], [2700, 2000], [1200, 0]], 1207, 400],
[145, 340, 3899]),
('partial repair guard 3',
[[[2700, 0], [0, 0], [2280, 480], [2019, 900], [2400, 480], [2700, 900], [2055, 0], [2169, 2000]], 1207,
400],
[0, 311, 3101]),
('boundary control 4', [[[2400, 0], [2400, 500]], 1207, 400], [0, 179, 100]),
('boundary control 5', [[[2400, 400], [0, 400]], 1207, 11200], [0, 0, 511]),
('normal control 6', [[[1200, 2000], [2700, 0], [2280, 240], [0, 0], [2700, 0]], 833, 600], [400, 0, 1901]),
('normal control 7', [[[2400, 0], [1200, 0], [0, 0], [0, 900], [1200, 0]], 1000, 600], [120, 0, 540]),
('normal control 8', [[[2400, 2000], [0, 900], [2700, 480]], 769, 1200], [0, 0, 2988])],
[('regression: cap before usage 1',
[[[2400, 240], [2280, 0], [1200, 0], [2700, 0], [2400, 0], [887, 900]], 769, 600], [0, 128, 356]),
('regression variant: cap before usage 2',
[[[1200, 0], [1754, 0], [2700, 0], [1200, 0], [1200, 0], [2280, 480], [2400, 240], [2700, 0]], 1000, 600],
[390, 433, 0]),
('partial repair guard 3', [[[2280, 0], [2700, 0], [1200, 240], [2400, 240]], 1000, 600], [360, 18, 0]),
('boundary control 4', [[[2400, 0], [2400, 0], [2400, 0]], 1207, 11200], [869, 0, 0]),
('boundary control 5', [[[2400, 0], [2400, 500]], 1207, 400], [0, 179, 100]),
('normal control 6', [[[2400, 0], [1200, 0], [0, 0], [0, 900], [1200, 0]], 1000, 600], [120, 0, 540]),
('normal control 7', [[[0, 240], [2280, 2000], [2400, 900], [2700, 480]], 1207, 400], [0, 0, 2730]),
('normal control 8', [[[1200, 900], [2700, 2000], [2700, 900]], 769, 11200], [0, 0, 3293])],
[('regression: cap before usage 1',
[[[2479, 2000], [2700, 0], [1200, 0], [1200, 0], [2400, 480], [2400, 0]], 1000, 400], [240, 350, 1833]),
('regression variant: cap before usage 2',
[[[2400, 480], [1434, 2000], [1200, 480], [0, 900], [2400, 0], [0, 0], [2400, 0], [2400, 480]], 1207, 400],
[0, 469, 3333]),
('partial repair guard 3',
[[[0, 0], [2400, 480], [2400, 240], [2400, 0], [1200, 240], [0, 480], [2280, 2000], [1200, 900]], 1207,
400],
[0, 84, 2991]),
('boundary control 4', [[[99, 0], [99, 0], [2, 0]], 5000, 100], [100, 0, 0]),
('boundary control 5', [[[2400, 0], [2400, 0], [2400, 0]], 1207, 11200], [869, 0, 0]),
('normal control 6', [[[2280, 240], [2400, 900], [2928, 240]], 769, 600], [0, 0, 795]),
('normal control 7', [[[2700, 240], [2280, 240], [2700, 900], [2400, 480], [2400, 2000]], 1000, 600],
[0, 0, 2612]),
('normal control 8', [[[0, 480], [2400, 0], [0, 0], [2700, 0], [0, 240], [1200, 0]], 769, 600],
[244, 0, 480])]]
for label, args, expected in fixtures[N-1]:
check(label, solve(*args), expected)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
| Boundary fixture | Actual | Expected | Outcome |
|---|---|---|---|
| regression: cap before usage 1 | [0, 179, 100] | [0, 179, 100] | Passed |
| regression variant: cap before usage 2 | [0, 215, 1031] | [0, 215, 1031] | Passed |
| partial repair guard 3 | [0, 666, 1400] | [0, 666, 1400] | Passed |
| boundary control 4 | [100, 0, 0] | [100, 0, 0] | Passed |
| boundary control 5 | [869, 0, 0] | [869, 0, 0] | Passed |
| normal control 6 | [1197, 0, 0] | [1197, 0, 0] | Passed |
| normal control 7 | [0, 0, 1886] | [0, 0, 1886] | Passed |
| normal control 8 | [0, 0, 1843] | [0, 0, 1843] | Passed |
SHA-256 / b7b80d415944d8aaa4b9b25e76cc5700483026145027596e5a0e1578b37fb249
Verification & scope
Stipulated toy labor rule for a bounded roster model; it is not legal advice and does not claim conformance with any jurisdiction, award, or collective agreement. This reproducer isolates one failure mechanism. Results cover the supplied fixtures. Variants within a family share a test contract and should remain grouped when constructing evaluation splits. Related mechanisms with a shared evaluation_group must also remain together; these controlled models are not independent production incidents.
Observations recorded using Python 3.12.14 at 2026-09-29T14:52:01.494990+00:00.
Case digest / 68322b09c257d4adae988c00eb532d223a3626c6c558ab86298d3a82d8816e5e