FAILURE MAP
← Case archive

FA-74886 / CRDT convergence / Open access

PN-Counter with local floor: negative slots flow against the sync direction · case 01

Receivers do not see the sender's decrements while the sender absorbs the receiver's.

Verified by executionVariant 1 · 9 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

The negative-map loop copies from destination to source, reversing the one-way sync for decrements only.

THE FAILURE

The negative-map loop copies from destination to source, reversing the one-way sync for decrements only.

Unsuccessful approach: Making the negative merge bidirectional lets the sender learn decrements without their matching increments.

Case contract

Replicas keep a positive map and a negative map (replica->total). ["inc", r, k] and ["dec", r, k] need k > 0, otherwise the event index is rejected. A decrement is also rejected when r's locally observed value is below k. ["sync", src, dst] merges both maps of src into dst by pointwise maximum. Return values and rejected indices.

Why this case matters

Positive/negative counters are the standard way to support decrements in a state-based counter.

1 / The failure

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(events, replicas):
    pos = {r: {} for r in replicas}
    neg = {r: {} for r in replicas}
    rejected = []
    def value(r):
        return sum(pos[r].values()) - sum(neg[r].values())
    for i, ev in enumerate(events):
        kind = ev[0]
        if kind == 'sync':
            s, d = ev[1], ev[2]
            for k, v in pos[s].items():
                pos[d][k] = max(pos[d].get(k, 0), v)
            for k, v in neg[d].items():
                neg[s][k] = max(neg[s].get(k, 0), v)
            continue
        r, amt = ev[1], ev[2]
        if amt <= 0:
            rejected.append(i)
            continue
        if kind == 'inc':
            pos[r][r] = pos[r].get(r, 0) + amt
        else:
            if value(r) < amt:
                rejected.append(i)
                continue
            neg[r][r] = neg[r].get(r, 0) + amt
    return {'values': [value(r) for r in replicas], 'rejected': rejected}
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
cases = {
    1: [('decrement to exactly zero is allowed', [[['inc', 'a', 1], ['dec', 'a', 1]], ['a', 'b', 'c']], {'values': [0, 0, 0], 'rejected': []}), ('decrement beyond local view is rejected', [[['inc', 'a', 1], ['dec', 'b', 1], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [1, 2]}), ('decrement funded by learned increments', [[['inc', 'a', 3], ['sync', 'a', 'b'], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'b', 'c'], ['sync', 'a', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('concurrent decrements converge', [[['inc', 'a', 4], ['sync', 'a', 'b'], ['dec', 'a', 1], ['dec', 'b', 2], ['sync', 'a', 'b'], ['sync', 'b', 'a'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('zero amounts are rejected', [[['inc', 'a', 0], ['inc', 'b', 3], ['dec', 'b', 0], ['dec', 'b', 1]], ['a', 'b', 'c']], {'values': [0, 2, 0], 'rejected': [0, 2]}), ('repeated decrement slot is refreshed', [[['inc', 'c', 4], ['dec', 'c', 1], ['sync', 'c', 'a'], ['dec', 'c', 1], ['sync', 'c', 'a'], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [2, 2, 2], 'rejected': []}), ('second decrement cannot pass the floor', [[['inc', 'a', 2], ['dec', 'a', 1], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [2]}), ('one-way sync does not pull from the receiver', [[['inc', 'a', 1], ['inc', 'b', 1], ['dec', 'b', 1], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [1, 1, 0], 'rejected': []}), ('mixed history', [[['inc', 'a', 3], ['dec', 'a', 1], ['sync', 'a', 'b'], ['inc', 'b', 1], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'a', 'c'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []})],
    2: [('decrement to exactly zero is allowed', [[['inc', 'a', 2], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [0, 0, 0], 'rejected': []}), ('decrement beyond local view is rejected', [[['inc', 'a', 2], ['dec', 'b', 1], ['dec', 'a', 3]], ['a', 'b', 'c']], {'values': [2, 0, 0], 'rejected': [1, 2]}), ('decrement funded by learned increments', [[['inc', 'a', 4], ['sync', 'a', 'b'], ['dec', 'b', 3], ['sync', 'b', 'a'], ['sync', 'b', 'c'], ['sync', 'a', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('concurrent decrements converge', [[['inc', 'a', 6], ['sync', 'a', 'b'], ['dec', 'a', 1], ['dec', 'b', 2], ['sync', 'a', 'b'], ['sync', 'b', 'a'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [3, 3, 3], 'rejected': []}), ('zero amounts are rejected', [[['inc', 'a', 0], ['inc', 'b', 3], ['dec', 'b', 0], ['dec', 'b', 1]], ['a', 'b', 'c']], {'values': [0, 2, 0], 'rejected': [0, 2]}), ('repeated decrement slot is refreshed', [[['inc', 'c', 5], ['dec', 'c', 1], ['sync', 'c', 'a'], ['dec', 'c', 1], ['sync', 'c', 'a'], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [3, 3, 3], 'rejected': []}), ('second decrement cannot pass the floor', [[['inc', 'a', 3], ['dec', 'a', 2], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [2]}), ('one-way sync does not pull from the receiver', [[['inc', 'a', 2], ['inc', 'b', 2], ['dec', 'b', 1], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [2, 3, 0], 'rejected': []}), ('mixed history', [[['inc', 'a', 3], ['dec', 'a', 1], ['sync', 'a', 'b'], ['inc', 'b', 2], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'a', 'c'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [2, 2, 2], 'rejected': []})],
    3: [('decrement to exactly zero is allowed', [[['inc', 'a', 3], ['dec', 'a', 3]], ['a', 'b', 'c']], {'values': [0, 0, 0], 'rejected': []}), ('decrement beyond local view is rejected', [[['inc', 'a', 3], ['dec', 'b', 1], ['dec', 'a', 4]], ['a', 'b', 'c']], {'values': [3, 0, 0], 'rejected': [1, 2]}), ('decrement funded by learned increments', [[['inc', 'a', 5], ['sync', 'a', 'b'], ['dec', 'b', 4], ['sync', 'b', 'a'], ['sync', 'b', 'c'], ['sync', 'a', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('concurrent decrements converge', [[['inc', 'a', 8], ['sync', 'a', 'b'], ['dec', 'a', 1], ['dec', 'b', 2], ['sync', 'a', 'b'], ['sync', 'b', 'a'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [5, 5, 5], 'rejected': []}), ('zero amounts are rejected', [[['inc', 'a', 0], ['inc', 'b', 3], ['dec', 'b', 0], ['dec', 'b', 1]], ['a', 'b', 'c']], {'values': [0, 2, 0], 'rejected': [0, 2]}), ('repeated decrement slot is refreshed', [[['inc', 'c', 6], ['dec', 'c', 1], ['sync', 'c', 'a'], ['dec', 'c', 1], ['sync', 'c', 'a'], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [4, 4, 4], 'rejected': []}), ('second decrement cannot pass the floor', [[['inc', 'a', 4], ['dec', 'a', 3], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [2]}), ('one-way sync does not pull from the receiver', [[['inc', 'a', 3], ['inc', 'b', 3], ['dec', 'b', 1], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [3, 5, 0], 'rejected': []}), ('mixed history', [[['inc', 'a', 3], ['dec', 'a', 1], ['sync', 'a', 'b'], ['inc', 'b', 3], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'a', 'c'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [3, 3, 3], 'rejected': []})],
    4: [('decrement to exactly zero is allowed', [[['inc', 'a', 4], ['dec', 'a', 4]], ['a', 'b', 'c']], {'values': [0, 0, 0], 'rejected': []}), ('decrement beyond local view is rejected', [[['inc', 'a', 4], ['dec', 'b', 1], ['dec', 'a', 5]], ['a', 'b', 'c']], {'values': [4, 0, 0], 'rejected': [1, 2]}), ('decrement funded by learned increments', [[['inc', 'a', 6], ['sync', 'a', 'b'], ['dec', 'b', 5], ['sync', 'b', 'a'], ['sync', 'b', 'c'], ['sync', 'a', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('concurrent decrements converge', [[['inc', 'a', 10], ['sync', 'a', 'b'], ['dec', 'a', 1], ['dec', 'b', 2], ['sync', 'a', 'b'], ['sync', 'b', 'a'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [7, 7, 7], 'rejected': []}), ('zero amounts are rejected', [[['inc', 'a', 0], ['inc', 'b', 3], ['dec', 'b', 0], ['dec', 'b', 1]], ['a', 'b', 'c']], {'values': [0, 2, 0], 'rejected': [0, 2]}), ('repeated decrement slot is refreshed', [[['inc', 'c', 7], ['dec', 'c', 1], ['sync', 'c', 'a'], ['dec', 'c', 1], ['sync', 'c', 'a'], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [5, 5, 5], 'rejected': []}), ('second decrement cannot pass the floor', [[['inc', 'a', 5], ['dec', 'a', 4], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [2]}), ('one-way sync does not pull from the receiver', [[['inc', 'a', 4], ['inc', 'b', 4], ['dec', 'b', 1], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [4, 7, 0], 'rejected': []}), ('mixed history', [[['inc', 'a', 3], ['dec', 'a', 1], ['sync', 'a', 'b'], ['inc', 'b', 4], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'a', 'c'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [4, 4, 4], 'rejected': []})],
    5: [('decrement to exactly zero is allowed', [[['inc', 'a', 5], ['dec', 'a', 5]], ['a', 'b', 'c']], {'values': [0, 0, 0], 'rejected': []}), ('decrement beyond local view is rejected', [[['inc', 'a', 5], ['dec', 'b', 1], ['dec', 'a', 6]], ['a', 'b', 'c']], {'values': [5, 0, 0], 'rejected': [1, 2]}), ('decrement funded by learned increments', [[['inc', 'a', 7], ['sync', 'a', 'b'], ['dec', 'b', 6], ['sync', 'b', 'a'], ['sync', 'b', 'c'], ['sync', 'a', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('concurrent decrements converge', [[['inc', 'a', 12], ['sync', 'a', 'b'], ['dec', 'a', 1], ['dec', 'b', 2], ['sync', 'a', 'b'], ['sync', 'b', 'a'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [9, 9, 9], 'rejected': []}), ('zero amounts are rejected', [[['inc', 'a', 0], ['inc', 'b', 3], ['dec', 'b', 0], ['dec', 'b', 1]], ['a', 'b', 'c']], {'values': [0, 2, 0], 'rejected': [0, 2]}), ('repeated decrement slot is refreshed', [[['inc', 'c', 8], ['dec', 'c', 1], ['sync', 'c', 'a'], ['dec', 'c', 1], ['sync', 'c', 'a'], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [6, 6, 6], 'rejected': []}), ('second decrement cannot pass the floor', [[['inc', 'a', 6], ['dec', 'a', 5], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [2]}), ('one-way sync does not pull from the receiver', [[['inc', 'a', 5], ['inc', 'b', 5], ['dec', 'b', 1], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [5, 9, 0], 'rejected': []}), ('mixed history', [[['inc', 'a', 3], ['dec', 'a', 1], ['sync', 'a', 'b'], ['inc', 'b', 5], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'a', 'c'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [5, 5, 5], 'rejected': []})],
}[N]
for label, args, expected in cases:
    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 fixtureActualExpectedOutcome
decrement to exactly zero is allowed{'rejected': [], 'values': [0, 0, 0]}{'rejected': [], 'values': [0, 0, 0]}Passed
decrement beyond local view is rejected{'rejected': [1, 2], 'values': [1, 0, 0]}{'rejected': [1, 2], 'values': [1, 0, 0]}Passed
decrement funded by learned increments{'rejected': [], 'values': [3, 1, 3]}{'rejected': [], 'values': [1, 1, 1]}Failed
concurrent decrements converge{'rejected': [], 'values': [1, 1, 4]}{'rejected': [], 'values': [1, 1, 1]}Failed
zero amounts are rejected{'rejected': [0, 2], 'values': [0, 2, 0]}{'rejected': [0, 2], 'values': [0, 2, 0]}Passed
repeated decrement slot is refreshed{'rejected': [], 'values': [4, 4, 2]}{'rejected': [], 'values': [2, 2, 2]}Failed
second decrement cannot pass the floor{'rejected': [2], 'values': [1, 0, 0]}{'rejected': [2], 'values': [1, 0, 0]}Passed
one-way sync does not pull from the receiver{'rejected': [], 'values': [0, 1, 0]}{'rejected': [], 'values': [1, 1, 0]}Failed
mixed history{'rejected': [], 'values': [3, 1, 4]}{'rejected': [], 'values': [1, 1, 1]}Failed

SHA-256 / f66bb839253ee39590aba92c1dc20c4fa23003d154dbba8c7d3992ba65aff73f

2 / The unsuccessful fix

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(events, replicas):
    pos = {r: {} for r in replicas}
    neg = {r: {} for r in replicas}
    rejected = []
    def value(r):
        return sum(pos[r].values()) - sum(neg[r].values())
    for i, ev in enumerate(events):
        kind = ev[0]
        if kind == 'sync':
            s, d = ev[1], ev[2]
            for k, v in pos[s].items():
                pos[d][k] = max(pos[d].get(k, 0), v)
            for k, v in neg[s].items():
                neg[d][k] = max(neg[d].get(k, 0), v)
            for k, v in neg[d].items():
                neg[s][k] = max(neg[s].get(k, 0), v)
            continue
        r, amt = ev[1], ev[2]
        if amt <= 0:
            rejected.append(i)
            continue
        if kind == 'inc':
            pos[r][r] = pos[r].get(r, 0) + amt
        else:
            if value(r) < amt:
                rejected.append(i)
                continue
            neg[r][r] = neg[r].get(r, 0) + amt
    return {'values': [value(r) for r in replicas], 'rejected': rejected}
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
cases = {
    1: [('decrement to exactly zero is allowed', [[['inc', 'a', 1], ['dec', 'a', 1]], ['a', 'b', 'c']], {'values': [0, 0, 0], 'rejected': []}), ('decrement beyond local view is rejected', [[['inc', 'a', 1], ['dec', 'b', 1], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [1, 2]}), ('decrement funded by learned increments', [[['inc', 'a', 3], ['sync', 'a', 'b'], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'b', 'c'], ['sync', 'a', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('concurrent decrements converge', [[['inc', 'a', 4], ['sync', 'a', 'b'], ['dec', 'a', 1], ['dec', 'b', 2], ['sync', 'a', 'b'], ['sync', 'b', 'a'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('zero amounts are rejected', [[['inc', 'a', 0], ['inc', 'b', 3], ['dec', 'b', 0], ['dec', 'b', 1]], ['a', 'b', 'c']], {'values': [0, 2, 0], 'rejected': [0, 2]}), ('repeated decrement slot is refreshed', [[['inc', 'c', 4], ['dec', 'c', 1], ['sync', 'c', 'a'], ['dec', 'c', 1], ['sync', 'c', 'a'], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [2, 2, 2], 'rejected': []}), ('second decrement cannot pass the floor', [[['inc', 'a', 2], ['dec', 'a', 1], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [2]}), ('one-way sync does not pull from the receiver', [[['inc', 'a', 1], ['inc', 'b', 1], ['dec', 'b', 1], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [1, 1, 0], 'rejected': []}), ('mixed history', [[['inc', 'a', 3], ['dec', 'a', 1], ['sync', 'a', 'b'], ['inc', 'b', 1], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'a', 'c'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []})],
    2: [('decrement to exactly zero is allowed', [[['inc', 'a', 2], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [0, 0, 0], 'rejected': []}), ('decrement beyond local view is rejected', [[['inc', 'a', 2], ['dec', 'b', 1], ['dec', 'a', 3]], ['a', 'b', 'c']], {'values': [2, 0, 0], 'rejected': [1, 2]}), ('decrement funded by learned increments', [[['inc', 'a', 4], ['sync', 'a', 'b'], ['dec', 'b', 3], ['sync', 'b', 'a'], ['sync', 'b', 'c'], ['sync', 'a', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('concurrent decrements converge', [[['inc', 'a', 6], ['sync', 'a', 'b'], ['dec', 'a', 1], ['dec', 'b', 2], ['sync', 'a', 'b'], ['sync', 'b', 'a'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [3, 3, 3], 'rejected': []}), ('zero amounts are rejected', [[['inc', 'a', 0], ['inc', 'b', 3], ['dec', 'b', 0], ['dec', 'b', 1]], ['a', 'b', 'c']], {'values': [0, 2, 0], 'rejected': [0, 2]}), ('repeated decrement slot is refreshed', [[['inc', 'c', 5], ['dec', 'c', 1], ['sync', 'c', 'a'], ['dec', 'c', 1], ['sync', 'c', 'a'], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [3, 3, 3], 'rejected': []}), ('second decrement cannot pass the floor', [[['inc', 'a', 3], ['dec', 'a', 2], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [2]}), ('one-way sync does not pull from the receiver', [[['inc', 'a', 2], ['inc', 'b', 2], ['dec', 'b', 1], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [2, 3, 0], 'rejected': []}), ('mixed history', [[['inc', 'a', 3], ['dec', 'a', 1], ['sync', 'a', 'b'], ['inc', 'b', 2], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'a', 'c'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [2, 2, 2], 'rejected': []})],
    3: [('decrement to exactly zero is allowed', [[['inc', 'a', 3], ['dec', 'a', 3]], ['a', 'b', 'c']], {'values': [0, 0, 0], 'rejected': []}), ('decrement beyond local view is rejected', [[['inc', 'a', 3], ['dec', 'b', 1], ['dec', 'a', 4]], ['a', 'b', 'c']], {'values': [3, 0, 0], 'rejected': [1, 2]}), ('decrement funded by learned increments', [[['inc', 'a', 5], ['sync', 'a', 'b'], ['dec', 'b', 4], ['sync', 'b', 'a'], ['sync', 'b', 'c'], ['sync', 'a', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('concurrent decrements converge', [[['inc', 'a', 8], ['sync', 'a', 'b'], ['dec', 'a', 1], ['dec', 'b', 2], ['sync', 'a', 'b'], ['sync', 'b', 'a'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [5, 5, 5], 'rejected': []}), ('zero amounts are rejected', [[['inc', 'a', 0], ['inc', 'b', 3], ['dec', 'b', 0], ['dec', 'b', 1]], ['a', 'b', 'c']], {'values': [0, 2, 0], 'rejected': [0, 2]}), ('repeated decrement slot is refreshed', [[['inc', 'c', 6], ['dec', 'c', 1], ['sync', 'c', 'a'], ['dec', 'c', 1], ['sync', 'c', 'a'], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [4, 4, 4], 'rejected': []}), ('second decrement cannot pass the floor', [[['inc', 'a', 4], ['dec', 'a', 3], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [2]}), ('one-way sync does not pull from the receiver', [[['inc', 'a', 3], ['inc', 'b', 3], ['dec', 'b', 1], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [3, 5, 0], 'rejected': []}), ('mixed history', [[['inc', 'a', 3], ['dec', 'a', 1], ['sync', 'a', 'b'], ['inc', 'b', 3], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'a', 'c'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [3, 3, 3], 'rejected': []})],
    4: [('decrement to exactly zero is allowed', [[['inc', 'a', 4], ['dec', 'a', 4]], ['a', 'b', 'c']], {'values': [0, 0, 0], 'rejected': []}), ('decrement beyond local view is rejected', [[['inc', 'a', 4], ['dec', 'b', 1], ['dec', 'a', 5]], ['a', 'b', 'c']], {'values': [4, 0, 0], 'rejected': [1, 2]}), ('decrement funded by learned increments', [[['inc', 'a', 6], ['sync', 'a', 'b'], ['dec', 'b', 5], ['sync', 'b', 'a'], ['sync', 'b', 'c'], ['sync', 'a', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('concurrent decrements converge', [[['inc', 'a', 10], ['sync', 'a', 'b'], ['dec', 'a', 1], ['dec', 'b', 2], ['sync', 'a', 'b'], ['sync', 'b', 'a'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [7, 7, 7], 'rejected': []}), ('zero amounts are rejected', [[['inc', 'a', 0], ['inc', 'b', 3], ['dec', 'b', 0], ['dec', 'b', 1]], ['a', 'b', 'c']], {'values': [0, 2, 0], 'rejected': [0, 2]}), ('repeated decrement slot is refreshed', [[['inc', 'c', 7], ['dec', 'c', 1], ['sync', 'c', 'a'], ['dec', 'c', 1], ['sync', 'c', 'a'], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [5, 5, 5], 'rejected': []}), ('second decrement cannot pass the floor', [[['inc', 'a', 5], ['dec', 'a', 4], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [2]}), ('one-way sync does not pull from the receiver', [[['inc', 'a', 4], ['inc', 'b', 4], ['dec', 'b', 1], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [4, 7, 0], 'rejected': []}), ('mixed history', [[['inc', 'a', 3], ['dec', 'a', 1], ['sync', 'a', 'b'], ['inc', 'b', 4], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'a', 'c'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [4, 4, 4], 'rejected': []})],
    5: [('decrement to exactly zero is allowed', [[['inc', 'a', 5], ['dec', 'a', 5]], ['a', 'b', 'c']], {'values': [0, 0, 0], 'rejected': []}), ('decrement beyond local view is rejected', [[['inc', 'a', 5], ['dec', 'b', 1], ['dec', 'a', 6]], ['a', 'b', 'c']], {'values': [5, 0, 0], 'rejected': [1, 2]}), ('decrement funded by learned increments', [[['inc', 'a', 7], ['sync', 'a', 'b'], ['dec', 'b', 6], ['sync', 'b', 'a'], ['sync', 'b', 'c'], ['sync', 'a', 'c']], ['a', 'b', 'c']], {'values': [1, 1, 1], 'rejected': []}), ('concurrent decrements converge', [[['inc', 'a', 12], ['sync', 'a', 'b'], ['dec', 'a', 1], ['dec', 'b', 2], ['sync', 'a', 'b'], ['sync', 'b', 'a'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [9, 9, 9], 'rejected': []}), ('zero amounts are rejected', [[['inc', 'a', 0], ['inc', 'b', 3], ['dec', 'b', 0], ['dec', 'b', 1]], ['a', 'b', 'c']], {'values': [0, 2, 0], 'rejected': [0, 2]}), ('repeated decrement slot is refreshed', [[['inc', 'c', 8], ['dec', 'c', 1], ['sync', 'c', 'a'], ['dec', 'c', 1], ['sync', 'c', 'a'], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [6, 6, 6], 'rejected': []}), ('second decrement cannot pass the floor', [[['inc', 'a', 6], ['dec', 'a', 5], ['dec', 'a', 2]], ['a', 'b', 'c']], {'values': [1, 0, 0], 'rejected': [2]}), ('one-way sync does not pull from the receiver', [[['inc', 'a', 5], ['inc', 'b', 5], ['dec', 'b', 1], ['sync', 'a', 'b']], ['a', 'b', 'c']], {'values': [5, 9, 0], 'rejected': []}), ('mixed history', [[['inc', 'a', 3], ['dec', 'a', 1], ['sync', 'a', 'b'], ['inc', 'b', 5], ['dec', 'b', 2], ['sync', 'b', 'a'], ['sync', 'a', 'c'], ['sync', 'b', 'c']], ['a', 'b', 'c']], {'values': [5, 5, 5], 'rejected': []})],
}[N]
for label, args, expected in cases:
    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 fixtureActualExpectedOutcome
decrement to exactly zero is allowed{'rejected': [], 'values': [0, 0, 0]}{'rejected': [], 'values': [0, 0, 0]}Passed
decrement beyond local view is rejected{'rejected': [1, 2], 'values': [1, 0, 0]}{'rejected': [1, 2], 'values': [1, 0, 0]}Passed
decrement funded by learned increments{'rejected': [], 'values': [1, 1, 1]}{'rejected': [], 'values': [1, 1, 1]}Passed
concurrent decrements converge{'rejected': [], 'values': [1, 1, 1]}{'rejected': [], 'values': [1, 1, 1]}Passed
zero amounts are rejected{'rejected': [0, 2], 'values': [0, 2, 0]}{'rejected': [0, 2], 'values': [0, 2, 0]}Passed
repeated decrement slot is refreshed{'rejected': [], 'values': [2, 2, 2]}{'rejected': [], 'values': [2, 2, 2]}Passed
second decrement cannot pass the floor{'rejected': [2], 'values': [1, 0, 0]}{'rejected': [2], 'values': [1, 0, 0]}Passed
one-way sync does not pull from the receiver{'rejected': [], 'values': [0, 1, 0]}{'rejected': [], 'values': [1, 1, 0]}Failed
mixed history{'rejected': [], 'values': [1, 1, 1]}{'rejected': [], 'values': [1, 1, 1]}Passed

SHA-256 / 7e47f1a829e70ddb1fe29fd75251beacaded38951375118b7bdac5e6b1248de3

HELD IN THE MEMBER ARCHIVE

The verified repair and its recorded checks are member-only.

This mechanism has 9 recorded checks per implementation. The open-access tier publishes the failure and the unsuccessful fix; the repaired source that passes every check, and the observations that prove it, are available to members.

Every case sharing this mechanism uses the same contract and the same repair, so this one record is held back for all of them.

Member access is invitation-based. Sign in with your invited account to inspect the repair.

Sign in to the archive ↗

Verification & scope

A deterministic, bounded teaching model of one replicated data type with stipulated operation and merge rules; it is not a production CRDT library and makes no claim of conformance to any specific published design. 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 2026-09-29T14:49:00.927853+00:00.

Case digest / 492396fcf1dcc7a734ed691c19b43dff73136bb0b1cd2ec8833b4835a0c13f2c