FA-75146 / CRDT convergence / Open access
Formatting mark spans: the last delivered operation wins · case 01
Replicas that receive the same formatting operations in different orders render different bold spans.
ROOT CAUSE
Each covered character takes whichever operation was applied last instead of the one with the largest stamp.
VERIFIED REPAIR
Keep, per character, the operation with the largest (counter, replica) stamp.
Unsuccessful approach: Comparing only counters leaves equal-counter operations from different replicas in arrival order.
Case contract
A text of `length` characters receives ["bold"|"unbold", start, end, stamp] operations over half-open ranges clamped to the text. Each character takes the operation with the largest stamp (counter, replica) covering it; it is bold when that operation is "bold". Return maximal bold spans as [start, end) pairs in order.
Why this case matters
Rich-text CRDTs must resolve overlapping formatting operations identically regardless of arrival order.
1 / The failure
Exit 1"""Failure Map reference implementation. Python standard library only."""
import json
N = 1
observations = []
def solve(length, ops):
best = [None] * length
for kind, start, end, stamp in ops:
st = tuple(stamp)
for i in range(max(start, 0), min(end, length)):
best[i] = (st, kind)
spans = []
for i, b in enumerate(best):
if b is not None and b[1] == 'bold':
if spans and spans[-1][1] == i:
spans[-1][1] = i + 1
else:
spans.append([i, i + 1])
return spans
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
cases = {
1: [('later unbold wins inside a bold range', [9, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [9, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [9, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [9, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [9, [['bold', 6, 9, [1, 'a']]]], [[6, 9]]), ('range end is exclusive', [9, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [9, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [9, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [9, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [9, [['bold', 8, 10, [1, 'a']], ['unbold', -1, 1, [2, 'a']]]], [[8, 9]])],
2: [('later unbold wins inside a bold range', [10, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [10, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [10, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [10, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [10, [['bold', 7, 10, [1, 'a']]]], [[7, 10]]), ('range end is exclusive', [10, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [10, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [10, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [10, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [10, [['bold', 9, 12, [1, 'a']], ['unbold', -2, 1, [2, 'a']]]], [[9, 10]])],
3: [('later unbold wins inside a bold range', [11, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [11, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [11, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [11, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [11, [['bold', 8, 11, [1, 'a']]]], [[8, 11]]), ('range end is exclusive', [11, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [11, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [11, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [11, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [11, [['bold', 10, 14, [1, 'a']], ['unbold', -3, 1, [2, 'a']]]], [[10, 11]])],
4: [('later unbold wins inside a bold range', [12, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [12, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [12, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [12, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [12, [['bold', 9, 12, [1, 'a']]]], [[9, 12]]), ('range end is exclusive', [12, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [12, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [12, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [12, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [12, [['bold', 11, 16, [1, 'a']], ['unbold', -4, 1, [2, 'a']]]], [[11, 12]])],
5: [('later unbold wins inside a bold range', [13, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [13, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [13, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [13, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [13, [['bold', 10, 13, [1, 'a']]]], [[10, 13]]), ('range end is exclusive', [13, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [13, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [13, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [13, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [13, [['bold', 12, 18, [1, 'a']], ['unbold', -5, 1, [2, 'a']]]], [[12, 13]])],
}[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 fixture | Actual | Expected | Outcome |
|---|---|---|---|
| later unbold wins inside a bold range | [[0, 2], [4, 6]] | [[0, 2], [4, 6]] | Passed |
| arrival order does not matter | [[0, 6]] | [[0, 2], [4, 6]] | Failed |
| equal counters break by replica | [[1, 3]] | [[1, 5]] | Failed |
| equal counters other arrival | [[1, 5]] | [[1, 5]] | Passed |
| mark reaching the end of text | [[6, 9]] | [[6, 9]] | Passed |
| range end is exclusive | [[2, 3], [5, 6]] | [[2, 3], [5, 6]] | Passed |
| adjacent ranges join | [[0, 4]] | [[0, 4]] | Passed |
| one-character gap stays split | [[0, 2], [3, 5]] | [[0, 2], [3, 5]] | Passed |
| range starting before the text is clamped | [[0, 2]] | [[0, 2]] | Passed |
| range past the end is clamped | [[8, 9]] | [[8, 9]] | Passed |
SHA-256 / cbcb06944c543548e5abf383e75e9e34908883072861cde0aae76adad6936cf3
2 / The unsuccessful fix
Exit 1"""Failure Map reference implementation. Python standard library only."""
import json
N = 1
observations = []
def solve(length, ops):
best = [None] * length
for kind, start, end, stamp in ops:
st = tuple(stamp)
for i in range(max(start, 0), min(end, length)):
if best[i] is None or st[0] > best[i][0][0]:
best[i] = (st, kind)
spans = []
for i, b in enumerate(best):
if b is not None and b[1] == 'bold':
if spans and spans[-1][1] == i:
spans[-1][1] = i + 1
else:
spans.append([i, i + 1])
return spans
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
cases = {
1: [('later unbold wins inside a bold range', [9, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [9, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [9, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [9, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [9, [['bold', 6, 9, [1, 'a']]]], [[6, 9]]), ('range end is exclusive', [9, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [9, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [9, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [9, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [9, [['bold', 8, 10, [1, 'a']], ['unbold', -1, 1, [2, 'a']]]], [[8, 9]])],
2: [('later unbold wins inside a bold range', [10, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [10, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [10, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [10, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [10, [['bold', 7, 10, [1, 'a']]]], [[7, 10]]), ('range end is exclusive', [10, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [10, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [10, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [10, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [10, [['bold', 9, 12, [1, 'a']], ['unbold', -2, 1, [2, 'a']]]], [[9, 10]])],
3: [('later unbold wins inside a bold range', [11, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [11, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [11, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [11, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [11, [['bold', 8, 11, [1, 'a']]]], [[8, 11]]), ('range end is exclusive', [11, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [11, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [11, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [11, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [11, [['bold', 10, 14, [1, 'a']], ['unbold', -3, 1, [2, 'a']]]], [[10, 11]])],
4: [('later unbold wins inside a bold range', [12, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [12, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [12, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [12, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [12, [['bold', 9, 12, [1, 'a']]]], [[9, 12]]), ('range end is exclusive', [12, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [12, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [12, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [12, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [12, [['bold', 11, 16, [1, 'a']], ['unbold', -4, 1, [2, 'a']]]], [[11, 12]])],
5: [('later unbold wins inside a bold range', [13, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [13, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [13, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [13, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [13, [['bold', 10, 13, [1, 'a']]]], [[10, 13]]), ('range end is exclusive', [13, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [13, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [13, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [13, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [13, [['bold', 12, 18, [1, 'a']], ['unbold', -5, 1, [2, 'a']]]], [[12, 13]])],
}[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 fixture | Actual | Expected | Outcome |
|---|---|---|---|
| later unbold wins inside a bold range | [[0, 2], [4, 6]] | [[0, 2], [4, 6]] | Passed |
| arrival order does not matter | [[0, 2], [4, 6]] | [[0, 2], [4, 6]] | Passed |
| equal counters break by replica | [[1, 5]] | [[1, 5]] | Passed |
| equal counters other arrival | [[1, 3]] | [[1, 5]] | Failed |
| mark reaching the end of text | [[6, 9]] | [[6, 9]] | Passed |
| range end is exclusive | [[2, 3], [5, 6]] | [[2, 3], [5, 6]] | Passed |
| adjacent ranges join | [[0, 4]] | [[0, 4]] | Passed |
| one-character gap stays split | [[0, 2], [3, 5]] | [[0, 2], [3, 5]] | Passed |
| range starting before the text is clamped | [[0, 2]] | [[0, 2]] | Passed |
| range past the end is clamped | [[8, 9]] | [[8, 9]] | Passed |
SHA-256 / c36a8f083a2fcddaf2b286e81369bb87ab250f302f22ccc3eaa305bb1636ecd1
3 / The verified repair
Exit 0"""Failure Map reference implementation. Python standard library only."""
import json
N = 1
observations = []
def solve(length, ops):
best = [None] * length
for kind, start, end, stamp in ops:
st = tuple(stamp)
for i in range(max(start, 0), min(end, length)):
if best[i] is None or st > best[i][0]:
best[i] = (st, kind)
spans = []
for i, b in enumerate(best):
if b is not None and b[1] == 'bold':
if spans and spans[-1][1] == i:
spans[-1][1] = i + 1
else:
spans.append([i, i + 1])
return spans
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
cases = {
1: [('later unbold wins inside a bold range', [9, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [9, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [9, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [9, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [9, [['bold', 6, 9, [1, 'a']]]], [[6, 9]]), ('range end is exclusive', [9, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [9, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [9, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [9, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [9, [['bold', 8, 10, [1, 'a']], ['unbold', -1, 1, [2, 'a']]]], [[8, 9]])],
2: [('later unbold wins inside a bold range', [10, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [10, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [10, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [10, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [10, [['bold', 7, 10, [1, 'a']]]], [[7, 10]]), ('range end is exclusive', [10, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [10, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [10, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [10, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [10, [['bold', 9, 12, [1, 'a']], ['unbold', -2, 1, [2, 'a']]]], [[9, 10]])],
3: [('later unbold wins inside a bold range', [11, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [11, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [11, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [11, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [11, [['bold', 8, 11, [1, 'a']]]], [[8, 11]]), ('range end is exclusive', [11, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [11, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [11, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [11, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [11, [['bold', 10, 14, [1, 'a']], ['unbold', -3, 1, [2, 'a']]]], [[10, 11]])],
4: [('later unbold wins inside a bold range', [12, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [12, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [12, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [12, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [12, [['bold', 9, 12, [1, 'a']]]], [[9, 12]]), ('range end is exclusive', [12, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [12, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [12, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [12, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [12, [['bold', 11, 16, [1, 'a']], ['unbold', -4, 1, [2, 'a']]]], [[11, 12]])],
5: [('later unbold wins inside a bold range', [13, [['bold', 0, 6, [1, 'a']], ['unbold', 2, 4, [2, 'b']]]], [[0, 2], [4, 6]]), ('arrival order does not matter', [13, [['unbold', 2, 4, [2, 'b']], ['bold', 0, 6, [1, 'a']]]], [[0, 2], [4, 6]]), ('equal counters break by replica', [13, [['bold', 1, 5, [3, 'b']], ['unbold', 3, 7, [3, 'a']]]], [[1, 5]]), ('equal counters other arrival', [13, [['unbold', 3, 7, [3, 'a']], ['bold', 1, 5, [3, 'b']]]], [[1, 5]]), ('mark reaching the end of text', [13, [['bold', 10, 13, [1, 'a']]]], [[10, 13]]), ('range end is exclusive', [13, [['bold', 2, 3, [1, 'a']], ['bold', 5, 6, [2, 'a']]]], [[2, 3], [5, 6]]), ('adjacent ranges join', [13, [['bold', 0, 2, [1, 'a']], ['bold', 2, 4, [1, 'b']]]], [[0, 4]]), ('one-character gap stays split', [13, [['bold', 0, 2, [1, 'a']], ['bold', 3, 5, [1, 'b']]]], [[0, 2], [3, 5]]), ('range starting before the text is clamped', [13, [['bold', -2, 2, [1, 'a']]]], [[0, 2]]), ('range past the end is clamped', [13, [['bold', 12, 18, [1, 'a']], ['unbold', -5, 1, [2, 'a']]]], [[12, 13]])],
}[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 fixture | Actual | Expected | Outcome |
|---|---|---|---|
| later unbold wins inside a bold range | [[0, 2], [4, 6]] | [[0, 2], [4, 6]] | Passed |
| arrival order does not matter | [[0, 2], [4, 6]] | [[0, 2], [4, 6]] | Passed |
| equal counters break by replica | [[1, 5]] | [[1, 5]] | Passed |
| equal counters other arrival | [[1, 5]] | [[1, 5]] | Passed |
| mark reaching the end of text | [[6, 9]] | [[6, 9]] | Passed |
| range end is exclusive | [[2, 3], [5, 6]] | [[2, 3], [5, 6]] | Passed |
| adjacent ranges join | [[0, 4]] | [[0, 4]] | Passed |
| one-character gap stays split | [[0, 2], [3, 5]] | [[0, 2], [3, 5]] | Passed |
| range starting before the text is clamped | [[0, 2]] | [[0, 2]] | Passed |
| range past the end is clamped | [[8, 9]] | [[8, 9]] | Passed |
SHA-256 / 21c8eaa45830efc45a72b6b5b3a268b3a1405e8bbf2c11fe819a0647666b0375
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 2026-09-29T14:49:03.514527+00:00.
Case digest / 98b8bec197e9527f9c560ca63680cf6bcf81d55de1fb6eb16ae0d1603af3942e