FA-48071 / Delimited text / Open access
Byte-span advancement forgets the record separator · case 01
A structured table violates the declared record or column contract.
ROOT CAUSE
The next span begins at the previous payload end rather than after LF.
VERIFIED REPAIR
Preserve the named invariant at the faulty decision: start += size + 1
Unsuccessful approach: Assuming CRLF overcounts a one-byte delimiter.
Case contract
For UTF-8 text with LF-delimited physical records, report each record as [text,start_byte,end_byte,line_number]. LF is excluded from text and end offset; a final LF does not add a phantom record. Empty interior records remain. Leading BOM is ordinary text here.
Why this case matters
Delimited interchange needs explicit framing, schema and field semantics at ingestion and emission boundaries.
1 / The failure
Exit 1"""Failure Map reference implementation. Python standard library only."""
import json
def _vary(value):
if value == '@END': return 3 + 4*N
if isinstance(value, str): return value.replace('@', 'cell' * N)
if isinstance(value, list): return [_vary(x) for x in value]
if isinstance(value, dict): return {_vary(k): _vary(v) for k,v in value.items()}
return value
N = 1
observations = []
def solve(data):
out, start, line = [], 0, 1
parts = data.split('\n')
if parts[-1] == '': parts.pop()
for text in parts:
size = len(text.encode('utf-8'))
out.append([text, start, start + size, line])
start += size
line += 1
return out
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('unicode', solve(_vary('é\n@')), _vary([['é', 0, 2, 1], ['@', 3, '@END', 2]]))
check('empty middle', solve(_vary('a\n\nb')), _vary([['a', 0, 1, 1], ['', 2, 2, 2], ['b', 3, 4, 3]]))
check('final LF', solve(_vary('a\n')), _vary([['a', 0, 1, 1]]))
check('empty', solve(_vary('')), _vary([]))
check('only LF', solve(_vary('\n')), _vary([['', 0, 0, 1]]))
check('astral', solve(_vary('😀\nz')), _vary([['😀', 0, 4, 1], ['z', 5, 6, 2]]))
check('no LF', solve(_vary('x')), _vary([['x', 0, 1, 1]]))
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 |
|---|---|---|---|
| unicode | [['é', 0, 2, 1], ['cell', 2, 6, 2]] | [['é', 0, 2, 1], ['cell', 3, 7, 2]] | Failed |
| empty middle | [['a', 0, 1, 1], ['', 1, 1, 2], ['b', 1, 2, 3]] | [['a', 0, 1, 1], ['', 2, 2, 2], ['b', 3, 4, 3]] | Failed |
| final LF | [['a', 0, 1, 1]] | [['a', 0, 1, 1]] | Passed |
| empty | [] | [] | Passed |
| only LF | [['', 0, 0, 1]] | [['', 0, 0, 1]] | Passed |
| astral | [['😀', 0, 4, 1], ['z', 4, 5, 2]] | [['😀', 0, 4, 1], ['z', 5, 6, 2]] | Failed |
| no LF | [['x', 0, 1, 1]] | [['x', 0, 1, 1]] | Passed |
SHA-256 / b5a39d17f070b3942d648e6e3ba6d98da8667788c304533a37606c1af8b21e63
2 / The unsuccessful fix
Exit 1"""Failure Map reference implementation. Python standard library only."""
import json
def _vary(value):
if value == '@END': return 3 + 4*N
if isinstance(value, str): return value.replace('@', 'cell' * N)
if isinstance(value, list): return [_vary(x) for x in value]
if isinstance(value, dict): return {_vary(k): _vary(v) for k,v in value.items()}
return value
N = 1
observations = []
def solve(data):
out, start, line = [], 0, 1
parts = data.split('\n')
if parts[-1] == '': parts.pop()
for text in parts:
size = len(text.encode('utf-8'))
out.append([text, start, start + size, line])
start += size + 2
line += 1
return out
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('unicode', solve(_vary('é\n@')), _vary([['é', 0, 2, 1], ['@', 3, '@END', 2]]))
check('empty middle', solve(_vary('a\n\nb')), _vary([['a', 0, 1, 1], ['', 2, 2, 2], ['b', 3, 4, 3]]))
check('final LF', solve(_vary('a\n')), _vary([['a', 0, 1, 1]]))
check('empty', solve(_vary('')), _vary([]))
check('only LF', solve(_vary('\n')), _vary([['', 0, 0, 1]]))
check('astral', solve(_vary('😀\nz')), _vary([['😀', 0, 4, 1], ['z', 5, 6, 2]]))
check('no LF', solve(_vary('x')), _vary([['x', 0, 1, 1]]))
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 |
|---|---|---|---|
| unicode | [['é', 0, 2, 1], ['cell', 4, 8, 2]] | [['é', 0, 2, 1], ['cell', 3, 7, 2]] | Failed |
| empty middle | [['a', 0, 1, 1], ['', 3, 3, 2], ['b', 5, 6, 3]] | [['a', 0, 1, 1], ['', 2, 2, 2], ['b', 3, 4, 3]] | Failed |
| final LF | [['a', 0, 1, 1]] | [['a', 0, 1, 1]] | Passed |
| empty | [] | [] | Passed |
| only LF | [['', 0, 0, 1]] | [['', 0, 0, 1]] | Passed |
| astral | [['😀', 0, 4, 1], ['z', 6, 7, 2]] | [['😀', 0, 4, 1], ['z', 5, 6, 2]] | Failed |
| no LF | [['x', 0, 1, 1]] | [['x', 0, 1, 1]] | Passed |
SHA-256 / 3b84822ccb7702a6ca29c9ad67023f2d42b0f7061d59c1ebb570f538d75cdbb5
3 / The verified repair
Exit 0"""Failure Map reference implementation. Python standard library only."""
import json
def _vary(value):
if value == '@END': return 3 + 4*N
if isinstance(value, str): return value.replace('@', 'cell' * N)
if isinstance(value, list): return [_vary(x) for x in value]
if isinstance(value, dict): return {_vary(k): _vary(v) for k,v in value.items()}
return value
N = 1
observations = []
def solve(data):
out, start, line = [], 0, 1
parts = data.split('\n')
if parts[-1] == '': parts.pop()
for text in parts:
size = len(text.encode('utf-8'))
out.append([text, start, start + size, line])
start += size + 1
line += 1
return out
def check(label, actual, expected):
observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('unicode', solve(_vary('é\n@')), _vary([['é', 0, 2, 1], ['@', 3, '@END', 2]]))
check('empty middle', solve(_vary('a\n\nb')), _vary([['a', 0, 1, 1], ['', 2, 2, 2], ['b', 3, 4, 3]]))
check('final LF', solve(_vary('a\n')), _vary([['a', 0, 1, 1]]))
check('empty', solve(_vary('')), _vary([]))
check('only LF', solve(_vary('\n')), _vary([['', 0, 0, 1]]))
check('astral', solve(_vary('😀\nz')), _vary([['😀', 0, 4, 1], ['z', 5, 6, 2]]))
check('no LF', solve(_vary('x')), _vary([['x', 0, 1, 1]]))
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 |
|---|---|---|---|
| unicode | [['é', 0, 2, 1], ['cell', 3, 7, 2]] | [['é', 0, 2, 1], ['cell', 3, 7, 2]] | Passed |
| empty middle | [['a', 0, 1, 1], ['', 2, 2, 2], ['b', 3, 4, 3]] | [['a', 0, 1, 1], ['', 2, 2, 2], ['b', 3, 4, 3]] | Passed |
| final LF | [['a', 0, 1, 1]] | [['a', 0, 1, 1]] | Passed |
| empty | [] | [] | Passed |
| only LF | [['', 0, 0, 1]] | [['', 0, 0, 1]] | Passed |
| astral | [['😀', 0, 4, 1], ['z', 5, 6, 2]] | [['😀', 0, 4, 1], ['z', 5, 6, 2]] | Passed |
| no LF | [['x', 0, 1, 1]] | [['x', 0, 1, 1]] | Passed |
SHA-256 / 5a6882ff32edfacc4f2d66200eb1f89df8b3f59bdc8b43ded53608af400d1d63
Verification & scope
Deterministic bounded in-memory model. No claim of complete CSV or external format conformance. 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 2026-09-29T14:44:47.455307+00:00.
Case digest / 5ec270dfd9f98f8800e24f90798f572e5c3ac2b6c789902b606b09f884d67952