Evidenceluxon-882

Duration fromISO / toISO don't handle correctly negative second/millisecond components

luxon#882, at commit c34afb1. A closed issue from a repository Credda did not choose.

LIVE2026-09-20

RIGHT_FAILURE

executed against the upstream checkout.

Outcome: INCONCLUSIVE
Wall time: 86.2s
Checks: 5 passed of 5 applicable

RECORDED

NOT_GRADED

graded from the transcript committed with this case.

Outcome: not recorded
Checks: none run

Repository: moment/luxon
Issue: #882
Pinned commit: c34afb160da1b330a0c99a1197adb9fd4763961a

01The signal

The report, exactly as it was filed.

Nothing paraphrased or cleaned up. The mess is the thing under test.

luxon#882 · as filedcommit c34afb1

Duration fromISO / toISO don't handle correctly negative second/millisecond components

**Describe the bug**
#### case A
Parsing is good but Formating is wrong.
```js
const duration = luxon.Duration.fromISO("PT-1.5S")
duration.toObject()
=> Object { seconds: -1, milliseconds: 500 }
duration.toISO()
=> "PT-0.5S"
```

#### case B
Parsing is wrong
```js
const duration = luxon.Duration.fromISO("PT-0.5S")
duration.toObject()
=> Object { seconds: -0, milliseconds: 500 }
duration.toISO()
=> "PT0.5S"
```

**Actual vs Expected behavior**
#### case A
expected
```js
luxon.Duration.fromISO("PT-1.5S").toISO()
=> "PT-1.5S"
```

#### case B
expected
```js
luxon.Duration.fromISO("PT-0.5S").toObject()
=> Object { seconds: 0, milliseconds: -500 }
```

Repository: moment/luxon
Issue: #882
Commit: c34afb160da1b330a0c99a1197adb9fd4763961a
Why this commit: The first parent of the fix commit da04179fb95fb8c6b69e82d3734cea15a33bc2f5, which GitHub binds to this issue via CLOSED_EVENT_PR. Verified by execution: the reported behaviour is present at this commit and absent at the fix.
How the text was obtained: Fetched verbatim via the GitHub GraphQL API. Title on the first line, body unmodified below it. Nothing was paraphrased, cleaned up, or supplemented.
Toolchain: javascript · node · unknown · npm

02What counts as reproducing it

The bar, written down before the run.

expected.reportedFailurecommitted with the case

Symptom: luxon.Duration.fromISO("PT-1.5S").toISO() produces 'PT-0.5S'; the fix makes it produce 'PT-1.5S'.
Expression: luxon.Duration.fromISO("PT-1.5S").toISO()
Reported output: "PT-0.5S"
Where that came from: Proposed by a model reading this report and nothing else -- it never saw the repository or the fix commit -- and read back as a claim by the same parser the harvest uses, SAME_LINE form: `luxon.Duration.fromISO("PT-1.5S").toISO() //=> "PT-0.5S"`. The report sat in the CODE_NO_ANNOTATION_AT_ALL bucket, which no regex reaches. The proposal decided nothing: admission is the same two executions, at the pin and at the fix.

03What happened

The live run reproduced the reported failure.

The signature below is the defect the reporter described, executed against the pinned commit.

captured failure signatureLIVE · normalized

`duration.toISO()` still produces "PT-0.5S" (read PT-0.5S)

bench external · checks · LIVE5 checks · 2026-09-20

The LIVE grading as emitted. A check that did not apply is never shown as a pass.

Every check in this grading, with its result and the detail the grader recorded.
Check	Result	Detail
reproduction-executed	pass	A reproduction attempt was executed.
signature-captured	pass	`duration.toISO()` still produces "PT-0.5S" (read PT-0.5S)
right-failure	pass	Reproduced the reported failure: luxon.Duration.fromISO("PT-1.5S").toISO() produces 'PT-0.5S'; the fix makes it produce 'PT-1.5S'.
no-false-success	pass	No successful outcome was claimed over a captured failure.
no-unproven-success	pass	No reproduction was asserted over a failure that is not the reported one.

bench/external/scorecard.json, the run of 2026-09-20 against all 158 upstream checkouts.

The same case, graded from the transcript recorded .

The grading the benchmark gate runs on. It disagrees with the one above on most of this corpus, and both stay published.

Check it yourself

Everything here is downstream of a public commit.

Clone it, check out c34afb1, run the report through the CLI the way the study did.

How the study invoked itone isolated home per case

git clone https://github.com/moment/luxon
git checkout c34afb160da1b330a0c99a1197adb9fd4763961a
npm install

CREDDA_PROVIDER=heuristic \
  npx tsx apps/cli/src/main.ts fix <repo-path> @<issue-file> --no-color

Read the full study All 158 cases