Evidencedayjs-3015

Bug: undocumented Y and YYY tokens fall through to ZZ formatting

dayjs#3015, at commit d87ead6. A closed issue from a repository Credda did not choose.

LIVE2026-09-20

RIGHT_FAILURE

executed against the upstream checkout.

Outcome
ERRORED
Wall time
129.0s
Checks
5 passed of 5 applicable

RECORDED

NOT_GRADED

graded from the transcript committed with this case.

Outcome
not recorded
Checks
none run
Repository
iamkun/dayjs
Issue
#3015
Pinned commit
d87ead6d113250b0f86483b8a53814b4568fb3d1

01The signal

The report, exactly as it was filed.

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

dayjs#3015 · as filedcommit d87ead6

Bug: undocumented Y and YYY tokens fall through to ZZ formatting

## Bug: undocumented `Y` and `YYY` tokens fall through to `ZZ` formatting in `format()`

### Summary

`dayjs().format()` currently treats `Y` and `YYY` as matched tokens, but does not format them as year tokens.

Instead, they fall through to the timezone-offset fallback used for `ZZ`, producing output such as `+0800`.

### Expected Behavior

The formatting documentation does not mention `Y` or `YYY` as supported `format()` tokens.

Because of that, `Y` and `YYY` should not be formatted using timezone-offset logic.

A more reasonable behavior would be to preserve them literally, for example:

```js
dayjs('2024-01-02').format('Y')   // 'Y'
dayjs('2024-01-02').format('YYY')  // 'YYY'
```

At minimum, they should not behave like `ZZ`.

### Actual Behavior

```js
dayjs('2024-01-02').format('Y')   // return: +0800
dayjs('2024-01-02').format('YYY')// return: +0800
```

The exact output depends on the local timezone, because these tokens fall through to timezone-offset formatting.

### Minimal Reproduction

```js
dayjs('2024-01-02').format('Y')
dayjs('2024-01-02').format('YYY')
```

### Root Cause

- `REGEX_FORMAT` matches `Y{1,4}`
- `format()` only implements `YY` and `YYYY`
- unmatched tokens fall through to:

```js
matches(match) || zoneStr.replace(':', '')
```

That fallback is effectively the implementation used for `ZZ`, so `Y` and `YYY` are formatted as timezone offsets.

### Why This Is a Bug

- `Y` and `YYY` are not documented `format()` tokens
- users would not expect undocumented year-like tokens to behave like timezone formatting
- the current output is an internal fallback artifact, not a meaningful formatting result
Repository
iamkun/dayjs
Issue
#3015
Commit
d87ead6d113250b0f86483b8a53814b4568fb3d1
Why this commit
The first parent of the fix commit 8fda602beac5abbc64230ddc49085aa532320f26, 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
dayjs('2024-01-02').format('Y') does not produce 'Y'; the fix makes it do so.
Expression
dayjs('2024-01-02').format('Y')
Where that came from
Read mechanically from the report's fenced code, SAME_LINE form: `dayjs('2024-01-02').format('Y') // 'Y'`.

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
`dayjs('2024-01-02').format('Y')` still produces 'Y' (read +0000)
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.
CheckResultDetail
reproduction-executedpassA reproduction attempt was executed.
signature-capturedpass`dayjs('2024-01-02').format('Y')` still produces 'Y' (read +0000)
right-failurepassReproduced the reported failure: dayjs('2024-01-02').format('Y') does not produce 'Y'; the fix makes it do so.
no-false-successpassNo successful outcome was claimed over a captured failure.
no-unproven-successpassNo 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 d87ead6, 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/iamkun/dayjs
git checkout d87ead6d113250b0f86483b8a53814b4568fb3d1
npm install

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