Evidencewrap-ansi-39

Newline handling in 6.0.0

wrap-ansi#39, at commit 7bcd854. A closed issue from a repository Credda did not choose.

LIVE2026-09-20

NO_FAILURE_OBSERVED

executed against the upstream checkout.

Outcome: INCONCLUSIVE
Wall time: 64.0s
Checks: 3 passed of 5 applicable

RECORDED

NOT_GRADED

graded from the transcript committed with this case.

Outcome: not recorded
Checks: none run

Repository: chalk/wrap-ansi
Issue: #39
Pinned commit: 7bcd854fb5ca90e5e8c4963a1398d46a306a4a66

01The signal

The report, exactly as it was filed.

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

wrap-ansi#39 · as filedcommit 7bcd854

Newline handling in 6.0.0

In `wrap-ansi@5.0.0` all newlines are converted properly:

```js
const wrapAnsi = require("wrap-ansi@5.0.0")
console.log(wrapAnsi('a\r\n\r\nb', 5)) // outputs "a\n\nb"
```

In `wrap-ansi@6.0.0` not all newlines are converted:

```js
const wrapAnsi = require("wrap-ansi@6.0.0")
console.log(wrapAnsi('a\r\n\r\nb', 5)) // outputs "a\r\n\nb"
```

Repository: chalk/wrap-ansi
Issue: #39
Commit: 7bcd854fb5ca90e5e8c4963a1398d46a306a4a66
Why this commit: The first parent of the fix commit adbac70bcac898db1ec9668c436c8c9af0b8f476, 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: wrapAnsi('a\r\n\r\nb', 5) does not produce "a\n\nb"; the fix makes it do so.
Expression: wrapAnsi('a\r\n\r\nb', 5)
Where that came from: Read mechanically from the report's fenced code, SAME_LINE form: `console.log(wrapAnsi('a\r\n\r\nb', 5)) // outputs "a\n\nb"`.

03What happened

No failure was captured.

Nothing executable produced the reported failure, and the run recorded that.

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	fail	The reproduction ran and demonstrated no failure.
right-failure	fail	Expected `wrapAnsi('a\r\n\r\nb', 5)` not producing "a\n\nb".
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 7bcd854, 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/chalk/wrap-ansi
git checkout 7bcd854fb5ca90e5e8c4963a1398d46a306a4a66
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