Sign in

Evidenceyargs-parser-118

odd behavior of narg:2 versus duplicate-arguments-array:false

yargs-parser#118, at commit 57b7883. A closed issue from a repository Credda did not choose.

LIVE2026-09-20

WRONG_FAILURE

executed against the upstream checkout.

Outcome
INCONCLUSIVE
Wall time
29.6s
Checks
4 passed of 5 applicable

RECORDED

NOT_GRADED

graded from the transcript committed with this case.

Outcome
not recorded
Checks
none run
Repository
yargs/yargs-parser
Issue
#118
Pinned commit
57b788303e1119ba5260b273e486b73d0d638ad8

01The signal

The report, exactly as it was filed.

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

yargs-parser#118 · as filedcommit 57b7883

odd behavior of narg:2 versus duplicate-arguments-array:false

```
const P = require('yargs-parser')
P('foo -p x y', { narg: { p: 2 } })
{ _: [ 'foo' ], p: [ 'x', 'y' ] }
```

i.e. this results in the expected `{x:'y'}` binding.

whereas:
```
P('foo -p x y', { narg: { p: 2 }, configuration: {  'duplicate-arguments-array': false } })
{ _: [ 'foo' ], p: 'y' }
```

which seems buggy, and contrary to the documentation. i would expect the resulting binding to be the same in both cases.
Issue
#118
Commit
57b788303e1119ba5260b273e486b73d0d638ad8
Why this commit
The first parent of the fix commit 47ccb0b7fcca1b989ef97bb084e4aa2aaf2a7666, 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
P('foo -p x y', { narg: { p: 2 }, configuration: { 'duplicate-arguments-array': false } }) produces { _: [ 'foo' ], p: 'y' }; the fix makes it produce { _: [ 'foo' ], p: [ 'x', 'y' ] }.
Expression
P('foo -p x y', { narg: { p: 2 }, configuration: { 'duplicate-arguments-array': false } })
Reported output
{ _: [ 'foo' ], p: 'y' }
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: `P('foo -p x y', { narg: { p: 2 }, configuration: { 'duplicate-arguments-array': false } }) //=> { _: [ 'foo' ], p: 'y' }`. 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

A real failure was captured. It was the wrong one.

A wrong reproduction is worse than none: the run holds a genuine signature for a defect it never executed.

captured failure signatureLIVE · normalized
`P('foo -p x y', { narg: { p: 2 } })` still produces { _: [ 'foo' ], p: [ 'x', 'y' ] } (read { _: [ 'foo' ], p: [ 'x', 'y' ] })
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.

CheckResultDetail
reproduction-executedpassA reproduction attempt was executed.
signature-capturedpass`P('foo -p x y', { narg: { p: 2 } })` still produces { _: [ 'foo' ], p: [ 'x', 'y' ] } (read { _: [ 'foo' ], p: [ 'x', 'y' ] })
right-failurefailExpected `P('foo -p x y', { narg: { p: 2 }, configuration: { 'duplicate-arguments-array': false } })` still producing { _: [ 'foo' ], p: 'y' }.
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 57b7883, 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/yargs/yargs-parser
git checkout 57b788303e1119ba5260b273e486b73d0d638ad8
npm install

CREDDA_PROVIDER=heuristic \
  npx tsx apps/cli/src/main.ts fix <repo-path> @<issue-file> --no-color
Read the full studyAll 158 cases