fix(rulesets): duplicated-entry-in-enum - support null values

[DougEdey-Slice]

Fixes #2959.

This handles null enum values, which is supported/required by OAS3.1 for nullable enums

This updates the tests & the ruleset to support nulls, which OAS3.0 does support: https://swagger.io/docs/specification/v3_0/data-models/enums/#nullable-enums

First time contributor so let me know if I need anything else :)

Checklist

• Tests added / updated
• Docs added / updated

Does this PR introduce a breaking change?

• No

[slegarraga]

Nice follow-up to the enum edge cases — null in enum arrays is easy to miss. From #2934 land: worth confirming resolved: false behavior on $ref-wrapped enums matches your test cases. Thanks for pushing this.

[slegarraga]

LGTM — null enum handling matches OAS 3.0 nullable enums and 3.1 semantics from #2959.

• Test matrix covers null in enum arrays without false positives on valid enums.
• validators.ts tweak is minimal and scoped to the duplicated-entry rule path.

No concerns from my side; thanks for the fix.

[slegarraga]

LGTM — null enum handling matches #2959; tests cover nullable enums without false positives. Thanks for the fix.

[slegarraga]

@DougEdey-Slice Thanks — this addresses the regression from #2934 (my merge). Suggest adding a test case for issue #2959 reproducer (components.examples.*.value with foo: null) so we don’t regress the JSONPath guard again. Approved from my side.

[slegarraga]

Opened #2963 with the #2959 reproducer test + same guard — happy to close #2963 if you prefer landing this PR instead.

[slegarraga]

@tomek-tursa-sb Maintainer review when possible — 2 contributor approvals, fixes #2959 (production crash). Alternative: #2963 has the same guard + explicit reproducer test from the issue. Either PR unblocks users on spectral-cli 6.16.x.

[slegarraga]

Reviewed this against #2959 and my own #2963, since they overlap.

The 3.1 coverage here is the right call: null is a valid enum member in OAS 3.1 (enum: [1, 2, 3, null]), and flagging duplicated nulls ([null, …, null] → items #0 and #6) is a real gap the original rule missed. Nice.

One coordination note: this and #2963 fix different facets of #2959 and will conflict on oas/index.ts and duplicated-entry-in-enum.test.ts. #2963 adds @ != null to the rule's given selector to stop a TypeError when the selector hits a null node (e.g. foo: null inside an example value, where reading @.enum throws). That null-node crash isn't exercised by the cases here (these are all schema enums), so I think both changes are needed: the selector guard to avoid the crash, plus your 3.1 null-entry semantics. Happy to rebase #2963 on top of this so it's a single clean change if that's easier to merge.