fix: restore legacy path compatibility regressions

[maxatwork]

Summary

• restore legacy Rails-style path parsing for nested object and indexed field names in the core parser
• fix the DOM radio coercion edge case and Rails-style field matching in js2form
• add regression coverage for the restored behaviors and nested DOM extraction

Fixed Issues

Fixes #74
Fixes #68
Fixes #41
Fixes #39
Fixes #22
Fixes #19

Test Plan

• npm test

---

Summary by cubic

Restore legacy Rails-style bracket parsing and name normalization in packages/core and packages/js2form, fix radio value coercion in packages/dom, and keep sibling Rails array fields on the same synthetic index. Resolves nested object/array regressions and adds tests to lock behavior.

• Bug Fixes
  • packages/core: Parses Rails-style keys with underscores, one-character names, root [prop], and mixed/consecutive indexed arrays.
  • packages/dom: Radio inputs now coerce to true/false/value/null correctly; an empty checked radio isn’t forced to false when true/false siblings exist.
  • packages/js2form: Normalizes Rails-style field names so inputs like a[b][c] populate, keeps items[][title] and items[][description] on the same item index, and preserves literal suffixes around bracketed indexes (e.g., items[5]_id).
  • Added regression tests for restored behaviors, nested DOM extraction, and suffix preservation.

^{Written for commit 95e4d1a. Summary will update on new commits.}

[cubic-dev-ai]

2 issues found across 6 files

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="packages/core/src/index.ts">

<violation number="1" location="packages/core/src/index.ts:94">
P1: Consecutive indexed segments are merged into one token, which breaks nested indexed path parsing (e.g. `foo[0][1][bar]`).</violation>
</file>

<file name="packages/js2form/src/index.ts">

<violation number="1" location="packages/js2form/src/index.ts:105">
P2: The new bracket normalization drops literal text outside bracket groups (e.g. suffixes), which can generate incorrect field names and break form field matching.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.}

[maxatwork]

cubic found two valid issues in the previous patch, and both are fixed in 73e77e3.

• Consecutive indexed segments now split into separate path tokens in the core parser, so nested indexed paths like foo[0][1][bar] preserve the intermediate array level.
• The js2form bracket normalization now preserves literal text around bracket groups instead of dropping suffixes while normalizing field names.

Added regression coverage for both cases and re-ran npm test.

@cubic-dev-ai re-run review

[cubic-dev-ai]

cubic found two valid issues in the previous patch, and both are fixed in 73e77e3.

• Consecutive indexed segments now split into separate path tokens in the core parser, so nested indexed paths like foo[0][1][bar] preserve the intermediate array level.
• The js2form bracket normalization now preserves literal text around bracket groups instead of dropping suffixes while normalizing field names.

...

@maxatwork I have started the AI code review. It will take a few minutes to complete.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: ac2848f90b

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Keep sibling [] Rails fields on the same array item

normalizeName() now routes Rails names like items[][title] and items[][description] through the empty-index array path, but each occurrence gets a fresh synthetic index. In mapFieldsByName() this produces items[0].title and items[1].description, so objectToForm() only fills one field for { items: [{ title, description }] } and leaves the other blank. This breaks a common Rails nested-form shape whenever array object fields omit explicit numeric indexes.

Useful? React with 👍 / 👎.

[maxatwork]

Valid issue. Fixed in 95e4d1a by keeping sibling Rails empty-index fields on the same synthetic array item until a sibling path repeats, so items[][title] and items[][description] map to the same object slot. Added a regression test for that shape in packages/js2form/test/js2form.test.ts and re-ran npm test.

[cubic-dev-ai]

No issues found across 6 files