fix(shared): preserve final quoted empty CSV records

[Lucenx9]

What changed

Keep the final quoted empty record in CSV and TSV table previews when the file has no trailing line ending. The parser now tracks the start of the source row instead of using decoded cell content to decide whether a final record exists.

Why

For name\r\nAlice\r\n"", the preview drops the last row and reports complete content. Adding a final CRLF makes that same row appear. A final record without a line ending is valid CSV, including a quoted empty field. See RFC 4180, section 2.

flowchart LR
    A[Final quoted empty record without a line ending] --> B[Before: empty decoded buffers]
    B --> C[Row omitted]
    A --> D[After: source row still present]
    D --> E[Row retained]

Loading

This changes only the shared parser and its focused tests. The web and desktop attachment and workspace table previews use it. Mobile does not currently render this table component. File contents, source view, wire contracts, product defaults, and preview limits are unchanged.

Verification

• The six new regression tests fail on the base commit. All nine tests pass with the fix.
• vp test run packages/shared/src/delimitedPreview.test.ts --maxWorkers=1 --no-file-parallelism
• vp lint packages/shared/src/delimitedPreview.ts packages/shared/src/delimitedPreview.test.ts
• vp fmt --check packages/shared/src/delimitedPreview.ts packages/shared/src/delimitedPreview.test.ts
• tsc --noEmit -p packages/shared/tsconfig.json passes. Existing Effect suggestions remain in unrelated files.
• 504 generated round-trip cases pass across both delimiters, LF/CRLF/CR, BOM, quoted line breaks, empty fields, and the 100-row limit.
• Fallow review scoped to packages/shared, one thread: low risk, no new APIs, dependencies, or diagnostic suppressions. No structural decisions flagged. Consumers outside that package were checked in source.

UI changes

Captured in the running T3 web app with Chromium at 2× device scale. The same draft attachment contains Participant\r\nAlice\r\nBob\r\n"", without a final newline. The before capture uses the parser from b1e223e2b0; the after capture uses caaf8daa44. No styles or components were changed for the captures.

Before: two data rows. The final empty record is missing.

After: three data rows. The empty row below Bob is retained.

Source view showing the quoted empty fourth line. Switching back to the table preserves all three data rows.

Checklist

• This PR is small and focused
• I explained what changed and why
• I included before/after screenshots for any UI changes
• No animation or timing changes, so a video is not applicable

Model: GPT-5 family. Harness: Codex.

Summary by CodeRabbit

• Bug Fixes

  • Improved delimited data previews for BOM-prefixed input, empty records, trailing line endings, and quoted empty values.
  • Corrected row-limit and truncation handling to ensure previews include the appropriate remaining data.

• Tests

  • Added coverage for empty input, BOM handling, row limits, trailing line endings, quoted empty records, and truncation behavior.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-12T14:08:58.185780Z	caaf8da	PR opened

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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[macroscopeapp]

Approvability

Verdict: Approved at caaf8da

Macroscope's review found this PR approvable — This focused parser fix preserves final quoted empty CSV/TSV records without changing parsing limits or introducing new capabilities. Regression tests cover the affected edge cases and confirm that empty input and trailing line endings remain unchanged.

^{You can add or adjust custom eligibility rules. Learn more.}

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Advanced

Run ID: 9da9efea-e684-4246-a27d-30b238a61656

📥 Commits

Reviewing files that changed from the base of the PR and between b1e223e and caaf8da.

📒 Files selected for processing (2)

• packages/shared/src/delimitedPreview.test.ts
• packages/shared/src/delimitedPreview.ts

Included review availability: Your plan provides up to 10 included reviews per hour; 8 remain after this review.

---

📝 Walkthrough

Walkthrough

parseDelimitedPreview now tracks the start of each row, including BOM-prefixed input, and uses remaining text to decide whether to emit a trailing row. Tests cover empty records, line endings, delimiters, BOM input, row limits, and truncation.

Changes

Delimited preview parsing

Layer / File(s)	Summary
Row tracking and trailing-row validation packages/shared/src/delimitedPreview.ts, packages/shared/src/delimitedPreview.test.ts	parseDelimitedPreview tracks row boundaries and emits trailing rows only when input remains. Tests cover comma and tab delimiters, quoted empty records, empty input, line endings, BOM input, row limits, and truncation.

Priority: ⬇️ Low

Estimated code review effort: 2 (Simple) | ~10 minutes

Change: Bug fix

Suggested reviewers: chrisdeeming

Merge Risk: ⚪ Minimal · up to 4e262

The parser change preserves final quoted empty records while retaining empty-input and trailing-line-ending behavior. No merge-blocking risk remains.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the main fix: preserving final quoted empty CSV records. It is concise and specific.
Description check	✅ Passed	The description explains what changed, why the fix is needed, affected consumers, verification steps, and UI evidence. It covers the required template sections and confirms that the change is focused.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 1 functions across 2 files.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches 💡 1

🛠️ Fix failing CI checks 💡

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

_{Comment @coderabbitai help to get the list of available commands.}

[greatitself]

Independent validation of caaf8daa448d3eb4a9b45d9eb97d25363cff020c (merge-base b1e223e2b0d8).

Generated 2,400 valid CSV/TSV inputs with an independent serializer: 1–100 rows, LF/CRLF/CR separators, optional final separator and BOM, and quoted cells containing delimiters, quotes, newlines, and Unicode. Every input ends with an explicitly quoted empty record. All 2,400 matched the expected table on this revision; the merge-base lost the final row in 1,200 cases.

Reproduction scripts and recorded results — first.mjs; the numbered metadata file pins this PR’s tested revisions.

Model: GPT-6. Harness: Codex.