Skip to content

Preserve review CSV identities and fullwidth formula protection - #8

Merged
dev-belly merged 1 commit into
mainfrom
codex/review-csv-identities
Oct 24, 2026
Merged

dev-belly merged 1 commit into
mainfrom
codex/review-csv-identities

Conversation

@dev-belly

Copy link
Copy Markdown
Owner

The review queue and high-risk extract use an apostrophe sanitizer that misses fullwidth formula prefixes; LF-only CSV writing also leaves lone CR characters unquoted. Use one review-only serializer with visible, collision-safe text: prefixes, protected headers and quoting for CR/LF/CRLF while keeping LF record endings. Numeric values, missing values and raw internal data files retain their existing behavior.

The handoff regression exports a fullwidth/multiline voucher ID and completes the existing outcome-capture path without changing the in-memory ID. The utility cases cover formula/control/BOM prefixes, prefix collisions, embedded quotes/newlines, numeric/missing values and an empty-table header. Existing ASCII-formula cases now assert the visible policy.

Validation: 508 full tests pass (5 new cases), including dashboard renders; Ruff and diff checks pass. Re-generating both default review exports leaves every committed report byte and queue checksum unchanged. Every uploaded Git blob hash matches its local file. Existing original CSV/selection-record pairs remain valid under their retained bytes.

@dev-belly
dev-belly merged commit 0f12cfb into main Oct 24, 2026
3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant