[daggerok]

What

• New maxRowLength(rows, min = 0) loop helper
• extractValidTableData: both Math.max(...rows.map(r => r.length)) calls use it
• Merged export: maxCols is a reduce over blocks with maxRowLength(block.rows) instead of nested spreads
• applyFilters uses the same helper instead of its own reduce (same result)

Why

CSV-H1: spreading one argument per row into Math.max throws RangeError: Maximum call stack size exceeded in V8 (Chrome, Edge, Node). A big export failed to load at all, and the merged export failed for a large block

Checks

• bun run build
• tsc --noEmit --strict --skipLibCheck --jsx react-jsx on src/main.tsx (no errors)
• Old and new parseCSVText + extractValidTableData transpiled with Bun.Transpiler and run in Node (V8):
  • old: 100k rows ok; 200k rows, 1M rows and a 300k single-column file throw Maximum call stack size exceeded
  • new: 100k, 200k, 1M rows and 300k single-column all parse
• Same output from old and new on a Fidelity file (trailing commas plus footer), a preamble, ragged rows, a single column and quoted fields
• Merged export width: old and new formula equal on empty, header-only, mixed-width and single-row inputs; old throws on a 300k-row block, new returns 2
• git diff --check
• No test file: the repo has no test setup (README "No tests")