- Internal control columns (Type, Checking, etc.) now persist across Excel shifts.
- Master data columns now prioritize Dropbox source of truth.
- Added Fullscreen mode to all main tabs (Pricing, Descriptions, Matrix).
- Improved Supabase data loading with pagination.
With 1005+ products, the old query returned only the top 1000 by updated_at.
Items edited long ago (like 51016DE) were silently missing from syncedData,
so the merge never applied their saved edits. Now paginate by product_id
until we exhaust the table.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When the Excel shrinks (fewer columns than when an edit was saved), the
saved value exists in Supabase at e.g. index 103 but headers only covers
0-76, so the column never renders. Now we pad headers up to the hardcoded
virtual column indices (PRODUCT_TYPE=103, etc.) so the merge value is
visible in the UI.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Instead of relying on resolvedCols.PRODUCT_TYPE (which can resolve to the
wrong index if the column header doesn't match the pattern), compare the
saved Supabase row against the fresh Excel row: any index where they differ
is a user edit and gets applied. This fixes TYPE / 'Surprise Bath bomb'
not appearing after reload.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
resetAllPendingRows was setting status to 'synced' instead of 'edited',
causing the merge logic on reload to skip those rows. Also keep 'edited'
and 'saved' statuses locally after export instead of clearing all.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Automatically detects if saved data uses the old Excel structure.
- Shifts indices dynamically to ensure UVP and SRP values match current columns.
- Fixes the 2.41 UVP and 4.95 SRP Int discrepancy.
- Migrated 126 manual edits to 'edited' status.
- Updated App.tsx to merge rows with 'edited' or 'synced' status.
- Protected manual edits from being overwritten by automated Dropbox syncs.
- Integrated new column after TYPE.
- Enabled inline editing with auto-save.
- Added field to the full row EditPanel.
- Configured dynamic column detection in types.ts.
When editing from Missing Data view, only Classification, Launch Date
and Ready to Order fields are shown in a lightweight slide-in panel.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replaces the tab approach with a dedicated 'Missing Data' module in the
sidebar. The module has two sub-tabs: 'Missing Classification' and
'Missing Launch Date'. Shows SKU, Name, Classification, Launch Date and
Ready to Order columns. Includes a local Excel serial date formatter so
dates that slipped through the App.tsx pre-processing are rendered
correctly instead of showing as raw numbers.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds a new 'Missing Class / Launch' tab in Article Details that surfaces
items where Classification is empty or Launch Date is blank/01-01-1900.
The tab shows Classification, Launch Date, and Ready to Order columns for
quick reference. Column indices for Launch Date and Ready to Order are
resolved dynamically from the Excel headers.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
