fix: flush pending drop cleanup to prevent fast-drag races

[samiodeh1337]

Type

• 🚀 Feature
• 🐛 Bug Fix
• ♻️ Refactor
• 📝 Documentation
• 🧹 Chore (dependencies, CI, tooling)

Description

Related Issue

Closes #

Changes

What

Fixes state corruption when a drop's deferred cleanup (the ~200ms drop-snap animation timeout) overlaps with a new drag or a cancel.

Why

dragEnd schedules finalizeDrop on a setTimeout(DROP_SNAP_MS). Until it ran, a fast second drag or a dragCancel could race it — the previous row stayed hidden or its late cleanup clobbered the new drag.

Changes

• Track the pending finalize + its timeout in refs (pendingFinalizeRef, dropAnimTimeoutRef).
• flushPendingDrop() clears the timeout and runs the finalize synchronously.
• Call flushPendingDrop() at the top of beginDrag (before dragEndFiredRef resets) and in dragCancel.
• Bail out of the post-shift requestAnimationFrame if dragEndFiredRef is already set.
• Bump devDeps: styled-components → 6.4.3, typescript-eslint → 8.57.0, vite → 8.1.3.

Testing

• Drag a row, drop, immediately drag another before the snap settles — no ghost/hidden row.
• Drag then cancel (Esc / pointer leave) — no late finalize fires.
• npm run lint + npm run build pass (CI gates this).

Screenshots / Demos

Checklist

• I have read the Contributing Guide
• My code follows the project's code style (ran npm run format)
• I have run npm run lint with no errors
• I have tested my changes locally
• I have updated the documentation if needed
• I have updated the CHANGELOG.md if applicable
• My changes do not introduce any new warnings