Refactor Store sorting to use symbol-based initialIndex strategy

This epic refactors the Store sorting mechanism to be more memory-efficient and robust by eliminating the duplication of data into initialData and instead using a lightweight, symbol-based index on records.

Problem

1. Crash on Reset: Store.sort({direction: null}) caused a crash (TypeError: me.initialData is not iterable) when initialData was null or not synced.
2. Memory Overhead: The previous approach duplicated the entire dataset into initialData solely for restoring sort order, doubling memory usage for large stores.
3. Synchronization Issues: Keeping initialData in sync with dynamic add/remove/filter operations was complex and error-prone.
4. UI Updates: The Grid failed to refresh the view when sorting was reset to the initial state.

Solution Architecture

1. Symbol-Based Indexing (initialIndexSymbol)

Instead of storing a separate array of data, we now tag each record with its insertion index using a non-enumerable symbol property.

• src/collection/Base.mjs: Introduced initialIndexCounter. The splice method now automatically assigns a unique, incrementing index to every new item using Symbol.for('initialIndex').
• src/data/RecordFactory.mjs: Updated the generated Record class to include a field for [initialIndexSymbol]. The constructor ensures this index is preserved when converting raw data objects into Record instances.

2. Streamlined Store Sorting

• src/data/Store.mjs:
  • Removed initialData config and all related synchronization logic (beforeSetInitialData, afterSetData).
  • Updated sort() to restore the natural order by sorting on initialIndexSymbol instead of clearing and reloading data.
  • Enabled me.fire('load') in onCollectionSort to ensure explicit sort operations trigger UI updates.

3. Optimized Event Flow

• src/collection/Base.mjs: Modified splice to perform sorting silently (silent=true). This prevents redundant sort events (and subsequent double-reloads) during data mutations, allowing Store to listen to sort events strictly for user-initiated sorting.

4. Grid Container Updates

• src/grid/Container.mjs: Updated onSortColumn to ensure the UI refreshes correctly even when the sort direction is null (reset), fixing the specific UI regression.

Benefits

• Performance: Reduced memory footprint by removing data duplication.
• Stability: Eliminated the crash vector associated with initialData syncing.
• Maintainability: Centralized index logic in Collection.Base, making it available to all collections, not just Stores.
• Correctness: Sorting reset now reliably updates the UI.