feat: native update merge (#14250)

#### PR Dependency Tree


* **PR #14250** 👈

This tree was auto-generated by
[Charcoal](https://github.com/danerwilliams/charcoal)

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **Backend Optimization**
  * Faster document retrieval via a native binary fetch path.
* Native-accelerated merging of document updates for improved
performance and consistency.
* **Indexing & Reliability**
* Indexing now only proceeds on valid parse results, with clearer
warnings and richer metadata on failures.
* More consistent sync behavior and enhanced diagnostic logging for
indexing operations.
* **Tests**
  * Expanded tests to cover native binary retrieval error handling.

<sub>✏️ Tip: You can customize this high-level summary in your review
settings.</sub>
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

packages/backend/server/src/core/doc/__tests__/reader-from-rpc.spec.ts

@@ -100,6 +100,9 @@ test('should throw error when doc service internal error', async t => {
   mock.method(adapter, 'getDoc', async () => {
     throw new Error('mock doc service internal error');
   });
+  mock.method(adapter, 'getDocBinNative', async () => {
+    throw new Error('mock doc service internal error');
+  });
   let err = await t.throwsAsync(docReader.getDoc(workspace.id, docId), {
     instanceOf: UserFriendlyError,
     message: 'An internal error occurred.',

packages/backend/server/src/core/doc/reader.ts

@@ -213,11 +213,9 @@ export class DatabaseDocReader extends DocReader {
     guid: string,
     fullContent?: boolean
   ): Promise<PageDocContent | null> {
-    const docRecord = await this.workspace.getDoc(workspaceId, guid);
-    if (!docRecord) {
-      return null;
-    }
-    return this.parseDocContent(docRecord.bin, fullContent ? -1 : 150);
+    const docBinary = await this.workspace.getDocBinNative(workspaceId, guid);
+    if (!docBinary) return null;
+    return this.parseDocContent(docBinary, fullContent ? -1 : 150);
   }
 
   protected override async getWorkspaceContentWithoutCache(

packages/backend/server/src/core/doc/storage/doc.ts

@@ -13,9 +13,15 @@ import {
 } from 'yjs';
 
 import { CallMetric } from '../../../base';
+import { mergeUpdatesInApplyWay } from '../../../native';
 import { Connection } from './connection';
 import { SingletonLocker } from './lock';
 
+async function nativeMergeUpdates(updates: Uint8Array[]): Promise<Uint8Array> {
+  // use native module to merge updates
+  return mergeUpdatesInApplyWay(updates.map(u => Buffer.from(u)));
+}
+
 export interface DocRecord {
   spaceId: string;
   docId: string;
@@ -95,6 +101,27 @@ export abstract class DocStorageAdapter extends Connection {
     return snapshot;
   }
 
+  /// get final binary only but not updating the snapshot in database
+  async getDocBinNative(
+    spaceId: string,
+    docId: string
+  ): Promise<Uint8Array | undefined> {
+    await using _lock = await this.lockDocForUpdate(spaceId, docId);
+
+    const snapshot = await this.getDocSnapshot(spaceId, docId);
+    const updates = await this.getDocUpdates(spaceId, docId);
+
+    if (updates.length) {
+      const docUpdate = await this.squash(
+        snapshot ? [snapshot, ...updates] : updates,
+        nativeMergeUpdates
+      );
+      return docUpdate.bin;
+    }
+
+    return snapshot?.bin;
+  }
+
   @Transactional<TransactionalAdapterPrisma>({ timeout: 60000 })
   private async squashUpdatesToSnapshot(
     spaceId: string,
@@ -223,8 +250,11 @@ export abstract class DocStorageAdapter extends Connection {
   ): Promise<boolean>;
 
   @CallMetric('doc', 'squash')
-  protected async squash(updates: DocUpdate[]): Promise<DocUpdate> {
-    const merge = this.options?.mergeUpdates ?? mergeUpdates;
+  protected async squash(
+    updates: DocUpdate[],
+    merge?: (updates: Uint8Array[]) => Promise<Uint8Array>
+  ): Promise<DocUpdate> {
+    const mergeFn = merge ?? this.options?.mergeUpdates ?? mergeUpdates;
     const lastUpdate = updates.at(-1);
     if (!lastUpdate) {
       throw new Error('No updates to be squashed.');
@@ -235,7 +265,7 @@ export abstract class DocStorageAdapter extends Connection {
       return lastUpdate;
     }
 
-    const finalUpdate = await merge(updates.map(u => u.bin));
+    const finalUpdate = await mergeFn(updates.map(u => u.bin));
 
     return {
       bin: finalUpdate,