feat(server): improve doc gc (#15363)

#### PR Dependency Tree


* **PR #15363** 👈

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

* **Bug Fixes**
* Improved validation of workspace roots and document projections, with
clearer failures for malformed or incomplete data.
  * Improved document reference rebuilding and cleanup reliability.
* Updated document update merging to better handle invalid binary data.

* **Performance**
* Avoided unnecessary document reconstruction when no updates are
pending.

* **Tests**
* Updated coverage for malformed workspace roots and document snapshot
parsing.

<!-- end of auto-generated comment: release notes by coderabbit.ai -->

packages/frontend/native/nbstore/src/indexer.rs

@@ -3,7 +3,7 @@ use memory_indexer::{SearchHit, SnapshotData};
 use napi_derive::napi;
 use serde::Serialize;
 use sqlx::Row;
-use y_octo::DocOptions;
+use y_octo::merge_updates_v1;
 
 // Increment this whenever there is a breaking change in the index format or how
 // updates are applied
@@ -120,7 +120,7 @@ impl SqliteDocStorage {
     }
     segments.extend(updates.into_iter().map(|update| update.bin.to_vec()));
 
-    merge_updates(segments, doc_id).map(Some)
+    merge_updates(segments).map(Some)
   }
 
   pub async fn init_index(&self) -> Result<()> {
@@ -249,7 +249,7 @@ impl SqliteDocStorage {
   }
 }
 
-fn merge_updates(mut segments: Vec<Vec<u8>>, guid: &str) -> Result<Vec<u8>> {
+fn merge_updates(mut segments: Vec<Vec<u8>>) -> Result<Vec<u8>> {
   if segments.is_empty() {
     return Err(ParseError::DocNotFound.into());
   }
@@ -258,15 +258,9 @@ fn merge_updates(mut segments: Vec<Vec<u8>>, guid: &str) -> Result<Vec<u8>> {
     return segments.pop().ok_or(ParseError::DocNotFound.into());
   }
 
-  let mut doc = DocOptions::new().with_guid(guid.to_string()).build();
-  for update in segments.iter() {
-    doc
-      .apply_update_from_binary_v1(update)
-      .map_err(|_| ParseError::InvalidBinary)?;
-  }
-
-  let buffer = doc
-    .encode_update_v1()
+  let update = merge_updates_v1(segments).map_err(|_| ParseError::InvalidBinary)?;
+  let buffer = update
+    .encode_v1()
     .map_err(|err| ParseError::ParserError(err.to_string()))?;
 
   Ok(buffer)
@@ -317,6 +311,13 @@ mod tests {
       .execute(&storage.pool)
       .await
       .unwrap();
+    sqlx::query(r#"INSERT INTO updates (doc_id, data, created_at) VALUES (?, ?, ?)"#)
+      .bind("demo-doc")
+      .bind(&[0, 0][..])
+      .bind(Utc::now().naive_utc())
+      .execute(&storage.pool)
+      .await
+      .unwrap();
 
     let result = storage.crawl_doc_data("demo-doc").await.unwrap();