From da8cd870b0afac4e84504fc30c4d94643019943a Mon Sep 17 00:00:00 2001
From: Matt Carey <mcarey@cloudflare.com>
Date: Wed, 8 Jul 2026 13:08:38 +0100
Subject: [PATCH] fix(sync): pull every path by default

---
 docs/02_sync_protocol.md             |  8 +++-----
 packages/dofs/src/sync/ignore.ts     |  2 +-
 packages/rpc/src/sync-driver.test.ts | 23 ++++++++++++++++++++---
 packages/rpc/src/sync-driver.ts      |  4 +++-
 4 files changed, 27 insertions(+), 10 deletions(-)

diff --git a/docs/02_sync_protocol.md b/docs/02_sync_protocol.md
index d5699e09..46c9f936 100644
--- a/docs/02_sync_protocol.md
+++ b/docs/02_sync_protocol.md
@@ -407,7 +407,6 @@ first-class conflict primitives.
 
 ## Ignore lists
 
-
 The `ignore` option hides path segments from the pull. Excluded
 paths are still written and read inside the container — the bytes just
 never cross the wire back to the DO. This is essential for any large
@@ -416,10 +415,9 @@ directory of derived files: `node_modules`, `.next`, `target`,
 push tens of thousands of small files through the sync wire on the
 next pull.
 
-The default is `["node_modules"]`, applied server-side when `ignore` is
-omitted. A caller-supplied list **replaces** the default — it does not
-extend it. Pass `[]` to disable ignoring entirely, or pass your full
-list (including `"node_modules"` if you still want it) to customise.
+No paths are ignored by default. A sync server can configure an ignore list,
+and an individual `fetchChanges` call can replace it. Pass `[]` to override a
+server-level list and request every path.
 
 ### Ignored entries
 
diff --git a/packages/dofs/src/sync/ignore.ts b/packages/dofs/src/sync/ignore.ts
index 91f7b841..47dd489c 100644
--- a/packages/dofs/src/sync/ignore.ts
+++ b/packages/dofs/src/sync/ignore.ts
@@ -8,7 +8,7 @@
 // node_modules_old or my_node_modules. Patterns are plain strings,
 // not globs; we can extend to globs later if a real case demands it.
 
-export const DEFAULT_IGNORE = ["node_modules"];
+export const DEFAULT_IGNORE: string[] = [];
 
 export function isIgnored(path: string, patterns: string[]): boolean {
   if (patterns.length === 0) return false;
diff --git a/packages/rpc/src/sync-driver.test.ts b/packages/rpc/src/sync-driver.test.ts
index ad78667b..278e6832 100644
--- a/packages/rpc/src/sync-driver.test.ts
+++ b/packages/rpc/src/sync-driver.test.ts
@@ -88,15 +88,31 @@ function trackFetchDisposal(
 }
 
 describe("sync driver — pullOnce", () => {
-  it("pulls a single entry from upstream", async () => {
+  it("pulls every path by default, including node_modules", async () => {
     const a = makePeer();
     const b = makePeer();
     try {
       const provider = new SQLiteWorkspaceProvider(a.db, { now: () => 1 });
       provider.writeFileSync("/hello.txt", "hello");
+      provider.mkdirSync("/node_modules/pkg", { recursive: true });
+      provider.writeFileSync("/node_modules/pkg/index.js", "dependency");
 
-      const applied = await pullOnce(b.db, a.rpc);
-      expect(applied.applied).toBe(1);
+      let requestedIgnore: string[] | undefined;
+      const remote = new Proxy(a.rpc as object, {
+        get(target, prop, receiver) {
+          if (prop === "fetchChanges") {
+            return (input: Parameters<SyncRPC["fetchChanges"]>[0]) => {
+              requestedIgnore = input.ignore;
+              return Reflect.get(target, prop, receiver).call(target, input);
+            };
+          }
+          return Reflect.get(target, prop, receiver);
+        },
+      }) as SyncRPC;
+
+      const applied = await pullOnce(b.db, remote);
+      expect(requestedIgnore).toEqual([]);
+      expect(applied.applied).toBeGreaterThan(1);
       expect(applied.skipped).toEqual([]);
       expect(fileEntries(b.db)).toContain("hello.txt");
       // Asserting the bytes arrived, not just the dirent. The
@@ -105,6 +121,7 @@ describe("sync driver — pullOnce", () => {
       // empty on the receiver — RPC reads landed HTTP 200 / 0 bytes.
       const providerB = new SQLiteWorkspaceProvider(b.db, { now: () => 1 });
       expect(providerB.readFileSync("/hello.txt", "utf8")).toBe("hello");
+      expect(providerB.readFileSync("/node_modules/pkg/index.js", "utf8")).toBe("dependency");
     } finally {
       a.close();
       b.close();
diff --git a/packages/rpc/src/sync-driver.ts b/packages/rpc/src/sync-driver.ts
index 75028ba3..1cf7176c 100644
--- a/packages/rpc/src/sync-driver.ts
+++ b/packages/rpc/src/sync-driver.ts
@@ -103,7 +103,9 @@ async function pullOnceImpl(
   // invariant trip, and any throw inside the batch loop. Disposing the
   // envelope tears down the contained stream stub, releasing the
   // remote iterator.
-  const fetchResult = await remote.fetchChanges({ after });
+  // Pull the complete remote tree. Sending [] explicitly also overrides the
+  // historical wsd default that filtered node_modules when ignore was omitted.
+  const fetchResult = await remote.fetchChanges({ after, ignore: [] });
   try {
     const { currentCursor, appliedPushCursor } = fetchResult;
     // Cross-side watermark divergence. Two shapes are recoverable: