Merge pull request #394 from automerge/denylist

Add a denylist of automerge URLs to Repo constructor

packages/automerge-repo/src/Repo.ts

@@ -31,7 +31,12 @@ import {
   DocSyncMetrics,
   SyncStatePayload,
 } from "./synchronizer/Synchronizer.js"
-import type { AnyDocumentId, DocumentId, PeerId } from "./types.js"
+import type {
+  AnyDocumentId,
+  AutomergeUrl,
+  DocumentId,
+  PeerId,
+} from "./types.js"
 
 function randomPeerId() {
   return ("peer-" + Math.random().toString(36).slice(4)) as PeerId
@@ -80,6 +85,7 @@ export class Repo extends EventEmitter<RepoEvents> {
     sharePolicy,
     isEphemeral = storage === undefined,
     enableRemoteHeadsGossiping = false,
+    denylist = [],
   }: RepoConfig = {}) {
     super()
     this.#remoteHeadsGossipingEnabled = enableRemoteHeadsGossiping
@@ -99,7 +105,7 @@ export class Repo extends EventEmitter<RepoEvents> {
 
     // SYNCHRONIZER
     // The synchronizer uses the network subsystem to keep documents in sync with peers.
-    this.synchronizer = new CollectionSynchronizer(this)
+    this.synchronizer = new CollectionSynchronizer(this, denylist)
 
     // When the synchronizer emits messages, send them to peers
     this.synchronizer.on("message", message => {
@@ -627,6 +633,13 @@ export interface RepoConfig {
    * Whether to enable the experimental remote heads gossiping feature
    */
   enableRemoteHeadsGossiping?: boolean
+
+  /**
+   * A list of automerge URLs which should never be loaded regardless of what
+   * messages are received or what the share policy is. This is useful to avoid
+   * loading documents that are known to be too resource intensive.
+   */
+  denylist?: AutomergeUrl[]
 }
 
 /** A function that determines whether we should share a document with a peer
@@ -670,3 +683,7 @@ export type DocMetrics =
       numOps: number
       numChanges: number
     }
+  | {
+      type: "doc-denied"
+      documentId: DocumentId
+    }

packages/automerge-repo/src/synchronizer/CollectionSynchronizer.ts

@@ -1,9 +1,9 @@
 import debug from "debug"
 import { DocHandle } from "../DocHandle.js"
-import { stringifyAutomergeUrl } from "../AutomergeUrl.js"
+import { parseAutomergeUrl, stringifyAutomergeUrl } from "../AutomergeUrl.js"
 import { Repo } from "../Repo.js"
 import { DocMessage } from "../network/messages.js"
-import { DocumentId, PeerId } from "../types.js"
+import { AutomergeUrl, DocumentId, PeerId } from "../types.js"
 import { DocSynchronizer } from "./DocSynchronizer.js"
 import { Synchronizer } from "./Synchronizer.js"
 
@@ -21,8 +21,11 @@ export class CollectionSynchronizer extends Synchronizer {
   /** Used to determine if the document is know to the Collection and a synchronizer exists or is being set up */
   #docSetUp: Record<DocumentId, boolean> = {}
 
-  constructor(private repo: Repo) {
+  #denylist: DocumentId[]
+
+  constructor(private repo: Repo, denylist: AutomergeUrl[] = []) {
     super()
+    this.#denylist = denylist.map(url => parseAutomergeUrl(url).documentId)
   }
 
   /** Returns a synchronizer for the given document, creating one if it doesn't already exist.  */
@@ -91,6 +94,19 @@ export class CollectionSynchronizer extends Synchronizer {
       throw new Error("received a message with an invalid documentId")
     }
 
+    if (this.#denylist.includes(documentId)) {
+      this.emit("metrics", {
+        type: "doc-denied",
+        documentId,
+      })
+      this.emit("message", {
+        type: "doc-unavailable",
+        documentId,
+        targetId: message.senderId,
+      })
+      return
+    }
+
     this.#docSetUp[documentId] = true
 
     const docSynchronizer = this.#fetchDocSynchronizer(documentId)

packages/automerge-repo/src/synchronizer/Synchronizer.ts

@@ -25,10 +25,15 @@ export interface SyncStatePayload {
   syncState: SyncState
 }
 
-export type DocSyncMetrics = {
-  type: "receive-sync-message"
-  documentId: DocumentId
-  durationMillis: number
-  numOps: number
-  numChanges: number
-}
+export type DocSyncMetrics =
+  | {
+      type: "receive-sync-message"
+      documentId: DocumentId
+      durationMillis: number
+      numOps: number
+      numChanges: number
+    }
+  | {
+      type: "doc-denied"
+      documentId: DocumentId
+    }

packages/automerge-repo/test/Repo.test.ts

@@ -1486,6 +1486,44 @@ describe("Repo", () => {
       teardown()
     })
   })
+
+  describe("the denylist", () => {
+    it("should immediately return an unavailable message in response to a request for a denylisted document", async () => {
+      const storage = new DummyStorageAdapter()
+
+      // first create the document in storage
+      const dummyRepo = new Repo({ network: [], storage })
+      const doc = dummyRepo.create({ foo: "bar" })
+      await dummyRepo.flush()
+
+      // Check that the document actually is in storage
+      let docId = doc.documentId
+      assert(storage.keys().some((k: string) => k.includes(docId)))
+
+      const channel = new MessageChannel()
+      const { port1: clientToServer, port2: serverToClient } = channel
+      const server = new Repo({
+        network: [new MessageChannelNetworkAdapter(serverToClient)],
+        storage,
+        denylist: [doc.url],
+      })
+      const client = new Repo({
+        network: [new MessageChannelNetworkAdapter(clientToServer)],
+      })
+
+      await Promise.all([
+        eventPromise(server.networkSubsystem, "peer"),
+        eventPromise(client.networkSubsystem, "peer"),
+      ])
+
+      const clientDoc = client.find(doc.url)
+      await pause(100)
+      assert.strictEqual(clientDoc.docSync(), undefined)
+
+      const openDocs = Object.keys(server.metrics().documents).length
+      assert.deepEqual(openDocs, 0)
+    })
+  })
 })
 
 const warn = console.warn