Fix incorrect IDs in AP federation

[github/Chocobozzz/PeerTube.git] / server / lib / activitypub / crawl.ts

diff --git a/server/lib/activitypub/crawl.ts b/server/lib/activitypub/crawl.ts
index 686eef04d9d0f4b77c5cdae05debe5c878320463..eeafdf4ba8d04848242020af795f8fb4388cda8b 100644 (file)
--- a/server/lib/activitypub/crawl.ts
+++ b/server/lib/activitypub/crawl.ts
@@ -3,7 +3,7 @@ import { doRequest } from '../../helpers/requests'
 import { logger } from '../../helpers/logger'
 import * as Bluebird from 'bluebird'
 import { ActivityPubOrderedCollection } from '../../../shared/models/activitypub'
-import { parse } from 'url'
+import { URL } from 'url'
 
 type HandlerFunction<T> = (items: T[]) => (Promise<any> | Bluebird<any>)
 type CleanerFunction = (startedDate: Date) => (Promise<any> | Bluebird<any>)
@@ -24,17 +24,26 @@ async function crawlCollectionPage <T> (uri: string, handler: HandlerFunction<T>
   const response = await doRequest<ActivityPubOrderedCollection<T>>(options)
   const firstBody = response.body
 
-  let limit = ACTIVITY_PUB.FETCH_PAGE_LIMIT
+  const limit = ACTIVITY_PUB.FETCH_PAGE_LIMIT
   let i = 0
   let nextLink = firstBody.first
   while (nextLink && i < limit) {
-    // Don't crawl ourselves
-    const remoteHost = parse(nextLink).host
-    if (remoteHost === WEBSERVER.HOST) continue
+    let body: any
 
-    options.uri = nextLink
+    if (typeof nextLink === 'string') {
+      // Don't crawl ourselves
+      const remoteHost = new URL(nextLink).host
+      if (remoteHost === WEBSERVER.HOST) continue
+
+      options.uri = nextLink
+
+      const res = await doRequest<ActivityPubOrderedCollection<T>>(options)
+      body = res.body
+    } else {
+      // nextLink is already the object we want
+      body = nextLink
+    }
 
-    const { body } = await doRequest<ActivityPubOrderedCollection<T>>(options)
     nextLink = body.next
     i++