From: Felix Ableitner <me@nutomic.com>
Date: Thu, 18 May 2023 11:22:05 +0000 (+0200)
Subject: Dont refetch post url metadata when post is received again
X-Git-Url: http://these/git/%7B%60%24%7BwebArchiveUrl%7D/%22%7B%7D/%22https:/nerdica.net/%7Biframely.thumbnail_url%7D?a=commitdiff_plain;h=436a293f382e132763de80f78b9475ee1b5e0ea5;p=lemmy.git

Dont refetch post url metadata when post is received again
---

diff --git a/crates/apub/src/objects/post.rs b/crates/apub/src/objects/post.rs
index 7bc93339..5b1583a2 100644
--- a/crates/apub/src/objects/post.rs
+++ b/crates/apub/src/objects/post.rs
@@ -185,6 +185,9 @@ impl Object for ApubPost {
       name = name.chars().take(MAX_TITLE_LENGTH).collect();
     }
 
+    // read existing, local post if any (for generating mod log)
+    let old_post = page.id.dereference_local(context).await;
+
     let form = if !page.is_mod_action(context).await? {
       let first_attachment = page.attachment.into_iter().map(Attachment::url).next();
       let url = if first_attachment.is_some() {
@@ -195,10 +198,13 @@ impl Object for ApubPost {
       } else {
         None
       };
-      let (metadata_res, thumbnail_url) = if let Some(url) = &url {
-        fetch_site_data(context.client(), context.settings(), Some(url)).await
-      } else {
-        (None, page.image.map(|i| i.url.into()))
+      // Only fetch metadata if the post has a url and was not seen previously. We dont want to
+      // waste resources by fetching metadata for the same post multiple times.
+      let (metadata_res, thumbnail_url) = match &url {
+        Some(url) if old_post.is_err() => {
+          fetch_site_data(context.client(), context.settings(), Some(url)).await
+        }
+        _ => (None, page.image.map(|i| i.url.into())),
       };
       let (embed_title, embed_description, embed_video_url) = metadata_res
         .map(|u| (u.title, u.description, u.embed_video_url))
@@ -245,8 +251,6 @@ impl Object for ApubPost {
         .updated(page.updated.map(|u| u.naive_local()))
         .build()
     };
-    // read existing, local post if any (for generating mod log)
-    let old_post = page.id.dereference_local(context).await;
 
     let post = Post::create(context.pool(), &form).await?;