From 916088a22fee4cca76de67d72224d7fab045cc1a Mon Sep 17 00:00:00 2001
From: Olivia Lee <olivia@computer.surgery>
Date: Fri, 29 Nov 2024 11:33:13 -0800
Subject: [PATCH] include mxcs from dangling thumbnails in
 service::media::iter_all

When requesting remote thumbnails over federation, we can end up with a
thumbnail in the media db without an associated original file. Because
of this, skipping thumbnails is insufficient to get a list of all MXCs.
---
 src/database/key_value/media.rs |  5 -----
 src/service/media.rs            | 20 ++++++++++++++++++--
 src/service/media/data.rs       |  4 +---
 3 files changed, 19 insertions(+), 10 deletions(-)

diff --git a/src/database/key_value/media.rs b/src/database/key_value/media.rs
index a7272042..53e2146e 100644
--- a/src/database/key_value/media.rs
+++ b/src/database/key_value/media.rs
@@ -208,11 +208,6 @@ impl service::media::Data for KeyValueDatabase {
                     let key = MediaFileKey::new(key);
 
                     let parts = MediaFileKeyParts::try_from(&key)?;
-                    if parts.width != 0 || parts.height != 0 {
-                        // Skip thumbnails
-                        return Ok(None);
-                    };
-
                     Ok(Some((parts.mxc, parts.meta, key)))
                 })
                 .filter_map(Result::transpose),
diff --git a/src/service/media.rs b/src/service/media.rs
index b9ccddae..836e8ef9 100644
--- a/src/service/media.rs
+++ b/src/service/media.rs
@@ -169,11 +169,27 @@ impl Service {
 
     /// List all media stored in the database.
     ///
-    /// Each MXC is list once. Thumbnails are not included separately from the
+    /// Each MXC is listed once. Thumbnails are not included separately from the
     /// original media.
     #[tracing::instrument(skip(self))]
     pub(crate) fn iter_all(&self) -> impl Iterator<Item = Result<OwnedMxcUri>> {
-        self.db.all_file_metadata().map(|media| media.map(|(mxc, ..)| mxc))
+        let mut prev_mxc = None;
+        self.db
+            .all_file_metadata()
+            .map(|media| media.map(|(mxc, ..)| mxc))
+            .filter(move |mxc| {
+                if let Ok(mxc) = mxc {
+                    // Skip mxcs that we have already seen. All files associated
+                    // with a given mxc should appear consecutively in the db
+                    // iterator, so we only need to check against the previous
+                    // value.
+                    if prev_mxc.as_ref() == Some(mxc) {
+                        return false;
+                    }
+                    prev_mxc = Some(mxc.clone());
+                }
+                true
+            })
     }
 
     /// Returns width, height of the thumbnail and whether it should be cropped.
diff --git a/src/service/media/data.rs b/src/service/media/data.rs
index 01b2ceb1..c3cff8cb 100644
--- a/src/service/media/data.rs
+++ b/src/service/media/data.rs
@@ -30,9 +30,7 @@ pub(crate) trait Data: Send + Sync {
         mxc: OwnedMxcUri,
     ) -> Result<Vec<(FileMeta, MediaFileKey)>>;
 
-    /// Returns an iterator over metadata for all media.
-    ///
-    /// Thumbnails are not included.
+    /// Returns an iterator over metadata for all media, including thumbnails.
     fn all_file_metadata(
         &self,
     ) -> Box<