@ai-matrx/messaging 0.7.0 → 0.8.0

package/CHANGELOG.md

@@ -1,5 +1,26 @@
 # Changelog — `@ai-matrx/messaging`
 
+## 0.8.0 — 2026-09-07
+
+**Every conversation in the first real inbox was titled with a raw UUID.** The second live defect
+of the adoption session, found by looking at the screen rather than at a test.
+
+### Fixed
+
+- **The participant profile is NESTED under `user`, and the top-level `id` is the participant
+  ROW.** The live `get_dm_conversations_with_details` aggregate is
+  `{ id, user_id, role, is_muted, last_read_at, user: { user_id, email, display_name, avatar_url } }`.
+  This package read `display_name` / `email` / `avatar_url` off the TOP level, found nothing, and
+  fell back to the user id — so a direct conversation's title, its avatar, and every participant
+  name rendered as a UUID. The nested profile is now the source, `user_id` is the identity, and
+  the bare `id` fallback survives only for an aggregate that carries neither — because here `id`
+  belongs to a different entity, and using it as a user id attributes a message to a row key.
+- A regression test built from the verbatim live aggregate; it fails against 0.7.0.
+
+### Consumer action (C28)
+
+None. Names and avatars simply appear now.
+
 ## 0.7.0 — 2026-09-07
 
 🚨 **Every RPC in this package was calling the wrong schema, so every read failed against the

package/dist/index.cjs

@@ -798,11 +798,13 @@ function projectParticipants(value, operation) {
   for (const entry of value) {
     if (typeof entry !== "object" || entry === null) continue;
     const record = entry;
-    const id = str(record, "user_id") ?? str(record, "id");
+    const nested = typeof record["user"] === "object" && record["user"] !== null ? record["user"] : {};
+    const id = str(record, "user_id") ?? str(nested, "user_id") ?? str(record, "id");
     if (id === null) continue;
     summaries.push(
       projectUserSummary({
         ...record,
+        ...nested,
         user_id: id
       })
     );