@remit/backend 0.0.15 → 0.0.17

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@remit/backend",
-  "version": "0.0.15",
+  "version": "0.0.17",
   "description": "Remit Mail Inspector API backend",
   "license": "MIT",
   "author": "",

package/src/handlers/me.test.ts

@@ -35,7 +35,7 @@ const entry = (over: Partial<QuarantineItem> = {}): QuarantineItem =>
 		quarantinedAt: 1_000,
 		attempts: 3,
 		failureStage: "BodyParse",
-		failureCode: "UnterminatedMultipartBoundary",
+		failureCode: "UnreadableBody",
 		failureMessage: "multipart boundary was never closed",
 		workerVersion: "worker 1.0.0",
 		structure: [{ depth: 0, contentType: "multipart/mixed" }],

package/src/handlers/thread.ts

@@ -270,11 +270,12 @@ export const buildListThreadMessagesOptions = (query: {
  * A ThreadMessage row is keyed by (threadId, messageId), and a messageId is
  * derived from the account plus the RFC 5322 Message-ID, so the same mail
  * synced from two folders is one row — the mailbox a thread spans is not what
- * duplicates it. `MessageMoveService.copyMessage` is the exception: it mints a
- * random id for the copy, so a copied message becomes a second row in the same
- * thread carrying the same `messageIdHeader`. Reading the whole thread rather
- * than one mailbox of it (#46) is what makes both rows visible at once, and two
- * entries for one message is never what a conversation should show.
+ * duplicates it. `MessageMoveService.copyMessage` is the exception: a copy is a
+ * per-folder placement with its own derived id (`deriveCopyMessageId`), so a
+ * copied message is a second row in the same thread carrying the same
+ * `messageIdHeader`. Reading the whole thread rather than one mailbox of it
+ * (#46) is what makes both rows visible at once, and two entries for one message
+ * is never what a conversation should show.
  *
  * The oldest row wins, so the original outlives the copy, with `threadMessageId`
  * breaking ties. Both are independent of sort order, so `asc` and `desc` keep