PeerStorage.delete() treats ENOENT as failure, spamming logs during catch-up

[pike00]

Summary

PeerStorage.delete() catches all exceptions from Deno.remove() and logs "delete failed" at LOG_LEVEL_NOTICE, returning false. When CouchDB tombstones arrive for paths that were already removed while the bridge was down (or never existed on this host), ENOENT is expected — the file is already gone. Treating it as a failure spams the log and returns false to the caller, which may affect downstream retry logic.

Reproduction

1. Delete a file on machine A while the bridge on machine B is down.
2. Start the bridge on machine B.
3. The CouchDB change feed delivers a tombstone for the deleted path.
4. PeerStorage.delete() calls Deno.remove() → ENOENT → logs "delete failed" at NOTICE level.

Fix

Catch Deno.errors.NotFound specifically and return true — the file is already absent, which is the desired end state.

 async delete(pathSrc: string): Promise<boolean> {
     try {
         await Deno.remove(path);
         this.receiveLog(\` ${path} deleted\`);
     } catch (ex) {
+        if (ex instanceof Deno.errors.NotFound) {
+            this.receiveLog(\` ${path} already absent\`);
+            return true;
+        }
         this.receiveLog(\` ${path} delete failed\`, LOG_LEVEL_NOTICE);
         Logger(ex, LOG_LEVEL_VERBOSE);
         return false;
     }
 }

[pike00]

I plan on implementing this. I'll submit a PR once it's ready.