fix: several bugs (#4500)

1. Fixes #4466 (we did not check validity of the scan cursor)
   We fix it by quietly returning "end of scan" result, i.e. no error are returned.
2. Add 34b1048274c8e50a0cc587a3ed9c383a82bb78c5 sha to the lua_undeclared_keys_shas list.
3. Pull helio that fixes bugs with WaitQueue and IdleTasks: romange/helio@2ab4412

src/core/dash.h

@@ -954,12 +954,13 @@ auto DashTable<_Key, _Value, Policy>::GetRandomCursor(absl::BitGen* bitgen) -> C
 template <typename _Key, typename _Value, typename Policy>
 template <typename Cb>
 auto DashTable<_Key, _Value, Policy>::Traverse(Cursor curs, Cb&& cb) -> Cursor {
-  if (curs.bucket_id() >= Policy::kBucketNum)  // sanity.
-    return 0;
-
   uint32_t sid = curs.segment_id(global_depth_);
   uint8_t bid = curs.bucket_id();
 
+  // Test validity of the cursor.
+  if (bid >= Policy::kBucketNum || sid >= segment_.size())
+    return 0;
+
   auto hash_fun = [this](const auto& k) { return policy_.HashFn(k); };
 
   bool fetched = false;

src/server/generic_family_test.cc

@@ -872,4 +872,9 @@ TEST_F(GenericFamilyTest, RestoreOOM) {
   EXPECT_THAT(resp, ErrArg("Out of memory"));
 }
 
+TEST_F(GenericFamilyTest, Bug4466) {
+  auto resp = Run({"SCAN","9223372036854775808"});  // an invalid cursor should not crash us.
+  EXPECT_THAT(resp, RespElementsAre("0", RespElementsAre()));
+}
+
 }  // namespace dfly

src/server/script_mgr.cc

@@ -43,6 +43,7 @@ ABSL_FLAG(
     std::vector<std::string>({
         "351130589c64523cb98978dc32c64173a31244f3",  // Sidekiq, see #2442
         "6ae15ef4678593dc61f991c9953722d67d822776",  // Sidekiq, see #2442
+        "34b1048274c8e50a0cc587a3ed9c383a82bb78c5"   // Sidekiq
     }),
     "Comma-separated list of Lua script SHAs which are allowed to access undeclared keys. SHAs are "
     "only looked at when loading the script, and new values do not affect already-loaded script.");